Wikipedia:Village pump (technical)/Archive 111

• Village pump
  • Policy
  • Technical
  • Proposals (persistent)
  • Idea lab
  • WMF
  • Miscellaneous

Village pump (technical) archive  

This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.

---

< Older discussions · Archives: A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, V, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX · 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212

Transliteration in searching

Hello, the search engine doesn't know to associate characters with their diacritics. For example if your search word has an "a", it won't suggest the same word that contains "á" instead of "a".

On Romanian Wikipedia, when I search for Carol Popp de Szatmary it shows only 3 results, and it won't include the ro:Carol Popp de Szathmári article, which is almost the same with the text I was searching. You can check that for yourself: [1] or search Szatmary

Another example: search for „Agri” and the result doesn't include ro:Ağrı.

I already reported the bug here: Help talk:Searching#Transliteration in searching - Help talk:Searching/Archive 4#Transliteration in searching. If the auto-suggest drop-down list doesn't know to make those associations - fine (even though 2013 sounds like a lot). But the search engine should definitely know how to make such associations, in my opinion. Thanks. —  Ark25  (talk) 06:04, 26 April 2013 (UTC)

Well, I made an error, I haven't noticed the extra "h" in the "Szathmári" word. However, the "Ağrı" example is valid. I think the search engine should associate "ğ" with "g". —  Ark25  (talk) 08:32, 26 April 2013 (UTC)

As far as I can tell, the problem character in your other example is not ğ but ı (lower case dotless I). A search on agrı does find ro:Ağrı, but it doesn't find ro:Agri. The character is treated oddly. If it's the first character in a wikilink then it automatically works as i, for example in ı or ıce. If it's not the first character then it makes a redlink, for example kıte. If you enter ı in the search box and there is a title match with i instead of ı then it goes there, for example "kıte" goes to kite (with no redirect). But a search on kıte finds nothing. ı and i appear identical to the "Go" feature of the search box, but they aren't associated at all in searches. PrimeHunter (talk) 10:26, 26 April 2013 (UTC)

You are right. Also it jumps directly if ı it's the last character: typing in the searchbox "Ağri" and pressing enter - will jump directly to Ağrı, just like in the ıce case (and I think it should not jump there, it should just show the article on the top of the search result list). But if you search for "Ağri" the Ağrı article will not show in the search result list. —  Ark25  (talk) 10:43, 26 April 2013 (UTC)

That ıce = Ice = ice is undoubtedly due to the fact that the uppercase mapping of ı, as defined in the Unicode standard, is I.[2]—Emil J. 11:43, 26 April 2013 (UTC)

In fact, this probably also explains why “kıte” goes to kite from the search box: the “Go” feature appears to fold all letters to uppercase before attempting a match—if you type in “kItE”, it also goes to the same article.—Emil J. 11:54, 26 April 2013 (UTC)

Ah yes, {{uc:ı}} returns I, the normal upper case letter. This means that {{lc:{{uc:ı}}}} returns i, the normal lower case letter and not the one you started with. bugzilla:33643 is related. I don't know whether there is a bugzilla entry for i and ı not being associated in searches. PrimeHunter (talk) 13:23, 26 April 2013 (UTC)

Section title "AD" disappears

I type:

Above section header
===AD===
Below section header

And I see rendered:

Above section header

AD

Below section header

AD is the prefix of some old vacuum tubes. I can't jump there from the table of contents, either. I had to insert a blank between A and D as a workaround.

Is that some misguided spam filter? --Mkratz (talk) 16:27, 26 April 2013 (UTC)

Oh, I just found out. It's the AdBlock Plus extension in Firefox. Lots of people have that installed. Now what? --Mkratz (talk) 16:40, 26 April 2013 (UTC)

I don't have any problems seeing it with AdBlock enabled (though I usually have it disabled on en.wikipedia.org). It will surely depend on which filters you use in AdBlock. — HHHIPPO 17:21, 26 April 2013 (UTC)

Right. Seems I've neglected to choose the filter, so AdBlock defaulted to FanBoy's. No problems anymore with EasyList.

Yet, I guess I better leave the workaround in place. --Mkratz (talk) 18:14, 26 April 2013 (UTC)

Indeed, FanBoy's filter is blocking the word "ad". I wonder if that makes sense. How about this workaround: ===AD === ? That looks basically the same, but tricks the filter. — HHHIPPO 20:51, 26 April 2013 (UTC)

Is FanBoy contactable? Could he be asked to correct his filters to not cause the false match? – PartTimeGnome (talk | contribs) 21:07, 26 April 2013 (UTC)

Might trick FanBoy, but there are lots of others. Better steer clear from the treshold. --Mkratz (talk) 22:18, 26 April 2013 (UTC)

Skin issues

I'm trying to replicate the appearance of the Classic skin within Vector, and would appreciate some guidance in customizing my (already-heavily-modified) css and js files. I was getting some help on meta:Tech, but that seems to have dried up for now. Note that I didn't write these modifications; other people very generously wrote them for me. Current issues include:

• I'd like a "my contributions" link on all pages.
• I'd like a "Watchlist" link on all pages.
• I'd like a "current version" link on pages where I compare two diffs.
• I'd like the "new messages" notifier to be at the very top of the page, rather than at the top of the article. Note that I already have modifications to remove the orange bar, because the orange bar is hideously ugly.
• The vertical separator bar is an ugly shade of light blue. I'd like it to be a pleasant gray (RGB: 128, 128, 128). I know how to change most of the font colors already (and am actively doing it because the colors that Vector uses are ugly), but not the separator bar colors.
• On a related point, the vertical separator bar goes too far down. It should stop just past the first four lines of the article (or equivalent). And no horizontal separator bar at the very bottom of the page, thank you.
• The relative font sizes of the various elements are wrong. Currently, the font size of article text is (X), but the font size of text in the edit window is (Y). I would like these to be inverted. I would also like the text in the sidebar to be (X) rather than (Y).
• I would like the sidebar to be somewhat narrower - like how it was in Classic.
• I would like to completely eliminate all use of the "star" icon as it pertains to the "watch page" function.
• There's an unsightly empty space at the top of the page (see File:Classic vs heavily-modified Vector (skin comparison) 1.jpg, File:Classic vs heavily-modified Vector (skin comparison) 2.jpg, and File:Classic vs heavily-modified Vector (skin comparison) 3.jpg -- but it's present in unmodified Vector too, just smaller). I'd like it refilled with the links shown in the comparison jpgs.

Thank you in advance for any advice or suggestions you can provide. DS (talk) 02:23, 27 April 2013 (UTC)

H.264 support

Tracked in Phabricator
Task T41869

Developments since Google reneged on dropping H.264 two years ago, possibly due to acquiring Motorola patents:

• Wikimedia rolls out TimedMediaHandler supporting H.264 transcoding
• WMF releases iPad app without Ogg or WebM support
• Mozilla will support H.264 when the underlying platform supports it
• Debian 7.0 "Wheezy" will decode H.264 media out of the box

Now that every modern platform supports it, so we're too late to pressure anyone into switching. So I suggest three transition phases: 1) Allow H.264 files masquerading as ogv/webm, 2) Allow uploading of H.264 and .mp4 extensions, 3) Enable transcoding to H.264. — Dispenser 05:38, 19 April 2013 (UTC)

This was never about pressuring anyone into switching. It's about Free and open information systems and not opening up ourselves to potential lawsuits by big techonlogy owners. Note that unlike for Mozilla for instance, we ARE part of those underlying systems (esp when it comes to transcoding), so there is quite a different risk assessment. —TheDJ (talk • contribs) 10:34, 19 April 2013 (UTC)

My post was to inform that all major platforms support H.264 video decoding, even Firefox on Debian the corner stop FOSS. H.264 is a "free" and open video codec, but patented (just like GIF). For the US, not even the oldest codec MPEG-1 (1992) is not patent free, and only more patent landmines have been laid. And even when Google purchases license from MPEG LA for everyone, it still doesn't the stop lawsuits. — Dispenser 23:03, 24 April 2013 (UTC)

Beyond the issues with proprietary formats, allowing "masqueraded" mime types seems very unwise. I have categorized those to be converted (some already have been), and filed bugzilla:47709 about those even making it to the site. Finally, GIF has never been free and open. Superm401 - Talk 08:06, 26 April 2013 (UTC)

Hasn't GIF been "free" since the patents expired in 2006? Anomie⚔ 21:27, 27 April 2013 (UTC)

Temporarily disabling WikiLove on English Wikipedia

Per Template:Bug and the discussion above, I'm going to temporarily disable the WikiLove extension until the underlying issues with mw.loader are resolved. Sorry for the inconvenience. If you need to spread WikiLove in the meantime, please reacquaint yourself with Category:WikiLove templates :) Kaldari (talk) 00:34, 26 April 2013 (UTC)

• Ha. Questionable timing, Kaldari--it's Friday, and who isn't in the mood for love on a Friday? I was just about to give it up to Wizardman. Drmies (talk) 17:43, 26 April 2013 (UTC)
  • Hopefully the WikiLove crisis will be over soon: https://gerrit.wikimedia.org/r/#/c/61057/ Kaldari (talk) 00:07, 27 April 2013 (UTC)
    • What do ya know. I haven't used WikiLove in ages. I was going to use it tonight, but I guess Wikipedia isn't feeling the love. MJ94 (talk) 01:38, 28 April 2013 (UTC)

Technical editors

Dear tech people:

Does Wikipedia have volunteer technical editors (for creating the scripts, tools, templates, etc.)? If so, where is the best place to inquire about becoming one of them? I have substantial, although somewhat outdated, experience in this area. —Anne Delong (talk) 14:45, 26 April 2013 (UTC)

What do you need? mabdul 16:10, 26 April 2013 (UTC)

I don't need anything in particular right now myself, I just enjoy coding, and wondered if this was a way I could help. If that's not allowed, just say so. — Preceding unsigned comment added by Anne Delong (talk • contribs) 19:47, 26 April 2013 (UTC)

Yes, there are ways for volunteer coders to contribute. The easiest (least sophisticated) way is through writing templates, which are just bits of specialized wikitext in the Template namespace. We also have Gadgets (written in Javascript), Modules (written in Lua), Skins (written with CSS). For more sophisticated programmers, one can also volunteer as a Mediawiki hacker (using PHP). With a bit more guidance about what you are looking for, perhaps we can point you in a more specific direction. Dragons flight (talk) 19:59, 26 April 2013 (UTC)

If you are able to edit the required page then you can just go ahead. Creating and editing templates, Lua modules or user scripts is like editing articles. There is no application process or official teams or leaders assigning tasks. PrimeHunter (talk) 23:17, 26 April 2013 (UTC)

Watchlisting this page would be a good start. People come here a lot to seek help from people such as yourself. Also, #mediawiki ^connect and #wikimedia-tech ^connect have lots of your types. :) Killiondude (talk) 01:08, 27 April 2013 (UTC)

Thank you. I would like to help out, but I know that errors can have more serious repercussions than when editing a regular text page. Are there standards and procedures essays that I should read before I make any changes? Or is everything usually located on the talk pages attached to the object? I am familiar with Javascript and PHP, but I would rather start out small. —Anne Delong (talk) 14:44, 27 April 2013 (UTC)

How can a page exist without any history

The page MediaWiki:Common.css/User:Paladox2014 exists (in that it's got an "edit this page" link not a "create this page" link), but it has no history. How can that happen? --Redrose64 (talk) 19:44, 27 April 2013 (UTC)

The same thing happens for all subpages of MediaWiki:Common.css. Assuming it's just a placeholder built into mediawiki. —Theopolisme (talk) 19:49, 27 April 2013 (UTC)

"MediaWiki:Common.css/User:Paladox2014" means the interface message "common.css" in the language with code "User:Paladox2014", just as "MediaWiki:Common.css/de" means the interface message "common.css" in German (code "de", from ISO 639). Of course, no language has code "User:Paladox2014", so it falls back to the default English message distributed with MediaWiki.

Since the interface message "common.css" does exist in MediaWiki, the software displays an edit link rather than a create page link. Anomie⚔ 21:43, 27 April 2013 (UTC)

As a little curio, non-existing languages cannot be chosen in preferences, but they can actually be chosen with uselang= in the url. For example, the page MediaWiki:Uploadtext/en-screenshot exists although "en-screenshot" is not intended as a language code. But if you replace http://en.wikipedia.org/wiki/Special:Upload with http://en.wikipedia.org/wiki/Special:Upload?uselang=en-screenshot then the software thinks you want to display Special:Upload in a language called en-screenshot, so you see MediaWiki:Uploadtext/en-screenshot. PrimeHunter (talk) 21:59, 27 April 2013 (UTC)

OK. Does anybody know why on earth Paladox2014 (talk · contribs) might need to create that page (see MediaWiki talk:Common.css/User:Paladox2014), or indeed what they're up to? The user account was created only yesterday, but from their edits so far (the very first was to create MediaWiki talk:Common.css/Wikipedia:Main Page/sandbox), I get the distinct impression that they want to jump right in with something hideously complicated - a redesign of the main page (without discussing it with others) is just one possibility. I might have formed entirely the wrong impression though. --Redrose64 (talk) 22:59, 27 April 2013 (UTC)

At the page you linked above, xe said "but how can I create for example Mediawiki:Common.css/User:Paladox2014/Main Page"...which makes it look like this was a failed attempt at that. WP:CLUE could be an issue. —Theopolisme (talk) 23:07, 27 April 2013 (UTC)

Removing the "Secure your account" information from the login screen

Hi. I've started a discussion at MediaWiki talk:Loginend#Future of this message about whether we want to continue including the "Secure your account" information on the login screen. Please discuss at the link provided. --MZMcBride (talk) 21:43, 27 April 2013 (UTC)

<Imagemap>

I recently created a clickable image for the infobox in Paris using Extension:ImageMap. However, this has added a couple of extra lines around the image, so is there a way for the infobox to look like it does in this revision but still to be clickable?--Gilderien Chat|List of good deeds 22:17, 27 April 2013 (UTC)

Done.[3] PrimeHunter (talk) 22:33, 27 April 2013 (UTC)

Thank you very much :) --Gilderien Chat|List of good deeds 22:38, 27 April 2013 (UTC)

Please say yes or no to this Afc proposal.

Dear editors: Last week I posted a proposal for an addition to the Afc submission process at my user page User:Anne Delong/AfcBox and asked the reviewers on the Afc talk page to respond at User talk:Anne Delong/AfcBox. After several of the reviewers showed interest, and with support from FoCuSandLeArN and some input from mabdul I asked for a technical assessment at Wikipedia:Village pump (proposals) and also at Village pump (technical).

Since then there has been a lot of discussion on all three talk pages about various ways to improve the Afc submission process, but aside from TheDJ, who indicated that my proposal was technically feasible, and Ypnypn , who agreed that PHP shouldn't be needed, all of the discussion has centred around alternative and more complicated ideas using bots, javascript, etc. These are likely good ideas, but don't provide feedback on my original simpler proposal.

Please will someone let me know if this simple proposal (rather than the other alternative ideas) is worth pursuing, or what's wrong with it if not, by posting your opinions at User talk:Anne Delong/AfcBox. If no one likes the idea, and people instead want to go in a different direction, I will delete it. If people agree that the proposal has merit. Petrb has agreed to set it up. I am posting this on all three talk pages hoping to get a decision one way or the other. Thanks for your time. —Anne Delong (talk) 22:56, 27 April 2013 (UTC)

Updating portal box images

I am looking to update the image displayed for the Portal:Coffee portal box (the box that's used in the See also sections of articles). I know that there's a page to update the image, but I just cannot find it at this time. I'm looking to update the generic portal puzzle image with File:A small cup of coffee.JPG. Northamerica1000^(talk) 03:17, 28 April 2013 (UTC)

See Template:Portal#Image Nanonic (talk) 04:20, 28 April 2013 (UTC)

 Resolved. Thanks very much Nanonic. Cheers, Northamerica1000^(talk) 05:07, 28 April 2013 (UTC)

Edit protect request to highly-visible template stalled?

Hi all! About a week ago, the Wikipedia:Route diagram template (RDT) project pushed out two edit requests to modify two highly-visible templates:

• Template talk:BSpx; and
• Template talk:BS-overlap.

However, it seems to me that these two requests were not addressed like other editrequests were, even after I wrote on WP:AN to ask to clear the backlog. May I know is there any special procedure to ask for large-scale modification to protected templates? Also can any administrator with the technical background to take a look on it? Thanks.

The background of this large-scale modification is at WT:RDT#Edit protected request 2013-04-17. — Peterwhy 16:42, 24 April 2013 (UTC)

Oh, I'm aware of it, just as I'm aware of all the other backlogged items at Category:Wikipedia protected edit requests, but I no longer touch anything that isn't 100% clear-cut cast-iron copper-bottomed uncontroversial; and in the case of templates (which these are) tested to destruction. This is because of the criticism that I have received when I fulfilled some requests that hadn't been discussed to death first. See also this discussion, if you have the patience. --Redrose64 (talk) 18:56, 24 April 2013 (UTC)

Hi Redrose, thank you for your response. I agree the original discussion has been quite limited to within WT:RDT, so I would like to know, what we should actually do to gain wider acceptance to this modification? And to eventually get the modification done? — Peterwhy 01:34, 25 April 2013 (UTC)

Admins unfamiliar with what a template does or how it works shouldn't carry out a proposed change without being certain that the change is beneficial, especially when the template is highly-visible. So, at the very least, I (and probably others) would like to see some WP:TESTCASES which both prove that these changes do what they're supposed to do, and also demonstrate that existing usage is not compromised.

I see that the proposed change to {{BSpx}} has been put into {{BSpx/sandbox}}, and that's great; but the proposed change to {{BS-overlap}} is not in {{BS-overlap/sandbox}}. Apparently it's in {{BS-overlap/sandbox2}} but that's more difficult to check since it doesn't have automatic links from the box at the bottom of Template:BS-overlap/doc nor from the {{editprotected}} box at Template talk:BS-overlap#Edit protected request 2013-04-19. The text after that states, rather vaguely, "we will have to modify some more unprotected templates afterwards" without detailing which ones nor in what way. I therefore cannot carry out any tests myself, which again means that testcases are needed.

As a frequent user of RDTs, I know what the overall suite is supposed to do, and for me there is one huge unanswered question: by changing the image size specifier from a height to a width, what will this do to those RDTs (such as this one) which use non-square icons? There are plenty of half-width icons such as   (exdKHSTla); there are also some double-width icons such as  (bvWSL-BS2+lr) (and even a few quarter-width icons like  (cSTRq)); these have had a constant height because {{BSpx}} contains x20px and not 20px - testcases will go a long way toward showing that these still work. --Redrose64 (talk) 06:37, 25 April 2013 (UTC)

Yes, I see I have to address your concerns before getting things changed.

• For location of sandbox of {{BS-overlap}}, that is in {{BS-overlap/sandbox2}}. The second sandbox was created by User:Sameboat before I started to work on this batch of modification, probably from a good intent to avoid erasing the previous content. If you know any way to merge the edit log of sandbox2 back to {{BS-overlap/sandbox}}, please do help to move that.
• For complete front to end test cases, please find them inside WT:RDT#Proposed changes to protected BS-overlap, BSpx. We have made some maps that uses {{BS-overlap/sandbox2}} and {{BSpx/sandbox}}. If you would like to create some tests, use the sandbox directory to find the sandboxes we used for testing.
• "we will have to modify some more unprotected templates afterwards": {{BSrow}}, {{BSrow-sc}}, {{BSrow-2}}, {{BSrow-replace}}, {{BSn}}, {{BSn-2}}, {{BSn-startCollapsible}}, {{BSn-replace}}. Again, these are the templates where sandboxes have been created to test for the new changes.
• Double/half/quarter/3-quarter width: The transition should be smooth, except probably between {{BSpx}} and {{BS-overlap}} are updated and other templates (see above) are updated. Technically, we only move the x of x20px to {{BS-overlap/sandbox2}}. The aim is listed at WT:RDT#About PX parameter and Template:BSpx.

Let me know if you have any other concerns. — Peterwhy 13:53, 25 April 2013 (UTC)

Additional note for the last point: we did not change "the image specifier from a height to a width", we change that from size to height. So this does not conflict with the original intention to add that x to {{BSpx}}. — Peterwhy 15:16, 25 April 2013 (UTC)

• Route maps 3x faster by {BS-overlap} update: Today, at 09:12, 28 April 2013‎, the route-map formatter Template:BS-overlap was updated (for the first time in 3 years) to use quick icon-overlay by {Template:Superimpose5} and omit {BS-alt} which was setting alt-text for each pictogram. Those changes have allowed many route-map diagrams to reformat about 3x faster, such as former 12-second maps now 4 sec. Unfortunately, the update also dropped the image-links (to each pictogram image-description page), which is likely improper handling of artwork-attribution links. Anyway, that update allows testing to confirm how the route diagrams can be edit-previewed 3x times faster now, and further updates can be discussed at WT:Route_diagram_template. Part of the problem had the fully-protected {BS-overlap}, and so there were several pent-up improvements which were delayed during the past 3 years. I have created a spike-solution variation to allow fuller testing of the complex features, and to format the pictogram alt-text much faster. However, the image-attribution links probably need to be re-added soon in Template:BS-overlap. -Wikid77 (talk) 10:25, 28 April 2013 (UTC)

Anyone willing and able to fix User:GregU/dashes.js?

User:GregU seems to have become inactive, and I've discovered a bug in this widely-used script (it has been incorporated in AutoEd, and is a js gadget).

The bug is pretty simple: it will modify the dashes and hyphens in section headers, and this breaks any wikilinks from redirects or other articles that link directly to the section. All it takes to fix it is to place {{anchor|original section header}} as the first line in the modified header whenever the tool modifies a section header. That way, people that worry about the appearance of dashes vs. hyphens stay satisfied and the incoming wikilinks and redirects stay working.

Anyone comfortable enough with Javascript editing of Wikipedia articles to take on the fix?—Kww(talk) 22:40, 26 April 2013 (UTC)

/me points to User:Writ Keeper. —Theopolisme (talk) 22:57, 26 April 2013 (UTC)

• How about first asking the right question? We need to find out whether anybody has witnessed the breaking of such a link, or potentially how many articles the alleged "problem" would affect? -- Ohconfucius ^{ping / poke} 01:41, 27 April 2013 (UTC)
• There are some other glitches that have already manifest as actual errors, as notified on the owner's talk page. -- Ohconfucius ^{ping / poke} 02:39, 27 April 2013 (UTC)
  • Storm in a teacup, if you ask me, Kww. Tony (talk) 10:22, 27 April 2013 (UTC)

• No one in the real world cares to force hyphens as dashes: Every so often we need people to come up for air, and learn that the world has stopped caring about the hyphen/dash issues. However, months ago someone had misled me into thinking the prestigious journal Nature had somehow advocated a style to use dashes between two-term expressions; however, after days of closer inspection, I discovered the truth is that Nature allowed the authors to use whichever dash/hyphen styles they preferred, even to mix hyphens and dashes in conflicting styles, such as page ranges "2-5" and "2–6" in the same peer-reviewed article. The 19th-century term "hyphenated Americans" states exactly what punctuation to use. Also, I learned, with the Michelson-Morley Experiment, even those two scientists spelled their Experiment with a hyphen (not a dash). Perhaps more confusing, the Adobe Reader will search with hyphen to match any dash or hyphen in the PDF-format text. Changing hyphens to dashes is just not as important as imagined some years ago, and has become viewed as balderdash. The world decided it did not care, and even hyphens have been removed from formerly hyphenated terms during the past 40 years. Spread the word down to the wiki-mines underground. -Wikid77 (talk) 22:17, 28 April 2013 (UTC)

Does anyone have a list of retired namespaces?

Due to the nature of this question, I believe that this is the correct forum for it. As an editor who did not become really involved in Wikipedia until 2012, but did look at Wikipedia from time to time from 2004 and onward, I thought I remembered seeing other name spaces on Wikipedia than the ones that exist today such as (article), Wikipedia:, Help:, MediaWiki:, Template:, etc. Essentially, I'm trying to build a proposal for a new namespace, but I want/need to make sure that what I'm trying to propose doesn't take the Wikipedia project backwards in time. So, does anyone have a list of any retired namespaces, or could someone either direct me to a page in Wikipedia where this list can be found? Steel1943 (talk) 23:57, 27 April 2013 (UTC)

Wikipedia:Namespace lists all of them, if I'm not mistaken. What exactly are you trying to propose? —Theopolisme (talk) 00:19, 28 April 2013 (UTC)

Thanks for that link Theopolisme, but that article only lists currently active namespaces; I'm looking for retired ones. And for your second question: basically, look at my user page for the answer to that one: I'm not trying to bring too much attention to that proposal since I have no idea yet if I would be taking steps backwards in the progress of this Wikipedia. Knowing about any retired name spaces could help with my latter concern. Steel1943 (talk) 00:42, 28 April 2013 (UTC)

An aside: I don't think that "stepping backwards" is always a bad thing. But in looking at your proposal and doing a bit of searching, I don't think that a specific Draft space has ever existed on Wikipedia. —Theopolisme (talk) 01:14, 28 April 2013 (UTC)

I am pretty confident that no namespaces have ever been "retired". — This, that and the other (talk) 01:53, 28 April 2013 (UTC)

indeed, no namespaces have ever been discontinued here. Graham87 04:49, 28 April 2013 (UTC)

We have switched the default names for certain namespaces ("image" has mostly been switched to "file"), but the same namespace still exists (and the old names\links still work). Could this be what you're thinking of? Andrew Gray (talk) 22:45, 28 April 2013 (UTC)

Score !!

Tracked in Phabricator
Task T2189
Resolved

Tracked in Phabricator
Task T49534

After many years, one of the most voted and oldest feature requests has been solved. As of today, Wikipedia finally has a renderer for music notation. See Mark's sandbox for an example. Congrats to the original filer xmlizer ! And a thank you to all who helped write the various generations of the extension and those that reviewed the code. —TheDJ (talk • contribs) 08:35, 23 April 2013 (UTC)

More at Help:Wiki markup#Musical notation. --Redrose64 (talk) 10:18, 23 April 2013 (UTC)

There is also a Wikibook in French which could be translated and added to English Wikibooks: b:fr:Introduction à LilyPond. Helder 13:13, 24 April 2013 (UTC)

Is this working? I tried adding the following code to Première rhapsodie:

<score raw=1>
\version "2.11.63"

\score {
  <<
    \new Staff="clar" \relative c'' {
      \clef treble
      \numericTimeSignature
      \time 4/4
      \key aes \major
      
      \mark \markup {\bold \small "Reveusement lent"} 
      R1 | r4 g\p_\markup {\italic \small "doux et expressif"}( bes c)
    }
    
    \new PianoStaff {
      <<
        \new Staff="one" \relative c'' {
          \clef treble
          \key ges \major 
          \numericTimeSignature
          \time 4/4
          
          r4\pp
          << { f2( f'4~) | f1 } \\
             { f,2.~ | f1 } \\
             { s2. | \times 2/3 { e'8\>( ees c~) } c2.\! } >>
        }
        
        \new Staff="two" \relative c' {
          \clef treble
          \key ges \major
          \numericTimeSignature
          \time 4/4
          
          << { f1--~ | f1 } \\
             { s1 | \times 2/3 { e8\>( ees c~) } c2.\! } >>
        }
      >>
    }
  >>
}
</score>

but it failed with the error

Processing `/tmp/MWLP.926c7a5f74a89391d70986be99418a50/file.ly'
Parsing...
Interpreting music... 
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `file.ps'...
Converting to PNG...GS exited with status: 9ERROR: In procedure delete-file:
ERROR: No such file or directory

Or is raw=1 not supported here? --SarekOfVulcan (talk) 16:16, 24 April 2013 (UTC)

raw is supported.—Emil J. 16:47, 24 April 2013 (UTC)

[[here says that raw would be used with lang="lilypond". But is estrange that no mark what error is here get an error after put a r8 and I don't know why. Is possible that have a bug in extension? --88.8.254.79 (talk) 16:11, 29 April 2013 (UTC)

lang="lilypond" is the default, it is not necessary to specify it explicitly. A simple test with raw=1 works fine:

<score raw="1">
\version "2.14.2"
\header { tagline = "" }
\score { \relative c' { c d e f } \layout { } \midi { } }
</score>

The error messages in both your and SOV’s examples suggest that the music was laid out all right, but the final conversion of the output from PostScript to PNG failed. This could indicate a problem with ghostscript rather than with lilypond proper.—Emil J. 16:31, 29 April 2013 (UTC)

FWIW, Ghostscript’s exit status 9 is invalid file access. There may be a problem with paths or file permissions. This should be reported to bugzilla, but I think it would be helpful if someone knowledgeable with lilypond syntax could make a simplified minimal example triggering the bug.—Emil J. 16:59, 29 April 2013 (UTC)

Hanging userpages

Tracked in Phabricator
Task T49457

I'm getting an odd effect when trying to access pages in userspace (doesn't affect any other namespace). The pages load, then appear to attempt to refresh, but do so unsuccessfully, leaving the browser hanging. Stoping the page load and hitting the Back button takes me to the correct page. I'm running Firefox 5.0.1. Anyone else spotted this, or know of a solution to it? Yunshui 雲‍水 08:01, 24 April 2013 (UTC)

Seems like the same issue as on Commons. It's reported, will take some time before it's actionable I think. —TheDJ (talk • contribs) 08:26, 24 April 2013 (UTC)

One fix on commons was to disable the wikilove option in prefs and another was right-click and open new window (not tab).--Canoe1967 (talk) 08:29, 24 April 2013 (UTC)

Wikimedia developers are working on this and trying to find the culprit, but haven't succeeded so far. See the bug report that was linked by TheDJ. --AKlapper (WMF) (talk) 10:12, 24 April 2013 (UTC)

While logged in with Firefox 20.0.1 it is not possible for me to open User:TheDJ and User:AKlapper (WMF) in new tabs on English Wikipedia. I can open User:Yunshui and User:Canoe1967 in new tabs. These are users commenting in this thread. I can open them all in the same tab without problems. (Note for techs working on this: I am using https and not http).

The Commons discussion is here: Commons:Village pump#April 21.

The problem is solved when i uncheck "Enable showing appreciation for other users with the WikiLove tab" in my preferences here: Special:Preferences#mw-prefsection-misc. --Timeshifter (talk) 13:12, 24 April 2013 (UTC)

• I believe this is the hanging resolvable by hitting the back button on my browser that I came here to report. I haven't noticed its being restricted to userspace pages, though. (Good that it's been reported, since I found it impossible to figure out what versions of my OS and browser I'm using - Win7 and Firefox something something). Yngvadottir (talk) 18:06, 24 April 2013 (UTC)

Cannot edit user talk pages - but fine elsewhere

At first I thought this was a browser issue, but seeing as my issue is limited to user talk pages only, it may not be...basically, I cannot edit user talk pages. I cannot even access them. As soon as I click on them I get half-a-second of normal display, before the screen goes blank. Mainspace? Fine. Other talk pages? Fine. User talk pages? No chance. Very odd. Help appreciated. GiantSnowman 13:08, 24 April 2013 (UTC)

See #Hanging userpages, above. If you have Wikilove enabled, try turning that off. -- John of Reading (talk) 13:21, 24 April 2013 (UTC)

Perfect, thanks. GiantSnowman 13:47, 24 April 2013 (UTC)

Unable to view user talk pages in firefox (problem with bits.wikimedia.org?)

Today I am unable to view most pages in the user talk namespace (other than my own talk page) - I get a brief flash of then page and then just a blank page while it apparently tries to "Read bits.wikimedia.org" according to the status bar, but the page never completes loading.

This is only happening in Firefox (20.0 on Xubuntu linux) and works fine in Konqueror and Links (the only other browsers I have access to) and is a new problem since last night when it was working fine (e.g. I was able to leave a message at User talk:CACook7).

I do run the NoScript extension but wikipedia.org and wikimedia.org are whitelisted, the script shows no blocked elements and temporarily disabling it does not make a difference. I am able to load the bits.wikimedia.org page (which redirects to www.wikimedia.org) without any issues. Thryduulf (talk) 15:16, 24 April 2013 (UTC)

Look up two or three sections for other reports of this. If you have Wikilove enabled, try disabling it. -- John of Reading (talk) 15:26, 24 April 2013 (UTC)

Thanks, disabling wikilove appears to have solved the problem. Thryduulf (talk) 15:30, 24 April 2013 (UTC)

#Hanging userpages -- You can be happy that you are an en.wp user. Obviously en.wp is more important than Commons. After AKlapper found the posts here, he changed the bugs relating to highest prio. This is the usual attitude towards Commons users. We get their test software first and if we find issues they are not important until they appear here. -- Rillke (talk) 15:32, 24 April 2013 (UTC)

I haven't been able to view lots of user talk pages today, but it is perfectly possible to edit the same pages (by manually adding "?action=edit" at the end of the URL). The wikilove icon is present on the edit page too, so why is the edit form still working? --Stefan2 (talk) 22:30, 24 April 2013 (UTC)

It's not just Firefox - I'm having the same problem with some user talk pages with IE9. Mine seems to be fine, as is User talk:Okeyes (WMF) and User talk:Rillke. However, I see the problem on User talk:Thryduulf and User talk:John of Reading. GoingBatty (talk) 23:55, 24 April 2013 (UTC)

Confirmed that disabling the WikiLove extension fixes the issue. GoingBatty (talk) 00:01, 25 April 2013 (UTC)

Worked for me too - thanks to all the above for identifying disabling WikiLove as a temp fix. -- stillnotelf is invisible 15:19, 25 April 2013 (UTC)

Thanks for the explanation. Is there any way to sign for a notification of when this problem is solved so that WikiLove can be reenabled? I find it very helpful in building a collegial, friendly atmosphere. --_{Piotr Konieczny aka Prokonsul Piotrus| reply here} 03:51, 25 April 2013 (UTC)

You could add your email address to the bug's cc list. That will send you an email every time there is a change to that bug, including a new comment, which is isn't quite what you are after. The only other thing I can think of is that I can try to remember to put a note on your talkpage when I see it's been fixed, but I can't promise to remember or to be online when it happens. Thryduulf (talk) 09:26, 25 April 2013 (UTC)

Fixed

We have deployed a change that should fix this problem, and have re-enabled WikiLove. If you see any more problems along these lines, please reply here, or at Template:Bug, or at #wikimedia-tech ^connect. Thanks. BJorsch (WMF) (talk) 23:22, 29 April 2013 (UTC)

User talk pages not loading with Firefox

Tracked in Phabricator
Task T49457

I can't get user talk pages to load today using Firefox. Other pages are fine, but user talk pages won't load as diffs or when I try to edit them. It's only happening with Firefox. SlimVirgin ^(talk) 21:21, 25 April 2013 (UTC)

Amusingly this is happening to me with user pages and some diffs. It has improved a bit during the day but not that much. Now it takes me a few tries. Snowolf ^{How can I help?} 21:23, 25 April 2013 (UTC)

There are a few threads about this up above from yesterday; try disabling the WikiLove tab (in Special:Preferences, on the "Misc." tab) and see if that helps. Writ Keeper ⚇♔ 21:24, 25 April 2013 (UTC)

I started seeing this too today. In Firebug Net tab, it says these URLs are loading endlessly:

http://commons.wikimedia.org/w/index.php?title=MediaWiki:Gadget-HotCat.js&action=raw&ctype=text/javascript

http://en.wikipedia.org/w/index.php?action=raw&ctype=text/css&title=MediaWiki:Gadget-navpop.css

http://en.wikipedia.org/w/index.php?action=raw&ctype=text/javascript&title=MediaWiki:Gadget-popups.js

If I interrupt loading quickly enough, it doesn't get stuck like that and I see the page normally. Really annoying, though. --Joy [shallot] (talk) 21:51, 25 April 2013 (UTC)

• I disabled WikiLove as Writ Keeper suggested, and the talk pages loaded again. I re-enabled it, and then they wouldn't load. So that seems to be the culprit. Thanks, WK. SlimVirgin ^(talk) 21:57, 25 April 2013 (UTC)

In the meantime I found bug 47457 (esp. comment #28). Ditto. --Joy [shallot] (talk) 22:15, 25 April 2013 (UTC)

This is handled in bugzilla:47457 and being worked on. Sorry for the inconvenience. Also see "Temporarily disabling WikiLove on English Wikipedia" below. --AKlapper (WMF) (talk) 10:03, 29 April 2013 (UTC)

Cat Scan

Is there a way i can find the articles of a category in english wikipedia that have an article in italian wikipedia? Xaris333 (talk) 11:47, 28 April 2013 (UTC)

Sounds like something you might want to ask the Wikidata folks. —Theopolisme (talk) 12:45, 28 April 2013 (UTC)

Done! Xaris333 (talk) 01:03, 30 April 2013 (UTC)

Wrong version of Edit tools

Resolved

Hi. Two or three days ago, the version of the Edit Tools that I get changed from the click-to-insert version depicted at MediaWiki:Edittools to the much more clunky copy-to-paste version. I haven't touched my preferences (and can't even see which preference would affect this). MSIE8 with active scripting enabled; Vector skin. Thanks for any help. --Stfg (talk) 13:21, 28 April 2013 (UTC)

Sounds like js in general might be failing somewhere - do other things still work, such as the search dropdown options, collapsible navigation, etc? -— Isarra ༆ 17:25, 28 April 2013 (UTC)

(edit conflict) what you describe is not what editTools supposed to do, and it's not what it does for me, so i think the problem is not general, but somehow specific to you. the first thing i'd do is try to see if it's related to your browser/computer or to your user account. the simplest way to test it is to log out, and try it again: edittools is on by default (too lazy to check, but i think it's a default gadget, rather than something added by Mediawiki:common.js ). if it has to do with your account, you probably want to go over your vector.js and common.js, and comment out one line at a time until you find the culprit, and discuss it with the owner of the imported script. if it happens when logged out also, please come back here and report - with as many details as you can (browser precise version, OS, which edittool is it that behaves this way, maybe even in which article did it happen, and any other detail you can think of). peace - קיפודנחש (aka kipod) (talk) 17:27, 28 April 2013 (UTC)

Search dropdown and collapsible navboxes work fine. Problem goes away when I log out. I've completely cleared my vector.js, common.js and, for good measure, vecrtor.css also (and remembered to clear the cache ) None of this removes the problem when logged in. By the way, I get the desired click-to-insert edit tool when editing these three files -- but only these three; editing anything else -- mainspace article, page in my user space, template -- still has the problem. I haven't done anything to my account for ages. I have Win XP (home) with SP3, fully patched. Browser version is MSIE 8.0.6001.18702. Thanks again for looking at this. --Stfg (talk) 19:26, 28 April 2013 (UTC)

I've bypassed the problem by removing one of the scripts from my vector.js and have started a discussion with the designer. Please don't invest any more of your time in it, and thanks for the help so far. --Stfg (talk) 16:36, 29 April 2013 (UTC)

Manipulating categories

Resolved

Is there a way (a Special page similar to Special:ComparePages, perhaps) to see an automatically generated list of all the members of e.g. $\left(CategoryA\lor CategoryB\right)$, $\left[\left(CategoryA\land CategoryB\right)\lor \left(CategoryC\land \neg CategoryD\right)\right]$, etc.? It Is Me Here ^{t / c} 15:03, 28 April 2013 (UTC)

Yep, Wikipedia:CatScan seems like exactly what you're looking for. —Theopolisme (talk) 15:08, 28 April 2013 (UTC)

Thank you It Is Me Here ^{t / c} 09:27, 29 April 2013 (UTC)

Parser function question

I notice that the #if: function would trim line breaks from parameters 2 and 3, as shown above with User:Peterwhy/sandbox. I would like to know if this is an intended effect? Thanks. — Peterwhy 18:37, 28 April 2013 (UTC)

It's a known and long-established feature of the MediaWiki template parser that leading and trailing whitespace is significant around positional parameters, but not around named parameters. It's unlikely to change. Parser functions such as {{#if:}} strip surrounding whitespace from their output. --Redrose64 (talk) 20:04, 28 April 2013 (UTC)

• Use the space-sensitive if-templates: There are some special templates which allow a space-sensitive result, with newline linebreaks, in conditional expressions. For example, using Template:ifeq or Template:ifexpr:

• "{{ifeq|{{{it|xx}}}|xx| It is xx. }}" → " It is xx. "
• "{{ifexpr|{{{| {{{age|19}}} >= 18}}} | Over 18 }}" → "Template:Ifexpr"

Those templates are extremely fast, so could be used hundreds of times with no significant delay. When the data contains an equals sign ("="), then pass the data inside a null-named parameter {{{|____}}}. -Wikid77 (talk) 23:05, 28 April 2013 (UTC)

Wikid77, please stop confusing the issue. There is no need for such complicated techniques; moreover, Peterwhy does not want to preserve the whitespace, because that breaks the image linking. What is needed is to trim the whitespace, and the construct that Peterwhy has already used (see the example that he began this thread with, his sandbox also this edit), i.e. {{#if:1|{{{paramname|}}} }} works perfectly well, and since it uses a single parser function, which is built in to the MediaWiki software, is bound to be faster than any template-based alternative. --Redrose64 (talk) 08:45, 29 April 2013 (UTC)

Problem editing a table

Today on List of vegans, when I added a name to the sortable table and previewed, it wasn't formatted as a table, but as lines of text:

|- !scope="row" | John Salley |Retired professional basketball player, actor and talk show host |United States |[1][2][3] - !scope="row" | Chad Ackerman |Singer, songwriter |United States |[4]

But when I saved, it was properly formatted. This problem only started today. It's happening repeatedly, with more than one browser, and with an account with different preferences. Can anyone think what the problem might be? SlimVirgin ^(talk) 21:20, 28 April 2013 (UTC)

Let me guess... it began happening after you had saved this edit. What you did there was to split the table into editable sections; as a result, when you edit any of the sections that occur after the start of the table, the table start code won't be present when you preview, but it will be there once saved. --Redrose64 (talk) 21:53, 28 April 2013 (UTC)

I noticed it today at 17:42, 28 April 2013‎; I think it was okay after the sections were first added, but I'm not sure. We need the subsections because loading the whole table is very slow and hard to edit with; it also leads to edit conflicts. But we also need the code to be present on preview, because people will think they've messed it up and won't save. So it's not just a nuisance factor, it will stop people from editing. What can be done, do you think? SlimVirgin ^(talk) 22:50, 28 April 2013 (UTC)

• Split into subarticles by alpha group: The only way to edit-preview part of a sortable table, with table headers, is to create transcluded sub-articles, where each page shows a part of the total list, but the overall list shows all subarticles transcluded together. For example, with letters A-F, create subarticle "List of vegans A-F" as if being a template, to allow edit-preview with table headers. Each subarticle page would be structured in the following manner:

         <noinclude>
         This is part of the '''[[List of vegans]], names A-F''' as a shorter
         section of the whole list.
         {| class="wikitable sortable plainrowheaders"
         |-
         ! scope="col" style="width:8em;"|Name
         ! scope="col" style="width:35em;"|Occupation
         ! scope="col" style="width:6em;"|Birthplace
         ! scope="col" class="unsortable" style="width:5em;"|Source
         </noinclude>
         |-
         person 1
         |- 
         person 2
         <noinclude>
         |}
         ==References==
         {{Reflist}}
         [[Category:...]]
         </noinclude>

Then, once separated into subarticles as the partial lists, each would be transcluded into the full "List of vegans" by treating each subarticle as if being a template file, trancluded by: Template:J in doubled curly braces. As the list grows, perhaps each subarticle would be sub-divided more, to span just 4 letters of the alphabet, rather than 6 letters A-F. So, plan now for how many letters in each subarticle. Note that each subarticle is truly a WP article (not a template), with lede paragraph and References, plus categories. However, that gives readers the option to view each subarticle if looking for vegans with names under just those letters. -Wikid77 (talk) 23:47, 28 April 2013 (UTC)

Thanks, Wikid. That sounds a bit complicated, and would it be confusing for new editors? I'm wondering whether it makes more sense just to have regular subsections (A-F, etc); that is, multiple tables. SlimVirgin ^(talk) 01:19, 29 April 2013 (UTC)

As a screen reader user, I'd prefer to just have multiple tables. The table does work with my screen reader JAWS, but it interprets the headings as part of the table header, meaning that I hear "A–F" along with the edit link when I navigate horizontally through the table; not a showstopper, but quite annoying. A big long table without headings shouldn't lead to edit conflicts, though, unless two users are trying to add people at the same point in the alphabet. Graham87 07:11, 29 April 2013 (UTC)

Personally I'd go with separate tables for each group, as we have done on pages like List of closed railway stations in Britain: D-F. But to return to the original matter; there is a simple technique, but it must be done carefully. When editing a section within the big long table, make your edit, and then go to the top of the edit window and insert one line immediately after the section header:

{|

and then go for Show preview Having verified the intended edit, make sure that you remove that {| before clicking Save page, because if you don't, you'll create a table-within-a-table, and almost certainly screw up the page formatting. --Redrose64 (talk) 09:10, 29 April 2013 (UTC)

In general, adding more complexity in order to create a solution to a complex problem is often not the best idea. Go with separate lists. —TheDJ (talk • contribs) 10:01, 29 April 2013 (UTC)

Many thanks to everyone for the input. Separate tables/sections seems to be the simplest solution. Thanks again. SlimVirgin ^(talk) 17:41, 29 April 2013 (UTC)

Pictures from Maps+ iPad app

Just a quick question: I have a couple of pictures taken from the Maps+ app on my iPad mini. Is it against any regulations to use them as pictures, and if it was permissible, under what licenses would they be put under. Also, I use a Safari browser on my iPad mini. --JB82 (talk) 23:02, 28 April 2013 (UTC)

That sounds like they would be copyrighted violations in most cases. The design of the app is owned by IZE, Ltd and the map imagery is owned by Google and it's map providers. There is some information on this topic in our WP:SCREENSHOT page. —TheDJ (talk • contribs) 07:39, 29 April 2013 (UTC)

Body font size

I am a beginner in CSS. Sorry if it sounds dumb to ask or if this is posted on a wrong page. I want to know how I can increase the body font size so that browsing on ipad could be a little easier. I tried messing with the CSS code but ended up with larger texts in all he toolboxes but not the article itself. — Preceding unsigned comment added by Lauivan (talk • contribs) 09:55, 29 April 2013 (UTC)

Open all articles of a category

Is there a way to open all articles of a catecory (not by click one by one)? With firefox for example? Xaris333 (talk) 18:22, 29 April 2013 (UTC)

WP:LINKY should do what you want in Firefox. jcgoble3 (talk) 19:29, 29 April 2013 (UTC)

Thx. Very useful. Xaris333 (talk) 22:21, 29 April 2013 (UTC)

Caterogies

Is there a way i can find all tha categories i have created? Xaris333 (talk) 19:09, 29 April 2013 (UTC)

Follow this link and hope the toolserver is in good mood... It seems the only hit is Category:Nea Salamina Famagusta managers. — HHHIPPO 19:22, 29 April 2013 (UTC)

Thx. Very useful. Xaris333 (talk) 22:21, 29 April 2013 (UTC)

Toolserver tools dysfunctional?

It appears that something strange has happened with the Toolserver. I was reviewing a AfC submission (Wikipedia talk:Articles for creation/Marc Danzeisen) and clicked the link (in the AFC template) to do WP:REFLINKS. To my astonishment I got a series of errors, eventually ending up at toolserver saying that the page doesn't exisist as a wiki. To verify I'm not going crazy I also checked Citation bot to see if something was up. It's been 10 minutes and Citation Bot has yet to finish. I do not think this is a web browser problem based on the fact that these links worked not 5 minutes before this oddity. Status Dashboard shows no issues. Hasteur (talk) 19:10, 29 April 2013 (UTC)

You're lucky that you got a message at all... I often get a blank screen, that's if it doesn't hang or timeout. If you look at past threads on this page, you'll realise that Toolserver has been having problems for months now. --Redrose64 (talk) 20:12, 29 April 2013 (UTC)

Copyright violation box

Dear tech people:

When I am reviewing articles at the Afc, I sometimes come across copyright violations. When I decline a submission for this reason, the dialogue box asks for the URL that has been copied. There's a box to type or paste it in, but it's very small. It's hard to tell if the URL has been entered correctly. Could the box be made wider so that more of the URL would show on the screen? —Anne Delong (talk) 00:30, 21 April 2013 (UTC)

I guess this is about MediaWiki:Gadget-afchelper.js. You can post to Wikipedia:WikiProject Articles for creation/Helper script/Development page#Feedback. PrimeHunter (talk) 00:39, 21 April 2013 (UTC)

With URLs I usually right click and open in new tab from the edit preview to see if the link goes to the correct site page.--Canoe1967 (talk) 01:24, 21 April 2013 (UTC)

I am not sure what you mean. Do you first decline the copyright infringement, and then right click on the URL in the resulting notice? If so, I would prefer to catch the error at an earlier stage. —Anne Delong (talk) 12:30, 21 April 2013 (UTC)

I have left a message as suggested. —Anne Delong (talk) 12:39, 21 April 2013 (UTC)

I have never seen the decline/accept edit screen. I was referring to when I do a regular edit and want to check wikilinks and urls. They can be right-clicked from the preview of the edit section.--Canoe1967 (talk) 13:39, 23 April 2013 (UTC)

Unfortunately, when working in the Afc script there is no preview. Back to my original question, can the entry box for the copyright violation URL be made longer? There's plenty of space in the pop-up box. If no one here knows, can you direct me where to ask? —Anne Delong (talk) 14:34, 27 April 2013 (UTC)

MY WISH HAS BEEN GRANTED! Thanks to whomever took care of it. —Anne Delong (talk) 03:58, 30 April 2013 (UTC)

That was mabdul in response to [4] at Wikipedia:WikiProject Articles for creation/Helper script/Development page. PrimeHunter (talk) 10:22, 30 April 2013 (UTC)

Test the new login and account creation page designs

Hi all,

After many weeks of testing, We (the editor engagement experiments team) are is getting close to enabling redesigns of the account creation and login pages. (There's more background about how we got here and why ‎our blog post.)

Right now are trying to identify any final bugs before we enable new defaults. This is where we really need your help: for now, we don't want to disrupt these critical functions if there are outstanding bugs or mistranslated interface messages. So for about a week, the new designs are opt-in only for testing purposes, and it would be wonderful if you could give them a try. Here's how:

• Go to account creation and add &useNew=1 to the URL, or just use this link
• Go to login and add &useNew=1 to the URL, or just use this link

If you have questions about how to test this or why something might be the way it is, I'd definitely check out our step-by-step testing guide and the general documentation.

Many thanks, Steven Walling (WMF) • talk 22:39, 25 April 2013 (UTC)

Will Special:PasswordReset be updated to reflect this new design as well? —Theopolisme (talk) 01:38, 26 April 2013 (UTC)

No, though if you think it should be, it should be simple enough. Steven Walling (WMF) • talk 04:59, 26 April 2013 (UTC)

I've taken the liberty of making the test links protocol-relative (i.e. they go to the HTTPS site if the user is on the secure server, otherwise they go to the HTTP site). Graham87 06:29, 26 April 2013 (UTC)

Thanks Graham. One of the nicer parts of the new login form IMO is that it includes a default link to the HTTPS connection, if you're not already on it. Steven Walling (WMF) • talk 06:32, 26 April 2013 (UTC)

The gray text used for the form field labels (e.g. "Username", "Password") is too light: it is difficult to read, and moreover, it is not compliant with WCAG accessibility guidelines. Please consider darkening the text colour a bit. — This, that and the other (talk) 06:41, 26 April 2013 (UTC)

Agree. How many times has this happened before ? Don't the designers have this on their checklist by now ? —TheDJ (talk • contribs) 10:29, 26 April 2013 (UTC)

The lighter color is very much intentional, and in several rounds of usability tests (here's a few who gave us permission to share), we've seen pretty much all users be able to read fields contents with this color. It's explanatory content, and is thus lower priority than the actual field titles. Making all text the same style reduces usability for the majority of users. Steven Walling (WMF) • talk 20:43, 26 April 2013 (UTC)

I don't object to using gray text, I think it looks good, but aren't you bothered by the rather low contrast level in terms of accessibility? It is absolutely crucial that our login form is as accessible to as many people as possible. — This, that and the other (talk) 01:02, 27 April 2013 (UTC)

And how many people in that group of testers had a severe form of colorblindness ? Just because nobody in the testgroup complains about the lack of a level floor in a new building doesn't mean much until you bring someone with a wheelchair. I don't blame people for NOT taking this into account when testing. I didn't even notice myself until TtatO brought it up and usually i'm one of the first to complain about these issues. There are WCAG checklists for a reason, namely the fact that many of those points are hard to notice for people that don't have disabilities. That's why you build a checklist. Not to cross off every point like a machine, but to make your mind actively consider the step (Ask airline pilots). —TheDJ (talk • contribs) 08:06, 27 April 2013 (UTC)

Actually, one of my team members is color blind. I agree we could consider readability enhancements, which are filed as bug 47777 if you'd like to add 2 cents there. Steven Walling (WMF) • talk 19:11, 30 April 2013 (UTC)

In general, like. I thought you guys were adding inline validation as well btw ? Or is that for a later phase ? If you take away the security advise etc from the interface, and don't give people tooltip advise, I'm sure some ppl will start wanting the headers with advice back. Also, slightly related: How about changing the "email (optional)" into "email (advisable/recommended)" ? —TheDJ (talk • contribs) 10:26, 26 April 2013 (UTC)

Also, AFAICS, there is no opportunity to add extra footer text: see [5] and [6]. Obviously it looks a lot better without footer text, but sometimes there may be a need for the community to add various timely messages to this part of the screen. — This, that and the other (talk) 10:41, 26 April 2013 (UTC)

I agree with "Email (recommended)" over (optional) as that would encourage people to provide their email address and they'd be able to get all the future benefits of Echo, Flow etc. Thehelpfulone 16:20, 26 April 2013 (UTC)

TheDJ: yeah, we deferred inline validation until another release. We'll either refactor using the validation HTMLForm already provides, or rebuild the backend of the form on top of the API (which supports CAPTCHA delivery now, thanks to Brion). Regarding the email label: remember that most users don't expect email to be optional, it's actually kind of unusual compared to the other popular sites on the Web. If we want to recommend email strongly, we could consider doing something like adding a suggestion to provide it, if people skip the field before submitting (this would require client-side validation, of course). An interim compromise could be "(Optional, but recommended)" which is clear but slightly wordy. Thoughts? Steven Walling (WMF) • talk 21:00, 26 April 2013 (UTC)

Error trying to restore an image

I had deleted Image:Gracyn Shinyei 2013.jpg earlier today as lacking permission. The uploader informed me that he had sent a letter of permission to OTRS and sent me a copy of it, which looked sufficient, so I went to restore the image and tag it as OTRS pending. However, when I did it, I got this error:

Error undeleting page

---

Then, I realized that I could view the deleted image, save it, and try to re-upload it myself, but I got this error:

Any thoughts other than move it to a different filename?

--B (talk) 21:26, 29 April 2013 (UTC)

Thanks for taking the time to report this! The problem you are reporting sounds like a potential issue in the code of the servers. If the problem is reproducible, it would be nice if somebody who has this issue could send the software bug to the 'Bugzilla' bug tracker under the product "Wikimedia" and component "File management" by following the instructions How to report a bug. This is to make developers of the software aware of the issue. If you have done so, please paste the number of the bug report (or the link) here, so others can also inform themselves about the bug's status. Thanks in advance! --AKlapper (WMF) (talk) 11:41, 30 April 2013 (UTC)

People were having the exact same problem with certain files about a year ago or so. If I remember correctly, the problem was that some files had wrong file access permissions on the servers where they were stored, with the result that the files couldn't be deleted, undeleted or moved. There were discussions about this on Commons at that time, and people were troubled about not being able to delete certain copyright violations. --Stefan2 (talk) 11:48, 30 April 2013 (UTC)

Need help removing the background sound from an audio

Hello,

Can anyone here please remove the background noises from this video please - [7]. I really require it, as I'm trying to get it translated into its article, and there is too much noise to understand it clearly.

Thanks, TheOriginalSoni (talk) 01:25, 30 April 2013 (UTC)

It is a tough problem, for sure. I combined the left and right channels into mono, which helped lower the noise floor somewhat, and greatly increased the signal level, but I was unable to pull the voices out of the static. The big problem is that the noise is in the same frequency range as the voices. Anyway, the resulting file I have is uncompressed WAV audio with a size of 16.4 MB. Let me know if you want to hear it. Binksternet (talk) 02:58, 30 April 2013 (UTC)

Of course I want to hear it :) Do upload it and link it to me. [Though for upload purposes compressing into mp3 will be better]

Also, the persion doing the tranlation says its most unclear near the beginning and end. So if you can do something about it, that will be good. TheOriginalSoni (talk) 04:10, 30 April 2013 (UTC)

Thanks for volunteering to help! (I'm the user doing the translation.) Actually, it's the first ~60 seconds that are giving me the most trouble because they are so quiet. The part at the end became more intelligible after I listened to it a few times. Tempodivalse [talk] 04:38, 30 April 2013 (UTC)

I share audio files all the time with people who know my real name, but I am not so familiar with using an anonymous account. Please see if you are able to get the audio from this shared folder: https://sites.google.com/site/audiobink/. It is a mono mp3 file which is slightly smaller than 4 MB. Best of luck! Binksternet (talk) 13:42, 30 April 2013 (UTC)

Binksternet, I can't see it, unfortunately. All I get is a blank page. Perhaps you could try another file sharing service? Tempodivalse [talk] 16:48, 30 April 2013 (UTC)

Or email? (Once you've exchanged emails, so you can email directly, not via the WP interface.) Rd232 ^talk 17:28, 30 April 2013 (UTC)

One more try before I start emailing people... Try this link. Binksternet (talk) 21:47, 30 April 2013 (UTC)

Link works find for me. Lets hope its clear enough for Tempodivalse. TheOriginalSoni (talk) 22:50, 30 April 2013 (UTC)

Download worked. The noise floor is more manageable now, although it will still be a challenge to decipher the individual words. I think I can catch most of the conversation if I listen long enough. Tempodivalse [talk] 03:41, 1 May 2013 (UTC)

Lines not wrapping in edit box

I am using IE10 on my laptop with Windows 7. The text in my edit window does not wrap, which is most inconvenient. Is there a setting somewhere that will correct this? • • • Peter (Southwood) ^(talk): 07:00, 30 April 2013 (UTC)

Replication lag

Anyone know what it's so high (14 hours)? Beyond My Ken (talk) 09:22, 30 April 2013 (UTC)

Toolserver has been having problems for months. --Redrose64 (talk) 09:36, 30 April 2013 (UTC)

• Cluster s1 had 0h replag, but s2/s5 have 14h: The enwiki wp:REPLAG has been showing zero hours for the s1 cluster, for many days during the past 2 weeks. I am not sure which wikis are on s2 and s5, with the 14-hour replag. However, I have advised the enwiki Lua-cite editors (for wp:CS1 cites) to avoid changing Lua-cite Module:Citation/CS1, as reduced to perhaps once per week or once per 2 weeks, to stop reformatting the 2 million CS1-cite pages "everyday" as before. Meanwhile, the category links still tend to be slow to reset within enwiki itself, even after changing a template used in just a few articles, and must null-edit those few dependent articles to force category-link updates. Also, many of the other-language wikipedias have caught the epidemic navboxaholism, and so a change to their main navbox or infobox templates might cause zillions of cache pages to become out-of-sync and increase the replag. -Wikid77 11:05, 30 April 2013 (UTC)

Can "×" be made interchangeable with "x" for searching?

Tracked in Phabricator
Task T49881

A new user pointed out that "This question has to do with all plants on wikipedia that are hybrids, and therefore can contain a special symbol called a cross (×) as part of their binomial name. While it's correct to have the special character '×' in the binomial name of a hybrid, it's neither consistent across all plants on Wikipedia, nor is it terribly useful for users trying to find a certain plant when Wikipedia gives a 404 if I put an 'x' instead of an '×'."

His example was Rubus × loganobaccus with this special character (×) (http://en.wikipedia.org/wiki/Rubus_%C3%97_loganobaccus), which did not work with 'x' (http://en.wikipedia.org/wiki/Rubus_x_loganobaccus).

This has since been handled by redirect, but it is not so easy to redirect them all, because (surprise surprise!) searching for × in the standard form gets you "An error has occurred while searching: The search backend returned an error:" (without giving an error). It can actually go to × if you type it in the go/search box and hit enter, though.

Is it possible to fix the search so that either:

a) any search with × that delivers either an error or no results is reattempted with x, and any search or URL with "x" delivering no results is reattempted with × (hmmm, that is a lot)

or at least

b) A search with × returns all article names containing it?

Wnt (talk) 14:30, 30 April 2013 (UTC)

I've filed a request on bugzilla as Template:Bug. Andrew Gray (talk) 14:55, 30 April 2013 (UTC)

Echo

When is Echo coming around? -- Ypnypn (talk) 17:22, 30 April 2013 (UTC)

Deployment is between 11-1pm today Pacific time, and thus is should begin appearing after that assuming no hiccups. Steven Walling (WMF) • talk 17:49, 30 April 2013 (UTC)

Notifications, or Echo, has now been released! You should start seeing things now: let us know on the talkpage if you see any bugs or have any feedback :) Thanks, Okeyes (WMF) (talk) 20:04, 30 April 2013 (UTC)

Bot

If someone have patient and time to help me creating a bot, pls answer me in my talk page. Xaris333 (talk) 19:34, 30 April 2013 (UTC)

Suggestion

Gadget suggestion, suppress the annoying "Your edit has been saved" bubble on every page save. Werieth (talk) 20:53, 30 April 2013 (UTC)

if you dig in WP:VPT archives, i think there's a CSS one-liner that will do that for you. peace - קיפודנחש (aka kipod) (talk) 21:03, 30 April 2013 (UTC)

This should do it:

.postedit { display: none; }

There isn't a gadget or preference to hide this because it's a system message, just like the notification bubble for watching/unwatching something. Steven Walling (WMF) • talk 21:14, 30 April 2013 (UTC)

I've corrected your CSS to include the braces, which I believe are required. – PartTimeGnome (talk | contribs) 22:29, 30 April 2013 (UTC)

They are, I was just in a hurry. Thanks for the fix. Steven Walling (WMF) • talk 22:31, 30 April 2013 (UTC)

CAPTCHA and Safari

Thrice over the past few weeks I've had new users in the IRC help channel say that they couldn't get past the CAPTCHA at the Article Wizard while using Safari; entering one CAPTCHA apparently simply brings up the next in an endless loop. I don't have Safari myself and thus can't test the effect myself, not even verify it exists, but someone else may want to take a look. Huon (talk) 22:25, 30 April 2013 (UTC)

I can help test this, but do you mean unregistered or registered folks? And by CAPTCHA I assume you mean a CAPTCHA on page creation? (It would help to know these things for someone to replicate.) Steven Walling (WMF) • talk 22:29, 30 April 2013 (UTC)

The way I understand it, registered but not yet autoconfirmed users writing their first draft via the Article Wizard. Today's example was Special:Contributions/LKK343 who apparently was successful after I advised them to use some other browser; they said they'd use Chrome. I've invited LKK343 to comment here and correct me if I got it wrong. Huon (talk) 22:45, 30 April 2013 (UTC)

Curious. I had a similar problem when (several years ago) I tried to register: one captcha lead to another. I think I was using Firefox at the time. Don't recall how I eventually got past it. ~ J. Johnson (JJ) (talk) 02:20, 1 May 2013 )UTC)

TorBlock problems

Tracked in Phabricator
Task T49626

It's been the third time this week I've received an unblock request on WP:UTRS from an established user who was running a Tor node which was not an exit node, but merely a relay node, and it was blocked by mw:Extension:TorBlock. This appears to be a systematic problem. I've filed a bugzilla request, but it hasn't yet been addressed. What should we do in the meantime? -- King of ♥ ♦ ♣ ♠ 02:53, 28 April 2013 (UTC)

Thanks for filing the bug report! It was filed four days ago and there was a weekend in-between, so I'd be a bit more patient. Andrew Garrett is CC'ed on the bug report, and as far as I know he'd be a good person to contact about it. --AKlapper (WMF) (talk) 10:06, 29 April 2013 (UTC)

OK, thank you. -- King of ♥ ♦ ♣ ♠ 07:37, 1 May 2013 (UTC)

Range viewing with IPv6

I use the CIDR range notation tool sometimes when viewing IP editing history (so I can find /24 addresses, for instance)... is there a similar tool for IPv6 addresses? Or is there a way to make it work with IPv6? Shadowjams (talk) 14:37, 28 April 2013 (UTC)

If you look at the gadget source you can see that only /27 to /32 actually do a CIDR range lookups (the ucuser parameter is limited to 50 simultaneous values, so 2^(32-27) = 32 is the most that can be done at once). /24 and /16 are special cases that do a ucuserprefix search instead. These behave functionally the same as an asterisk suffixed; Eg: 1.2.3.4/16 == 1.2.*. For some reason, the mw:API can't do ucuserprefix on IPv6 addresses. Compare: Splark 92.12 2001:9d8:2005 (2001:9d8:2005:12::3 has edits). If it did, then simply using 2001:9d8:2005* should work. Possibly someone should open a WP:BUG (Product: MediaWiki. Component: API). --Splarka (rant) 21:47, 1 May 2013 (UTC)

It works if you use correct capitalization: 2001:9D8:2005 Anomie⚔ 00:24, 2 May 2013 (UTC)

Interesting. Requiring uppercase seems to violate RFC 5952 and "The hexadecimal digits are case-insensitive, but IETF recommendations suggest the use of lower case letters." It even contradicts the format of the capitalization of IPv6 addresses in Recent Changes and article histories. Though it does agree with the API internal format (and IPv6 users' talk page titles). The API could possibly be modified to detect IPv6 fragments and ignore the case in a prefix search, as could the gadget, but neither solution is very elegant. --Splarka (rant) 01:17, 2 May 2013 (UTC)

The problem with that is how does it decide something is an "IPv6 fragment" rather than a username that happens to consist of numbers and letters a-f? Anomie⚔ 01:34, 2 May 2013 (UTC)

Unwanted global account(s)

I contribute to eight language-specific Wikipedias (CY, DE, EN, ES, FR, GA, NL, SV), with accounts at each, mostly under different user names in each case.

Lately, after logging in to (at least) my CY, NL, and SV accounts I am also being logged in (without being asked, as far as I can see) into

• meta.wikimedia.org
• wiktionary.org
• wikibooks.org
• wikiquote.org
• wikisource.org
• commons.wikimedia.org
• wikinews.org
• wikiversity.org
• mediawiki.org
• wikidata.org
• species.wikimedia.org
• incubator.wikimedia.org
• wikivoyage.org

That might be harmless enough, I suppose, except that now whenever I switch from cy.wikipedia, nl.wikipedia, or sv.wikipedia to one of my other accounts I find myself logged in there under the name I used at the wikipedia I have just been visiting, and have to log out and in again under my "real" username for that language -- which is, to put it at its mildest, bloody annoying!

Can anyone tell me, please, if this situation is reversible. I suspect there may be some connection between what is happening here and the "global account" concept (which, as I say, I have never knowingly signed up to). -- Picapica (talk) 11:27, 29 April 2013 (UTC)

This is one of the "benefits" brought by WP:SUL. My login was created after its introduction, so I'm used to the idea that I only need to log in once and I'm automatically logged in to all the others (although it does have its hiccups). Before we had Wikidata, it was a great boon when fixing interlanguage links (try some of the links at User:Redrose64#Non-English editing). But although I do very little on other Wikipedias now, it's still a benefit for editing on Commons, Meta and Wikidata. --Redrose64 (talk) 12:37, 29 April 2013 (UTC)

The only way to reverse this "problem" is to use one account. Whether that means you ask local bureaucrats to rename the account where it is not a part of the SUL, or abandon the old accounts, is up to you. I would personally just abandon the old accounts and ask for a bureaucrat or bureaucrat-analog to move the rights from old accounts to whichever you standardize on. --Izno (talk) 12:43, 29 April 2013 (UTC)

Also, this feature has been in place for a while. It surprises me that you are just discovering it now. --Izno (talk) 12:45, 29 April 2013 (UTC)

For "a while" read "almost five years". May 27, 2008 is almost a year before my account was created (5 May 2009) but four years after yours. --Redrose64 (talk) 12:59, 29 April 2013 (UTC)

Try the links on WP:SUL you may be able to unify them yourself. - X201 (talk) 13:01, 29 April 2013 (UTC)

Thanks for the answers, folks. The thing is, though: I don't want to unify my accounts; I want to UNunify them!

Izno, this problem has (re)surfaced only today. Checking back, I see that I first got rid of it around four and a half years ago (see paragraphs below). I have a strong suspicion that its re-appearance could be connected to my use (for the first time) today of the relatively new "Edit links" feature for updating interwikis, which I did from the Dutch-language nlwiki.

Here is a copy of a previous exchange on this subject from 2008. (I hope the solution offered there will work again!):

I very much regret falling for the invitation to go for a unified login. I have Wikipedia accounts in eight languages, mostly with different usernames. Now when I switch between different language Wps I frequently have to log out and log in each time in order to use these accounts. I don't wish to rename these accounts even if I could (new accounts using my main, English-language name have been set up without my asking for them).

Is there any way I can undo the unified login process so that I can be returned to the happy state I was in before I ever heard of unified login? If not, let this be a reminder to others of the value of the advice "If it ain't broke, don't fix it". :{ -- Picapica 11:24, 5 September 2008 (UTC)

Yes, You can, please request it here, also make sure to have a valid, confirmed email in Your special:preferences for Your own safety, thanks, --birdy geimfyglið (:> )=| 11:26, 5 September 2008 (UTC)

The global account for Picapica was deleted in 2008 as you say [8] after a request at meta:Steward requests/SUL requests. It hasn't been recreated so I guess you refer to one or more other usernames at CY, NL, and SV. What are the usernames? PrimeHunter (talk) 15:49, 29 April 2013 (UTC)

Those usernames are CY Jac-y-do, NL Torenkraai, SV Femten. It would make life a lot easier for me if none of these was a global account!

Does editing interwikis via the new procedure have anything at all to do with this? Or did I just click on a wrong button somewhare?

(My other usernames BTW -- I've now remembered a couple more -- are: BR Arvran, DE Picapica, ES La voz del oyente, FI Picapica, FR Le choucas, GA Picapica, NO Picapica) -- Picapica (talk) 16:23, 29 April 2013 (UTC)

Yes, that is what caused the problem. There's simply no (permanent) fix for it other than using one account and one account only, or never editing Wikidata again, which is frankly not going to be an option when phase 2 is in full swing.... --Izno (talk) 21:13, 29 April 2013 (UTC)

Which is the whole reason behind unified global accounts to begin with. One person, one username across the entire set of projects. It's the ideal - no need to forget (or remember) other userid's. (✉→BWilkins←✎) 21:20, 29 April 2013 (UTC)

Also see the announcement about enforcing single sign-on here: http://lists.wikimedia.org/pipermail/wikitech-l/2013-April/068937.html --AKlapper (WMF) (talk) 11:39, 30 April 2013 (UTC)

Wow. Am I reading that right? If an en.wp user has the same account name as anybody who ever SULed on any project, he gets it pushed to some long and unwieldy substitute? That seems like a really, really drastic change from before when you could only co-opt a nearly unused username. And how would you arrange the tracking of all those contributions? And they'd have to log in that way - I'd think half the users affected are going to think they forgot their password. And then they're all going to want to do renames, which they'll need assistance for ... sounds like a class 5 foulup waiting to happen. Wnt (talk) 16:33, 30 April 2013 (UTC)

See meta:Single User Login finalisation announcement; you may want to comment on its talk page. — Scott • talk 11:51, 1 May 2013 (UTC)

I would turn it around. If you are one of the few people who after a (in my opinion, grace period of) 6 years still didn't voluntarily created a global account, then now one will be created for you. If a global account with your name is no longer available, you will get a postfix attached to that global account and that is what you will have to use. There will only be one place for account renames. The talk page says, that they are looking at automatically redirecting users to their SUL account when they login with their old name and apparently people will be individually contacted as well. —TheDJ (talk • contribs) 18:47, 1 May 2013 (UTC)

To the OP: the weird thing is that it seems like whenever I enable scripts in Firefox (which I only do for Lua programming) I have to manually log in whenever I go to a page on a different project, despite having a SUL. But it works normally the 99% of the time when they're disabled. Wnt (talk) 16:33, 30 April 2013 (UTC)

That's because of how you are now protected by browsers from crossing domains with credentials. The technique we used over the past couple of years has actually stopped working in quite a few browsers (SUL2 is now being developed to properly solve this). The newest versions of Chrome, Safari and Firefox will only allow you to take your credentials to domains that you visited before. So if you regularly visit the sisterprojects, then it will work most of the time, if you don't then you might have to log in once in a while on a sisterproject. —TheDJ (talk • contribs) 19:11, 1 May 2013 (UTC)

My thanks for the replies and apologies for being slow to get back here to read them. Ho hum: looks like this is another case of the techies solving a problem I never had, while creating a new one for me in the process. I guess I'll just have to live with logging-out/logging-in on every occasion, until such time (it seems) as things get still worse. Such is Wp-life... :( -- Picapica (talk) 05:10, 9 May 2013 (UTC)

Prosesize JavaScript

I've been trying to get the JavaScript at User:Dr pda/prosesize to work for me. I've followed the installation instructions (including clearing my cache), but it still doesn't work. Would anyone have any suggestions for what I should do? Alphius (talk) 22:05, 29 April 2013 (UTC)

(I commented at the Help Desk, as the thread about it there is already active.) --Stfg (talk) 08:42, 1 May 2013 (UTC)

Announcing the release of the the Wikimedia Commons app for iOS and Android

The Wikimedia Foundation mobile features team is proud to announce the release of the official Commons app for iOS and Android! This mobile application allows users to log in with their Wikimedia account and view, upload, and share their Commons uploads with the image sharing sites of their choice. On Android, you can also batch upload and add categories to your images. We'll be working to expand the features on both versions of the app so that, in the future, you’ll also be able to participate in events like Wiki Loves Monuments and other focused uploading campaigns directly from this app. To let more Wikimedians with compatible mobile devices know about these apps, we'll be running banners on Commons and English Wikipedia to logged in users this week.

Those who are interested in uploading images via mobile on non-iOS/Android (e.g., using BlackBerry devices, running Firefox browsers) can still upload to Commons on the Wikimedia mobile web. Just visit the mobile site of your home wiki, log in, and go to the Uploads page in the left navigation menu.

If you have questions or feedback, please let us know! In addition to Bugzilla, we have a mailing list (mobile-llists.wikimedia.org) and an IRC channel, ##wikimedia-mobile ^connect, where we would be happy to hear from you. Maryana (WMF) (talk) 19:09, 29 April 2013 (UTC)

Aren't we opening ourselves up to an even bigger torrent of copyvio uploads from users who find images on the net, save them to their camera roll and upload them using the new app? --ukexpat (talk) 17:16, 30 April 2013 (UTC)

There are some checks that can be done that the photo was generated by the mobile device... not sure what is done, it would be good to know. See also lengthy related discussion on Commons about Mobile Web uploads, leading to disabling mobile web uploads for new users. Rd232 ^talk 17:26, 30 April 2013 (UTC)

Looking at the deletion rates for the various mobile upload methods on the mobile dashboard that does not seem to be happening :) YuviPanda (talk) 06:09, 3 May 2013 (UTC)

SUL account not attached to my account

Resolved

Can somebody explain me, why my account at nl.wikitionary wasn't corectly attached to my SUL account, although it was automatically created (as stated in the log)? I know that I visited the page logged in a few months ago and sadly there is no email set in the preferences. mabdul 16:04, 30 April 2013 (UTC)

It might be the same problem that I had with Urdu Wikipedia. --Redrose64 (talk) 16:48, 30 April 2013 (UTC)

I had a similar problem on Meta. I could see my account in Special:ListUsers, but couldn't log in to it. Meta didn't show on my Special:CentralAuth. Thankfully, I had an e-mail address set, so a password reset followed by Special:MergeAccount fixed it (hence Meta shows as "confirmed by password" in my Special:CentralAuth). The Meta account was "created automatically" on 1 August 2011, but only attached to my SUL account on 29 September 2011 (when I used Special:MergeAccount). Though I fixed my issue, I am still very curious to know what caused it.

If you don't have an e-mail address set, I suspect this can only be solved by a steward or a developer. – PartTimeGnome (talk | contribs) 22:48, 30 April 2013 (UTC)

Or ask a bureaucrat to rename the existing account and see if it works better next time the account is created. PrimeHunter (talk) 23:08, 30 April 2013 (UTC)

I simply will wait until the unification of all accounts to SUL; the other account will be moved and then I will have a completed SUL. Doesn't matter, I don't speak Dutch (although I might be able to read it) and I'm not active there. mabdul 20:17, 2 May 2013 (UTC)

 Fixed. The account was auto-created by CentralAuth, but due to a (now, to my knowledge, solved) bug it was created unattached and without password or e-mail set (making it impossible to log in and re-attach). I've imported the e-mail address from the attached account on en.wikipedia.org for User:Mabdul into the empty e-mail record for nl:wikt:User:Mabdul. Please use the "Password reset" feature and attach it with nl:wikt:Special:MergeAccount. Krinkle (talk) 22:09, 2 May 2013 (UTC)

Great. Thanks. mabdul 06:35, 3 May 2013 (UTC)

References Section

This new article version http://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Articles_for_creation/Irvine_historical_society&oldid=552773930 failed to show the AfC submission box - I tracked it down to the use of <references> instead of <references /> - it just causes the rest of the page to blank and shows no error. I wonder how many articles could be in Wikipedia with such a simple error...?  Ronhjones  ^(Talk) 23:56, 30 April 2013 (UTC)

Probably a lot fewer than unclosed <ref> tags, which also eats the following text. If I ever find a way to detect the first, then we can do the same for this one. --  Gadget850 ^talk 16:09, 1 May 2013 (UTC)

"Detect" via what sort of process? A bot that watches recentchanges? An edit-filter? Periodic offline scanning of the database dump? Couple of regex approaches come to mind for either the wikisource or the screenscraped result. DMacks (talk) 05:33, 2 May 2013 (UTC)

A database dump is probably the easiest. Just find articles with "<references>" but no accompanying "</references>". Legoktm (talk) 21:41, 2 May 2013 (UTC)

Maybe a more general idea (to catch a whole bunch of related mistakes) is to fire up weblint on the database dump. Could catch "open-with-no-close" in general, of which these are two examples. And also <s> and <center> and... DMacks (talk) 01:04, 3 May 2013 (UTC)

VisualEditor fortnightly update - 2013-04-29 (MW 1.22wmf3)

Hey all,

Here is a copy of the regular (every fortnight) update for the VisualEditor project, so that you all know what is happening (and make sure you have as much opportunity to tell us when we're wrong, as well as help guide the priorities for development and improvement):

VisualEditor was updated as part of the wider MediaWiki 1.22wmf3 branch deployment on Monday 29 April. In the two weeks since 1.22wmf2, the team have worked on deploying the VisualEditor opt-in alpha to more wikis, and on the new features for VisualEditor's wider launch as the default way users will edit our wikis: Categories, Templates, References and Images.

Fixed a number of bugs and regressions, including the 'bold' button not allowing you to un-bold some text (47680), fixing the link inspector throwing out "undefined" in some cases and other issues (47413), cleaning up the deployment including making the feedback link language-variable and having VE allow users preferences on non-deployed wikis (42936), avoiding a Firefox bug where shift-right caused characters to be deleted (47711), fixing a failure to let users convert between paragraphs and headings when next to an inline node (41203), and changing the integration so that the VisualEditor is now the editor behind the "Edit" tab (47396).

Additionally, we adjusted our code that faces mw:Parsoid to adapt to changes there and not fail when editing non-current pages (47434), fixed a problem of blank paragraphs getting inserted in some cases (46800), worked around a bug in jQuery that meant that references and other tags were getting corrupted (47417 and later 47737), avoided leaving trailing paragraphs when inserting content at the start or end of a paragraph (46799), made some speed improvements in the editing surface and the back-end (47343), and informed the user when a failure happens on serialisation (47581).

In experimental code that is not live on the Wikimedia servers yet, on images (37870) we added initial support for displaying and editing inline and thumbnail images, resizing of images, selection of images and other generated content with cursor keys (38129), and drag-and-drop relocation of images and other nodes. For references (39599) we added some infrastructure support for references. For templates (39598) we ensured that templates re-serialise to their original HTML if unchanged (47394).

A complete list of individual code commits is available in the 1.22/wmf3 changelog, and all Bugzilla bugs closed in this period on Bugzilla's list.

Per the MediaWiki deployment roadmap, this should be deployed here on Monday 6 May.

Hope this is helpful! As always, feedback gratefully received, either here or on the enwiki-specific feedback page.

Jdforrester (WMF) (talk) 04:00, 1 May 2013 (UTC)

Why is this being made a default instead of having two edit tabs side by side or a dropdown menu? – Allen4names (IPv6 contributions) 04:40, 2 May 2013 (UTC)

The idea is to have two edit tabs side by side at least for the short-to-medium-term as people get used to it, with "Edit" going to VisualEditor and "Edit source" to the wikitext editor - this second function will generally be more for "power users" and so it's appropriate for the longer, more precise and slightly 'scarier' labelling to apply to that one.

In the longer term, we imagine that the "Edit source" button will go "below the fold" on the Vector skin - i.e., to move to the drop-down menu - for pages where there will be VisualEditor available (which will not cover all of them - for example, the MediaWiki: namespace would be very hard to make work visually given that it's about editing fragments of wikitext, and the Template: namespace does the same to a smaller extent and, more importantly, includes a very large amount of highly-complex ParserFunction calls that would be a mess to edit.

I hope this answers your question! Jdforrester (WMF) (talk) 05:41, 2 May 2013 (UTC)

Will VisualEditor not have a button (like CodeEditor does) to toggle between "visual" and wikitext views? Although the CodeEditor button could use some UI love, it's not at all obvious that's what it does. Anomie⚔ 11:20, 2 May 2013 (UTC)

Potentially in the longer-term, as this requires a bunch of messiness in terms of juggling wikitext that isn't stored in the databases through the Parsoid service. It's do-able but not a priority. Jdforrester (WMF) (talk) 21:40, 2 May 2013 (UTC)

Please, please, please can we have an option somewhere to disable vised completely and to revert to the "old" raw edit mode for users who want to. I have tried vised, but I prefer "the old ways". Thanks.--ukexpat (talk) 17:41, 2 May 2013 (UTC)

second. mabdul 20:04, 2 May 2013 (UTC).  

You will be able to use the "Edit source" button to edit in wikitext rather than using VisualEditor for the foreseeable future; I can't make an absolute guarantee, but there are certainly no plans to change it in the next few years. Jdforrester (WMF) (talk) 21:40, 2 May 2013 (UTC)

Linking to foreign Wikipedia pages

An idea just came to my mind. If a page not yet exists on the English Wikipedia, but the page exists in on Wikipedia in a foreign language, I think it will be great if there is link to that or those foreign Wikipedia page(s) somewhere while searching for the article. For instance; the page "Edwin Evers" don't exist on the en.wikipedia.org, but it exist on nl.wikipedia.org. I think it will be nice if there is a link to (in this case) the Dutch page or a link to the matching Wikidata page, on the search results page or on the non-existing page. Sander.v.Ginkel (talk) 08:43, 1 May 2013 (UTC)

• This is follow-up to discussion at Template talk:No article text#Edit request 30 April 2013. --Redrose64 (talk) 16:30, 1 May 2013 (UTC)

I agree, a link that performs the search on Wikidata at the top of the search results page would be helpful, especially for person articles. --JFH (talk) 20:22, 2 May 2013 (UTC)

Distribution of number of references by article?

Is there an easy way to get the distribution of the number of references in articles? I need it for a rather minor point, but I can imagine that the data might be otherwise useful. In my specific case, I would like to know who many of the 4 million articles have four or fewer references. --SPhilbrick(Talk) 22:29, 1 May 2013 (UTC)

I would ping database reports as my first stop for such a question. --Izno (talk) 22:38, 1 May 2013 (UTC)

Probably better to scan a database dump looking for ref tags, I'd have thought. Or did you want to include bibliographies? - Jarry1250 ^{[Vacation needed]} 23:10, 1 May 2013 (UTC)

If you scan articles looking for <ref>...</ref> tags, you might not get articles like NBR 224 and 420 Classes, and you won't pick up Actuary at all. But neither of these is in any way unreferenced. --Redrose64 (talk) 06:33, 2 May 2013 (UTC)

You might be interested that 1.9 million article have at least one Citation Style 1 template. Of course that leaves two million articles that either have included citations in some other way (e.g. manually, rarer templates, bare urls) or have no citations at all. Dragons flight (talk) 05:05, 2 May 2013 (UTC)

I recently asked a similar question after finding the interesting list at Wikipedia:Articles with the most references (Wow! A lot of references per article...  :-) I am looking for a stat/ report/ study about the growth of references in enWP in total, for the last year or something similar.

• On german Wikipedia, users goiken and Svebert looked at the growth of "<ref" in article name space between 10. April 2011, 29. Oktober 2011 and 13. April 2012 (based on dumps, reported here: [9], [10], [11]). They found a growth of references of 36,4 % (from 2,6 to 3,6 million), while the number of german WP articles grew 13,5 %. The average number of references per article was 2,1 in 2011 and 2,5 in 2012.
• I found some very old stats from enWP of 2008/2009: User:Dr_pda/Article_referencing_statistics and User:WolterBot/Cleanup_statistics. "the average number of citations per paragraph was 2.07 for FA, 2.06 for GA, 0.87 for A class, 0.51 for B, 0.26 for Start, 0.14 for Stub and 0.15 for Unassessed articles. This was out of a total of 2,251,862 articles (disambig pages and obvious lists excluded); 1,625,072 (72%) had no refs" [12].

It would be great to have some recent stats! --Atlasowa (talk) 08:00, 2 May 2013 (UTC)

Thanks for all the information supplied. For my specific question, the answers suffice. However, given the critical nature that references play, I'm a little surprised we don't have more formal studies. Like Atlasowa, I'd be interested in more recent (and more complete) stats. --SPhilbrick(Talk) 13:10, 2 May 2013 (UTC)

Category intersection - today

Hi - input and ideas welcome for a new idea for category intersection:

• Wikipedia_talk:Category_intersection#A_working_category_intersection_today

This leverages the catscan tool, but provides users an easy way to get to it. proposed as a sort of interim solution until wikidata is fully up and running, to tackle gender/ethnic divisions in our categorization tree. --Obi-Wan Kenobi (talk) 08:04, 2 May 2013 (UTC)

Unwanted additional space

in Usage_share_of_web_browsers#Summary_table there is unwanted space between "9.07" and "%" - note that in other cells there is no space before percentage sign. It seems to be caused by overuse of autoformatting templates and I have no idea how to fix this (except converting this template orgy into proper table). Problem seems to not be browser dependent. 89.74.119.184 (talk) 10:35, 2 May 2013 (UTC)

The source simply had a space. I removed it.[13] Another place in the section said 9.07% without a space. Maybe you were looking at that. PrimeHunter (talk) 10:50, 2 May 2013 (UTC)

Ops and thanks (I checked 9.07% without a space). Now I feel stupid 89.74.119.184 (talk) 12:32, 2 May 2013 (UTC)

We all make mistakes, and this one wouldn't crack my top ten :) --SPhilbrick(Talk) 13:17, 2 May 2013 (UTC)

Glossary templates

Hi, I am have a difficult problem with templates (difficult for me that is) and am hoping someone here can help. I am using {{term}} and {{defn}} to structure this glossaary. I have created myself the templates {{subterm}} and {{subdefn}} which merely add some extra left margin and then call term and defn respectively. They both work fine. I have also created {{glossback}} to provide backlinks back to where the reader jumped from in the article. This also works fine. However, there is a strange problem that I can't solve when they are both used in the same glossary. After subdefn is used, every line that uses glossback subsequent to the subdefn call has a huge amount of whitespace inserted between the backlink and the term it is linked to. Please take a look. Thanks, SpinningSpark 14:43, 2 May 2013 (UTC)

When {{{4|}}} etc. were not specified, a blank line was output. The easiest fix is to hide the linebreaks. --Redrose64 (talk) 18:15, 2 May 2013 (UTC)

Thanks very much for fixing. But I don't understand why the glossary was not broken as well before the first call to subdefn. The entries using glossback were ok where the previous line called defn instead of subdefn. SpinningSpark 19:23, 2 May 2013 (UTC)

Issue with Main Page on mobile, viz. it hasn't changed since Tuesday

Tracked in Phabricator
Task T50062

Pull up the mobile site en.m.wikipedia.org/wiki/Main_Page. You'll find that Adrian Boult is still the Featured Article, the Icelandic election is the top news story, and the ITN picture is still the unfortunate minaret (rather than the new King of the Netherlands). Why, and how can this be fixed? Lockesdonkey (talk) 00:16, 3 May 2013 (UTC)

How about now? I logged in on the mobile site and purged the page. Looks fine to me now. jcgoble3 (talk) 01:08, 3 May 2013 (UTC)

I purged my whole cache. No dice. Also, this is happening on two devices in two browsers: one on a MacBook Pro running Chrome (OS: OS X Lion) and on an iPhone 4 running Safari (iOS 6.1.3). Lockesdonkey (talk) 02:13, 3 May 2013 (UTC)

I can confirm that this problem exists for me viewing [14] in Firefox 23 on Windows 7, something I've never done. Definitely a server-side problem.  — TORTOISEWRATH 02:22, 3 May 2013 (UTC)

Firefox 20.0.1 on Windows 8 here, and after going back to the page now, it's back to the old version again. But if I log in, it shows me the current version. Sounds like bugzilla:44391 has now manifested itself on the mobile site (also see the discussion from the original version of this bug). jcgoble3 (talk) 03:17, 3 May 2013 (UTC)

Indeed.  — TORTOISEWRATH 03:47, 3 May 2013 (UTC)

Thanks for reporting this! The problem sounds like an issue in the MediaWiki server configuration. It would be nice if somebody who has this issue could send the software bug to the 'Bugzilla' bug tracker by following the instructions How to report a bug. This is to make developers of the software aware of the issue. If you have done so, please paste the number of the bug report (or the link) here, so others can also inform themselves about the bug's status. Thanks in advance! --AKlapper (WMF) (talk) 12:17, 3 May 2013 (UTC)

Reported as bugzilla:48062.  — TORTOISEWRATH 01:26, 4 May 2013 (UTC)

Where is the audit trail for all actions on an article under pending changes?

Resolved

 – It is called the review log. Orange Suede Sofa (talk) 19:36, 3 May 2013 (UTC)

On List of countries by GDP (PPP), I noticed that another reviewer had accepted a change that shouldn't have been accepted. I unaccepted the change and then rejected it myself. I want to notify the other reviewer of what I did and why, but the other reviewer's acceptance disappeared from the normal page history (after I unaccepted & rejected the change) and I can't find a record of the fact that the other reviewer touched the page at all. It doesn't show up in the other reviewer's contribs. Where can I find that record? Regards, Orange Suede Sofa (talk) 19:05, 3 May 2013 (UTC)

Um. The Pending changes log is at (curiously) Special:Log/stable; input the article name in the "Target (title or User:username for user):" window, like this. However, I don't think that's what you're after. --Redrose64 (talk) 19:22, 3 May 2013 (UTC)

My bad. I've updated the section title accordingly. What I'm after is an audit trail of the pending changes and their acceptances/rejections, so that I can find a complete record of the situation I described above. Regards, Orange Suede Sofa (talk) 19:29, 3 May 2013 (UTC)

(edit conflict) Worked it out. You want the "Review log", Special:Log/review. Near upper right of the article you should see a little box showing "Accepted (latest)" with a down arrow after it. Hover over that arrow and a further box appears; click the word "accepted". --Redrose64 (talk) 19:32, 3 May 2013 (UTC)

That's exactly it. Thank you! Orange Suede Sofa (talk) 19:36, 3 May 2013 (UTC)

Searching for string literals containing non-alphanumeric characters

Are there any tools that can be used to perform a full-text search for a string containing non-alphanumeric characters. Say, hypothetically, I wanted a list of all pages in mainspace containing the string !!. Doing this doesn't work, for understandable reasons; grep only searches page titles. This would be equivalent to the pseudo-SQL SELECT * FROM `pages` WHERE `namespace`="(Article)" AND (`body` LIKE "%!!%" OR `title` LIKE "%!!%"), but performing direct SQL queries on the Wikimedia servers is something I predict only Jimbo is allowed to do.  — TORTOISEWRATH 03:43, 4 May 2013 (UTC)

The only method I know of is to download your own copy of Wikipedia's text and then to scan it with the AWB Database scanner. Alternatively, drop a note on my talk page and I could run a search for you in my copy (a snapshot from 4th April). -- John of Reading (talk) 05:50, 4 May 2013 (UTC)

Hiding the BLP editnotice

Can someone help me figure out how to hide the BLP editnotice? As I edit pages using a magnified page view, it really takes up a disproportionate percentage of my screen. It should be possible to hide it using either CSS or JS (I think), but I can't quite figure out how to do either. Thanks! – Philosopher ^{Let us reason together.} 19:51, 3 May 2013 (UTC)

For reference, the editnotice template is at Template:BLP editintro and is placed on the edit screen by MediaWiki:Common.js. – Philosopher ^{Let us reason together.} 19:54, 3 May 2013 (UTC)

Yep, putting .editnotice_BLP_editintro{display:none;} into your CSS page should do the trick. Writ Keeper ⚇♔ 19:58, 3 May 2013 (UTC)

Your name seems to keep showing up on my userscript pages. Thanks again! – Philosopher ^{Let us reason together.} 20:09, 3 May 2013 (UTC)

Generally speaking, if you know where the message comes from - and you've already worked out that it's Template:BLP editintro - look inside that for a class=value that encompasses the whole message; in this case we have <div class="editnotice_BLP_editintro">. The CSS to use is the name of the class, minus any quotes, preceded by a period and followed by {display:none;} which gives

.editnotice_BLP_editintro { display:none; }

Some messages have an id=value instead of a class=value and for these, the technique is almost the same, except that you begin with a hash # instead of a period . --Redrose64 (talk) 20:32, 3 May 2013 (UTC)

I look at the html source of the rendered page to find such things. That can also work when you don't know where it comes from. Indeed, the html source says <div class="editnotice_BLP_editintro">. PrimeHunter (talk) 20:41, 3 May 2013 (UTC)

Yeah, that's what I do (and did above). Right-click on it and select "inspect element" (if you're in Chrome; I use Firefox, so I get a similar thing through the Firebug extension). Writ Keeper ⚇♔ 20:43, 3 May 2013 (UTC)

I often use a magnified page view too. My User:Bgwhite/vector.css has options to remove alot of the "distractions" to give me more room. There are comments to say what each option does. I also use User:PleaseStand/hide-vector-sidebar.js in my vector.js file to remove the left sidebar. A tab option is placed at the top so you can access the sidebar when needed. Bgwhite (talk) 21:36, 3 May 2013 (UTC)

Thanks, everyone! I did find the class element, but wasn't sure what to do with it. Now I've got a (semi-permanent) reminder on my page in this script. Awesome!

@BGwhite: I'll take a look at those, thanks for pointing them out! – Philosopher ^{Let us reason together.} 22:23, 3 May 2013 (UTC)

...and now my User:Philosopher/common.css and User:Philosopher/common.js are both longer and the 'pedia is more easily editable. – Philosopher ^{Let us reason together.} 22:58, 3 May 2013 (UTC)

Firefox has an inbuilt element inspector, which is IMO the best out there. Right-click and "Inspect element" or Alt+Q.  — TORTOISEWRATH —Preceding undated comment added 01:39, 4 May 2013 (UTC)

Yeah, but in the change from v19 to v20 certain aspects of it were altered and sometimes it's more difficult to find what you want. The worst is that the pane on the lower left (where the "source" is displayed) sometimes doesn't show the line being inspected, but the very top of the page - the <html> tag, <head>...</head> element and <body> tag, followed by some <div>...</div>s. Other negative changes include the pane which shows how the styles were determined: it used to be full-height, now it's crammed into the corner and a lot of scrolling is needed. The only positive I've found so far is the "box model" tab.

I didn't mention the "inspect element" feature before, because not all browsers have it, and for those that do, it varies widely. I perhaps should have mentioned "View source", since all major browsers have it; but it can be daunting for somebody unused to HTML. --Redrose64 (talk) 07:54, 4 May 2013 (UTC)

I completely agree; I noticed these negative changes, too, but I forgot about them since v20 (I'm on the nightly channel, so it's been a few months for me since the change). I still think it's the best one available, just not as good as it used to be. It's a shame that nobody seems to have made an extension to get the v19 one back.  — TORTOISEWRATH 00:38, 5 May 2013 (UTC)

Wikidata items on article

Hello. In greek wikipedia we can enable (at preferences) to show items from Wikidata, under the title of a page. How can Ι enable this in english WP? I can't find it. Thx. Xaris333 (talk) 01:30, 4 May 2013 (UTC)

You can enable that by adding the following code to your common.js, or request to add the tool as an gadget. As far as I know, it is not enabled as an gadget on enwiki.

// [[d:User:Yair rand/WikidataInfo.js]]
importScriptURI("//www.wikidata.org/w/index.php?title=User:Yair rand/WikidataInfo.js&action=raw&ctype=text/javascript");

--Snaevar (talk) 01:41, 4 May 2013 (UTC)

Couldn't do it. Can you help? My commons.js in my page. Xaris333 (talk) 13:07, 4 May 2013 (UTC)

Yes, I can help. The code needs to be at the bottom of your common.js file, below the bracket and semicolon - };, not above it like you did. So, the bottom of your common.js should look like so:

    'listOfUnavailablePagesOn': 'List of not available pages on'
};

// [[d:User:Yair rand/WikidataInfo.js]]
importScriptURI("//www.wikidata.org/w/index.php?title=User:Yair rand/WikidataInfo.js&action=raw&ctype=text/javascript");

--Snaevar (talk) 18:26, 4 May 2013 (UTC)

Thx! Xaris333 (talk) 18:57, 4 May 2013 (UTC)

Watch tab star

Hi - since switching to Vector recently, I've noticed there seems to be no rhyme or reason to the watch/unwatch tab at the top of the page. Sometimes it's written in text, sometimes as a filled or hollow star. It doesn't seem to be namespace specific, nor (as I first suspected) is it browser dependent (a lot of my browsing is done on a clapped out IE7 PC which I can't upgrade). Does everyone get this? Is there any reason for the inconsistency, and is there any css/js tweak I can do to make it appear the same way? An optimist on the run!   06:58, 4 May 2013 (UTC)

In your vector.js you have a script $('#ca-watch').removeClass('icon'); to remove the icon. However, it can only do this after the page has fully loaded, which can take a while. That explains the inconsistent result you are observing. —TheDJ (talk • contribs) 07:06, 4 May 2013 (UTC)

I should have thought of that - thanks. An optimist on the run!   07:58, 4 May 2013 (UTC)

I only get the star, never text. I'm using Firefox 23 on Windows 7.  — TORTOISEWRATH 17:03, 4 May 2013 (UTC)

Move page

I made an error when i moved a page from User:Edinburgh Wanderer/2013–14 Partick Thistle F.C. season and ended up at Edinburgh Wanderer/2013–14 Partick Thistle F.C. season. I then tried to move it to the correct page 2013–14 Partick Thistle F.C. season and got a database error. The database error created 2013–14 Partick Thistle F.C. season but as a redirect to Edinburgh Wanderer/2013–14 Partick Thistle F.C. season i.e not moving the page. Need an admin to fix it and not sure what caused the database error.Blethering Scot 22:34, 4 May 2013 (UTC)

Fixed. PrimeHunter (talk) 22:43, 4 May 2013 (UTC)

climate table at Homer, Alaska

I've just removed a rather nice table at this article because it was causing a serious formatting error that generated massive whitespace in the article. This is the second time this has been added, the second time it has broken the article, and the second time I have spoken to the person who added it about it but they seemed uninterested in doing anything to rectify the situation. It's a nice table and I'd like to have it in the article, but not if it is going to ruin the formatting. Anyone know how to fix it? Beeblebrox (talk) 19:34, 4 May 2013 (UTC)

It's fairly easy to fix - the problem is that the weather table has to be a particular width, and it "collides" with the infobox and the images on the right, causing it to be shoved down until after the images - hence the whitespace.

I've moved the images below the box for now (it may still have some whitespace on wider screens due to the infobox); the other solution would be to move the weather table much further down the page. Andrew Gray (talk) 20:34, 4 May 2013 (UTC)

That works for me, thank you. Beeblebrox (talk) 22:34, 5 May 2013 (UTC)

Generating a JS library for cleanup scripts

While I was fixing some bugs at our WP:AFC helper script (WP:AFCH) I was realizing that there are many scripts trying to archive many kind of cleanup the same way and reinventing the wheel. I want that these developers work together (more?) and generate a library for cleanup uncontroversial stuff like converting HTML markup to wiki markup, fixing common spelling errors, removing unnecessary HTML comments (the AFC project is generating many of them) and doing similar stuff.

Potianial script for such a library (I know of) would be:

• Yet Another AfC Helper Script
• wikEd/AutoEd
• User:Cameltrader/Advisor
• WP:EngvarB
• maybe the MOSNUM script

and additional this also could be added to Twinkle as a general cleanup while tagging pages.

Would anybody experienced with JavaScript and/or Regex would help me adding as many as possible uncontroversial cleanup lines to this library and improving Wikipedia by crowed developing? :-)

mabdul 22:13, 4 May 2013 (UTC)

My experience with both is limited, but I would be happy to contribute where I can. Technical 13 (talk) 22:55, 4 May 2013 (UTC)

Is it a case of 'reinventing the wheel, or is it simply 'many hands make light work'? I run several scripts. Each script has a unique scope with some minimal overlap. Of those, my formatting script does things similar to AWB general fixes, and it is possibly the one you would target as being of the most general in cleaning up function. The scripts all employ regexes, so copying them over per your thoughts should be relatively simple. Maybe some others can chip in about the collaboration model: for security reasons, scripts are currently protected and tied to the user. -- Ohconfucius ^{ping / poke} 01:11, 5 May 2013 (UTC)

One possible solution for using the central depository might be to call up script functions from another. This can be done so long as it's imported into the driver script, in the same manner as importing into vector or monobook. Tony1 achieves a 'one-touch' facility by installing my control script – a composite I wrote for that purpose. -- Ohconfucius ^{ping / poke} 02:09, 5 May 2013 (UTC)

You probably should start with AutoEd. It is a javascript program that calls other javascript modules. There is a module to convert HTML markup to wiki code, some others to correct formatting.

Typos project is used by AWB, WikEd and WPCleaner for spelling mistakes. Bgwhite (talk) 08:36, 5 May 2013 (UTC)

Thumbnail generation issue

I just uploaded this file for use here; the thumbnail in the page where it is used is displayed fine, but the server is refusing to generate previews of the image for use on the image page or non-full-resolution links therefrom. What is causing this; is it related to the database lockdown an hour or so ago?  — TORTOISEWRATH 01:37, 4 May 2013 (UTC)

Seems to work for me (thumbnail is displayed). If a specific thumbnail size does not display for you, please provide the size(s). --AKlapper (WMF) (talk) 12:51, 6 May 2013 (UTC)

Adding new content (with reference)

While moving pictures and copyediting in the Home Front (World War II) article, I forgot to add: 'New material to the "Britain" ' section, on 4 May 2013. Trying three or four times, both yesterday and today, for some reason, it will not accept my input when I press 'Save' (it's only on this page). But the new content is there; I've looked. RASAM (talk) 12:20, 5 May 2013 (UTC)

Are you trying to change the edit summary without changing content? --  Gadget850 ^talk 12:29, 6 May 2013 (UTC)

Looks likely that that's it. A way round it is to delete your new material, and then save, followed by re-adding it with the edit summary. Or easiest, just don't worry about it. If you're not planning to stand for an admin job in the near future, just leave it. No-one's going to block you or shout at you for missing one summary. I don't think summaries without content will go into the record. Peridon (talk) 15:09, 6 May 2013 (UTC)

Hang on - you did it at Home front during World War II not Home Front. You moved pics and left an edit summary there all right on that day. Perhaps that's the trouble... Peridon (talk) 15:15, 6 May 2013 (UTC)

Sorted. Peridon (talk) 08:58, 7 May 2013 (UTC)

Toolserver problems

Referring back to Toolserver funkiness from April 21, it's still crap. Don't know how much I've tried since then. But the message that keeps coming up today when I tried to access "Contributors" from an article's history page is,

Bad Gateway

An error occurred while communicating with another application or an upstream server.

There may be more information about this error in the server's error logs.

It only has to do with the Contributors, all other options under history have always worked fine. If I go into the Firefox "Page Info" or "Page Source", it had exactly that same message, nothing else. — Maile (talk) 14:59, 5 May 2013 (UTC)

I've been having random problems for the past couple days whenever I try to use anything on Toolserver. I'm getting "Bad Gateway," "Gateway Time-out," "Page not found," "en.wikipedia.org is not a valid wiki," whatever it feels like giving me. What's gone wrong this time?  — TORTOISEWRATH 01:43, 6 May 2013 (UTC)

I got it to give me X!'s Edit Counter this time, but the replication lag is terrible right now. Is this related to the database lock a couple days ago, perchance?  — TORTOISEWRATH 01:48, 6 May 2013 (UTC)

I began noticing this weeks before I first posted about it on April 21. It comes. It goes. And just when you need it, it malfunctions. Since I posted above today, I went over to toolserver and tried tools I knew about. More than half the time, I got that "Bad Gateway" message. Right now, it's working. But later, it won't. Very unreliable. — Maile (talk) 02:01, 6 May 2013 (UTC)

I've gotten nothing but "Page not found" errors at least since yesterday. Anybody here in contact with anyone over there? Rivertorch (talk) 05:08, 8 May 2013 (UTC)

Preferences date/time-setting

I changed the format and TZ of date/time in Preferences, but the software still shows the UTC time. Should be CEST since I'm in Italy. Please give me a valid reason for this behavior. All 4 other great Wikipediae do it right, only the English is different. The number format (.=>,) as well, but I can accept this as a geek :-), Kind regards from Tuscany,  Klaas|Z4␟V:  15:35, 5 May 2013 (UTC)

It works for me. What exactly does "Server time", "Local time" and "Time zone" say at Special:Preferences#mw-prefsection-datetime? Do you see the UTC time in page histories and user contributions like Special:Contributions/ZeaForUs? Signatures and other content based on wiki source is not affected by the setting (but a gadget can affect it). PrimeHunter (talk) 18:17, 5 May 2013 (UTC)

Browser info (name and version) also welcome. --AKlapper (WMF) (talk) 12:55, 6 May 2013 (UTC)

Patrolling new pages

I used to be able to patrol new pages (allthough I never did). Some time ago, however, I lost the ability. I can't mark a page as patrolled anywhere (as I can see), and the red and green circles to the right which used to be there when I looked at new pages don't show anymore. Suggestions? Regards, Iselilja (talk) 19:32, 5 May 2013 (UTC)

That has happened to me as well, and I did used to patrol them.--Gilderien Chat|List of good deeds 22:51, 5 May 2013 (UTC)

• • Have either of you disabled JavaScript? I believe that the badge that pops up when I mark a page as patrolled (there's a tiny link at the bottom of the page) is a JavaScript thingie, so possibly the entire thin is. But wait a minute... 1) did you know there's a link at the bottom of the page? and 2) are you using Page Curation? In which case what I just said may have no bearing. Hope this helps. Ignatzmice•talk 22:55, 5 May 2013 (UTC)
    • I was using the new NewPagesFeed and I got quite I large bar menu on the right. I have JavaScript on.--Gilderien Chat|List of good deeds 23:08, 5 May 2013 (UTC)
      • Mm, yeah, that's Page Curation (or some variant thereof) I believe. I don't use that. See if there's a link that says [Mark this page as patrolled] when you click on an unpatrolled page from Special:NewPages. Ignatzmice•talk 23:14, 5 May 2013 (UTC)

@Gilderien, Iselilja: When you visit Special:NewPagesFeed and review a specific page, is there a "Curate this page" link in the toolbox section on the lefthand side? If so, clicking that link should make the toolbar that you can use to review pages re-appear.--Eloquence* 01:08, 6 May 2013 (UTC)

• Thanks for input. No, I cannot see a "Curate this page" link. And I don't get this "Mark this page as patrolled" link at unreviewed pages anymore. (Maybe I should mention that my skin is Cologne Blue, but this has been so all the time). I believe I have the JavaScript installed as I am able to use my internet bank. Regards, Iselilja (talk) 05:15, 6 May 2013 (UTC)*
• Eloquence, Ignatz, Gilderien. NOW, I found it: A "curate this page" link, and after I clicked it the right side bar is showing up again. Thank you, very much, Iselilja (talk) 22:22, 6 May 2013 (UTC)

Hidden categories text size

I'm trying to find a way to make the hidden categories have a smaller font size than the main categories. Surely there's some CSS hack I'm not aware of. I've found <div id="mw-hidden-catlinks" class="mw-hidden-catlinks mw-hidden-cats-user-shown"> in the HTML code, if that helps. FallingGravity (talk) 05:58, 6 May 2013 (UTC)

Judging by this edit, try

div#mw-hidden-catlinks { font-size: 88%; }

which works for me. I put it in Special:Mypage/skin.css but I don't see why it couldn't go in Special:Mypage/common.css --Redrose64 (talk) 09:33, 6 May 2013 (UTC)

That worked, thanks! FallingGravity (talk) 18:53, 6 May 2013 (UTC)

Email notification

Hi, is the email notification system working? I don't seem to be receiving items although my email appears to be working, so I don't think it's a problem on my own emails. SagaciousPhil - Chat 09:51, 6 May 2013 (UTC)

Have you checked your notification preferences? Also, has your email adress been confirmed? — Edokter (talk) — 10:33, 6 May 2013 (UTC)

Yes, I've re-checked preferences and notifications and everything is set correctly. It was working fine yesterday evening but just stopped and there is still nothing coming through - for instance, I didn't receive notification of your response, just picked it up by checking my watch list. SagaciousPhil - Chat 10:53, 6 May 2013 (UTC)

All of a sudden in the last few minutes, email notifications have started to arrive. SagaciousPhil - Chat 16:45, 6 May 2013 (UTC)

Resolved

Skip to TOC - Skip to Bottom

Skip to TOC - Skip to Bottom appears at the top of any of my user space pages (i.e. sandboxes etc.). The problem is, it lays itself right across the Edit link for the lead, making it impossible for me to click on that. This has only started happening recently, but I'd like to know how to get rid of it. Only on user space pages. I've changed skins to see if it will at least appear at a different place on different skins. Nope, that thing totally covers up my Edit link for the lead section. Zooming in or out with my browser does not change the position. I have absolutely no need for this "Skip to..." message, and it just hinders my editing ability on my own pages. Please tell me how to turn it off. — Maile (talk) 11:54, 6 May 2013 (UTC)

And...it's not on all my user space pages. It's just hit and miss as to which one it appears on. No pattern I can see what makes it appear on one user space page and not another. Has nothing to do with whether or not there is a lead section. And it has nothing to do with whether or not there even is a TOC on the page. This is strange. — Maile (talk) 12:17, 6 May 2013 (UTC)

Added nostb for you. — Bility (talk) 18:14, 6 May 2013 (UTC)

Thank you very much. That worked. — Maile (talk) 18:20, 6 May 2013 (UTC)

Well, it worked on my main user page. It did not work on This One. So, I'm wondering what I did in setting up this particular page that triggered it. — Maile (talk) 18:27, 6 May 2013 (UTC)

OK. Well, I see that something in the navboxes was triggering it. So, I removed the navboxes. End of problem. — Maile (talk) 18:29, 6 May 2013 (UTC)

There was already an instance of {{Noticeboard links}} on the page, so adding another one with |nostb=yes didn't do anything. Just needed to add the parameter to the template that was already on there. Cheers, — Bility (talk) 18:31, 6 May 2013 (UTC)

Invalid Search redirect to C2.org

Typing Wiki:SOFIXIT (okay I forgot what the proper link for shortcut, it should be WP:SOFIXIT) into the search bar on FF or the search bar on wikipedia itself links to C2.org instead of a page saying no link exists. ηoian ‡orever ηew ‡rontiers 19:53, 6 May 2013 (UTC)

It's a feature. Typing wiki:anything you like will always go to http://c2.com/cgi/wiki - it's one of the prefixes listed at m:Interwiki map. Same goes for wikilinks like wiki:anything. --Redrose64 (talk) 20:13, 6 May 2013 (UTC)

Indeed. But here's a slightly confusing detail: if you type <known interwiki prefix>:<any text> and press search rather than go, our search results page tells you "There is a page named "xx:yyy" on Wikipedia". This is wrong in two ways: (1) the page might or might not exist, it's only the prefix that exists for sure; (2) the page will not necessarily be on a site called Wikipedia.

It's obvious where this comes from, since the prefix is known the whole thing is a valid link from Wikipedia's point of view, so the search engine claims we have that page. But the phrasing is not quite right in the case of interwiki links. — HHHIPPO 20:28, 6 May 2013 (UTC)

The default Vector skin doesn't have search and go buttons. The search button in other skins corresponds to "containing..." in the drop-down box. ηoian is far from the first to be confused by wiki: going to c2.com. There is a suggestion to remove this prefix at meta:Talk:Interwiki map#Wiki. The discussion could actually use some technical expertise. Anyone have a method to determine the number of uses in Wikimedia wikis? If the prefix is removed then presumably a bot should change the uses at all wikis to one of the three other prefixes for c2.com. Can the uses be identified? PrimeHunter (talk) 22:45, 6 May 2013 (UTC)

https://

Resolved

 – Back up since circa 14:00, 7 May 2013 (UTC)

The https:// version of the site is inaccessible. Is this happening to anyone else?--Gilderien Chat|List of good deeds 19:54, 6 May 2013 (UTC)

• Not me, sorry. Ignatzmice•talk 19:56, 6 May 2013 (UTC)
• Nope. ~ Amory (u • t • c) 20:04, 6 May 2013 (UTC)
• Inaccessible to me too. --Roberto Segnali all'Indiano 21:11, 6 May 2013 (UTC)
• I'm having the same problem. The regular version works fine, but the HTTPS version times out (which is kind of a pain when you have http=>https rewrite rules in effect in your browser). Kolbasz (talk) 21:47, 6 May 2013 (UTC)
  • http://status.wikimedia.org/ says it's working. Steven Walling (WMF) • talk 23:14, 6 May 2013 (UTC)
    • It just came back up. Let's hope it stays that way. Kolbasz (talk) 14:06, 7 May 2013 (UTC)
• Works fine for me. I'm on it right now  — TORTOISEWRATH 23:47, 6 May 2013 (UTC)
• Not working here in the UK. Gilerian and Roberto both appear to be in Europe (not working) whereas Tortoise (working) seems to be in the US. Also seems to be slower than it used to be. Martin451 (talk) 00:01, 7 May 2013 (UTC)
  • https:// works for me too. Routing information etc. and trying with a different internet provider first is very welcome. --AKlapper (WMF) (talk) 08:48, 7 May 2013 (UTC)
• Down for me, too -- https://91.198.174.225 timeouts.
  Judging by Special:Contributions/Amalthea (bot) since 18:10 UTC yesterday. :/
  Amalthea2 (talk) 13:09, 7 May 2013 (UTC)

tracert
5 35 ms 42 ms 29 ms r1fra1.core.init7.net [80.81.192.67] 6 30 ms 48 ms 38 ms r1fra3.core.init7.net [77.109.128.222] 7 42 ms 43 ms 43 ms r1fra2.core.init7.net [77.109.128.245] 8 56 ms 48 ms 46 ms r1ams2.core.init7.net [77.109.128.201] 9 48 ms 45 ms 47 ms gw-wikimedia.init7.net [77.109.134.114] 10 45 ms 42 ms 42 ms wikipedia-lb.esams.wikimedia.org [91.198.174.225]

reverting single edit (not last one)

When someone does something obnoxious like this 8 Feb edit, which removed all ref tags, and some other code, I know one can restore to the version prior to the edit, but that will wipe out the couple dozen subsequent edits. Is there a solution which undoes that one edit without disturbing the subsequent edits? I'm fairly sure the answer is no, but want to ask, just in case.--SPhilbrick(Talk) 22:20, 6 May 2013 (UTC)

You can sometimes undo the edit, as long as there is no conflict with any of the following edits. However, it doesn't look like that will work in this case. --Bongwarrior (talk) 23:55, 6 May 2013 (UTC)

Yeah, I tried that, but it failed. Oh well.--SPhilbrick(Talk) 00:44, 7 May 2013 (UTC)

Resolved

by doing it manually.--SPhilbrick(Talk) 12:45, 8 May 2013 (UTC)

Template:Commons category multi

Could someone take a look at Template:Commons category multi? When there are 2 categories listed, the box is not "clearing right". It's aligned right, but it floats to the left of other right-aligned boxes. See Category:Horticulture and gardening, it should be under the "infobox catalog", not to the left of it. Thanks, --Funandtrvl (talk) 00:06, 7 May 2013 (UTC)

I think I took care of it. --Izno (talk) 00:24, 7 May 2013 (UTC)

Yes, I think that worked. Thanks! --Funandtrvl (talk) 00:28, 7 May 2013 (UTC)

Custom CSS - color change

Using: Firefox, latest version

What I want is to make the background display the color I choose. My CSS code so far (within "body { }brackets):

   font-family: DejaVu, serif;
   font-size: 125%;
   color: #FFD6AD;
   background-color: #FFD6AD;
   content: #FFD6AD;

This turns everything my chosen color except the background of the article itself -- that's still white. What am I missing? --Bluejay Young (talk) 02:27, 7 May 2013 (UTC)

Try this outside body brackets:

#content { background : #FFD6AD !important; }

PrimeHunter (talk) 03:19, 7 May 2013 (UTC)

You should also remove the

content: #FFD6AD;

from its present position. The content: property can't control colour, and so it isn't expecting a colour value. Another point: if you have settings like

    color: #FFD6AD;
    background-color: #FFD6AD;

what you're asking for is text of this colour on a background of this colour which looks like this and that has serious WP:CONTRAST issues, even if you have 20/20 vision. --Redrose64 (talk) 09:49, 7 May 2013 (UTC)

Yikes, no, that wasn't what I had in mind. I'm trying to make Wikipedia display everything in my Irlen color (or something close to it). What about the very top area, where (at least in Vector) your "talk," "sandbox," "preferences," "watchlist" etc. links & your searchbox are? What is that called/how would I change it from white to my color? Thanks! --Bluejay Young (talk) 13:17, 16 May 2013 (UTC)

Template: Conjugate

Could someone of HTML/CSS experts review my template {{conjugate}}? You can see the actual use in composition algebra. When you helped me with possible shortcomings, I would back-port the style to {{sqrt}} and {{radic}} which suffer from (^e√þ same) √disease as {{overline}}. Incnis Mrsi (talk) 07:24, 7 May 2013 (UTC)

I would advice against the packport. Both implementations may show better results on one, and worse result on other poeple's computers. It all depends on what fonts are used on the user's display. They actually look very similair to me. — Edokter (talk) — 23:46, 7 May 2013 (UTC)

Thanks, but does any reliable mean to control the gap between the top of glyphs and the overline exist? Or any calculation for upper border departs from some fixed line such as cap height and hence, a uniform solution for uppercase and lowercase letters is not possible? Edokter, please, watch your spelling. Making three typos in a short reply is awful. Incnis Mrsi (talk) 09:32, 8 May 2013 (UTC)

There is no 100% failproof way to position the overline. You can play with line-height, but that also varies with the used font. — Edokter (talk) — 10:15, 8 May 2013 (UTC)

Line-height has no effect either. Padding is already zero. If your goal is to make the line lower, then there is no way. — Edokter (talk) — 10:20, 8 May 2013 (UTC)

On my screen, {{overline}} looks actually slightly better in composition algebra then {{conjugate}}; the overline is slightly closer and scales better with font-size. — Edokter (talk) — 10:37, 8 May 2013 (UTC)

Universal account

Where can I discuss my situation regarding my accounts on other Wikimedia projects? Simply south...... eating shoes for just 7 years 10:02, 7 May 2013 (UTC)

There was some related discussion at Wikipedia:Village pump (technical)/Archive 111#Unwanted global account(s), which may contain helpful links. --Redrose64 (talk) 10:42, 7 May 2013 (UTC)

Thanks although mine is a different situation. I have now expanded a little and left a note at WT:Unified login/Finalisation if you want to reply there. Simply south...... eating shoes for just 7 years 10:53, 7 May 2013 (UTC)

Mass removal of old indefinite rangblocks under controlled conditions

Hello,

There is currently an ongoing proposal at Village Pump (proposals) to review and remove most old indefinite rangeblocks on IPs, and to have their edits monitored for an interim period of time. The proposal has received nearly unanimous support from about a dozen editors, but there has not been any discussion so far on how it could be possibly implemented. Anyone with any technical know-how is welcome to comment on the proposal as well as to discuss on how to implement it, should it pass.

Thank you,

TheOriginalSoni (talk) 10:02, 7 May 2013 (UTC)

[Please comment on that discussion only]

Twinkle D-batch

See User talk:MZMcBride#Advice. I can't seem to get the D-batch function to work on Category:Candidates for speedy deletion or any other page for that matter. There's a bunch of G8 speedies for broken redirects that I wanted to batch delete. How to get D-batch to work, or other methods of batch-deleting? INeverCry 19:30, 7 May 2013 (UTC)

Create a new "trusted user" permission group.

See this discussion. The proposal is to create a new group which bundles a few miscellaneous rights that don't fit in any of the existing groups. User:PinkAmpersand suggests "Suppressredirect, move-subpages, markbotedits, unwatchedpages, maybe tboverride." The reason I'm requesting this is that I'd like to be granted "move-subpages" permissions, but it isn't currently granted for any groups but admins. In the discussion linked to, I proposed adding it to Filemover, but that didn't fly. Klortho (talk) 12:55, 8 May 2013 (UTC)

Require IP/anon edits to enter an edit summary

Just wondering if the Wiki software was capable of enforcing the entry of edit summaries when users are not logged in (i.e. IP/anonymous users)? Registered users would be unaffected; they could continue to leave blank edit summaries based on their judgement (and lack of criticism from other editors). One notorious problem is that IPs will change information on articles (vital stats or facts) without explanation - often it is introduction of errors; sometimes it is choosing data from conflicting and/or unreliable sources; occasionally a valid point is being made, but the editor is not indicating that. The countervandalism and edit review process would be expedited if IPs had to enter an edit summary, ideally with a prompt which explains the need for reliable sourcing and the value of edit summaries. Or we could keep the status quo and continue to waste a lot of collective time attempting to read minds when dealing with most IP edits. Dl2000 (talk) 21:21, 8 May 2013 (UTC)

You can't force a useful summary...so why force a summary? --Onorem♠Dil 21:25, 8 May 2013 (UTC)

agree with Onorem. we currently get a lot of the "often it is introduction of errors; sometimes it is choosing data from conflicting and/or unreliable sources; " that ARE accompanied by edit summaries. And those summaries are rarely "I am adding bad information". -- TRPoD aka The Red Pen of Doom 21:29, 8 May 2013 (UTC)

Example. This is in fact useful, because there is a lot of unsourced WP:CRYSTAL being added to articles about British heritage railways - and although the IP address changes at least once a day, the edit summary is normally "Update!!" --Redrose64 (talk) 21:35, 8 May 2013 (UTC)

Echo

Modern skin and Echo

I made some local CSS changes to the Modern skin, so that these users get a bit better Echo experience. If ppl report issues with Modern skin not relating to Echo in the coming hours, this might be a candidate for reverting. —TheDJ (talk • contribs) 22:35, 1 May 2013 (UTC)

• When I click on the notifications button, the left hand one-third of the window that opens is off the left-hand side of my screen [15]. this is the first time I've done this, so I don't know if it worked before your change. Black Kite (talk) 10:53, 2 May 2013 (UTC)
  • Try bypassing your browser cache. If that doesn't help then it's because my changes don't work on IE and unfortunately I don't own Windows, so it will be difficult for me personally to help you along. Which version of IE are you using specifically ?—TheDJ (talk • contribs) 11:05, 2 May 2013 (UTC)
    • Bypassing the cache didn't work - it's IE9 (at work). Here at home, I'm on Firefox and it's fine. Black Kite (talk) 18:16, 2 May 2013 (UTC)
      • As I'm also a modern user, I will test it at work tomorrow. (only having access to a Mac right now) mabdul 19:54, 2 May 2013 (UTC)

Orange message bar

Where did the orange message bar go? All I get now is this little blue number in the upper right corner. Prettier, certainly, but far less effective. The orange bar was ugly as sin, but there's no way a new editor could claim he didn't know he had a message.—Kww(talk) 06:08, 2 May 2013 (UTC)

Just found Wikipedia talk:Notifications. I really don't understand how things like this just show up one day. How could anyone think making it difficult for a new editor to realize he's about to be blocked was a good idea?—Kww(talk) 06:20, 2 May 2013 (UTC)

It's gone. Choyoołʼįįhí:Seb az86556 ^{> haneʼ} 06:21, 2 May 2013 (UTC)

• LikeI actually quite like the powerful new functionality. Now a message on our talk page will give rise to a notification, ditto if someone reverts a change we make somewhere, and if you are mentioned on the talk page of another. There may be others which I have not yet discivered, but what's there is good. The little red blob with a number on it is discrete yet instantly noticeable, and will be more so once we get used to it. -- Ohconfucius ^{ping / poke} 06:31, 2 May 2013 (UTC)

I'll chime in with general support. Unfortunately, one of my first notifications was a revert, by an IP, of a recent edit. Embarrassingly, a good revert, but while it was on my watch list, my watch list is out of control, and I might not have seen it otherwise. Not opposed to restoring the orange bar, though.--SPhilbrick(Talk) 13:14, 2 May 2013 (UTC)

This appears to be a case of registered editors' preferences being put first, when there were very good reasons for not doing so. I'm certain that a comfortable majority of established editors would like to use the new implementation, indeed I would myself. But communication between all types of editor is essential, and the orange bar is without question easiest way of making sure that someone receives a message in good time (assuming they're logged in, of course). This is particularly true of newly registered and IP editors, but doesn't exclusively apply to them. In my experience established editors like to get rid of the orange bar as quickly as possible, and I doubt the same will apply to the relatively discrete number in the top corner. —WFC— _{FL wishlist} 05:47, 3 May 2013 (UTC)

• Question Is there a way I can get back the orange bar? "I am used to it" might not be the best justification, but I don't know whether I will recognize the small number. -- Toshio Yamaguchi 07:46, 3 May 2013 (UTC)
• The orange bar should be there for users that are not yet autoconfirmed, to ensure that they get their notifications. Remember that not all users even know they have a talk page. As for the rest of us, I'm sure a script will show up sooner or later to give it back to us if we want it. – Philosopher ^{Let us reason together.} 19:46, 3 May 2013 (UTC)
  • I wrote a script the other day that replicates the orange bar functionality; it's at User:Writ Keeper/Scripts/orangeBar.js. You can install it by adding the line importScript("User:Writ Keeper/Scripts/orangeBar.js"); to your common.js page. Writ Keeper ⚇♔ 19:49, 3 May 2013 (UTC)
    • Brilliant, thanks! Just to reiterate, though, I do think it should be turned on by default for users who are not yet autocinformed. – Philosopher ^{Let us reason together.} 20:01, 3 May 2013 (UTC)
    • Agreed, very cool. Thank you. -- Toshio Yamaguchi 20:03, 3 May 2013 (UTC)
      • Brilliant! Thank you! The new tiny little number (which is very very easy to overlook) is perhaps the single most stupid change in Wikipedia for a very long time (despite stiff competition). Good to have the orange banner back! Grutness...wha? 05:53, 4 May 2013 (UTC)

There is a new gadget (basically a copy-paste of Writ Keeper's script, except for a bit saying to get familiar with Notifications and a link to the doc page that has instructions to turn it off) has been created. Being a gadget, it is fully opt-out-able. The script can be reviewed at MediaWiki:Gadget-OBoD.js. It was ON by default, so new users would see it, but Edokter raised concerns about that. Should it be default-on? Comment at WT:Notifications#Pseudo OBOD Gadget is live. Ignatzmice•talk 19:50, 4 May 2013 (UTC)

There are far too many discussions going on in far too many places at this point. Can we just create a separate page to talk about the Orange Bar and related RFCs? Surely we can copy/paste all the current text to that and point links in all these pages / threads to that page. Now we have two gadgets for this, as well? (This is my understanding.) Killiondude (talk) 20:29, 4 May 2013 (UTC)

• Comment The 'official' discussion is at Wikipedia talk:Notifications (and THEY do seem to be listening to us), but the briefly turned on replacement has been slated for review here. Someone decided that a 107 to 19 consensus wasn't enough, or that the gadget was untested and needed reviewing over here. Peridon (talk) 20:55, 4 May 2013 (UTC)
  • I am not disputing the outcome of the RFC; just the way the code is deployed. Gadgets are subject to review, especially if they are going to be turned on be default. I could find no technical review, do I disabled the [default]. — Edokter (talk) — 21:06, 4 May 2013 (UTC)
  • It is urgent to restore a means of communicating with newbies. The WMF refuse to restore the "proper" orange bar, and are talking of the week after next for their solution. Please let us not start another discussion here: comment at WT:Notifications#On by default? JohnCD (talk) 21:17, 4 May 2013 (UTC)

I just got a message about vandalism on a library computer before signing in. So all is well, apparently.— Vchimpanzee · talk · contributions · 20:09, 8 May 2013 (UTC)

Notifications gadget's z-index

Tracked in Phabricator
Task T50078

Notifications gadget doesn't appear properly because z-index is same like the editor toolbar - 100. --Rezonansowy (talk) 12:01, 6 May 2013 (UTC)

I suspect you use the Visual Editor toolbar. I have already file a bug. — Edokter (talk) — 12:16, 6 May 2013 (UTC)

Gadget review please

I'd like some technical review for Mediawiki:Gadget-Notification.js and Mediawiki:Gadget-Notification.css for the potential event that it may be enabled by default. It's basically two lines of JavaScript, so I don't expect any problems. This may put a stop to a very large discussion at Wikipedia talk:Notifications. The gadget is intended to be active until the WMF team has decided on a solution for a new message indicator discussed at Wikipedia:Notifications/New message indicator‎. — Edokter (talk) — 15:11, 6 May 2013 (UTC)

Note: Screenshot here, if anyone cares to know. Ignatzmice•talk 15:15, 6 May 2013 (UTC)

Temporary notification system

I am going to enable a default gadget that will pop up when you receive a notification. I do this because there is a valid concern new editors may not notice the standard red blip on top of the screen. While there is consensus that some form of notification alert is needed, the current discussion at Wikipedia talk:Notifications/New message indicator is going downhill fast.

For more information, see Wikipedia:Notifications/Popup documentation. — Edokter (talk) — 00:30, 7 May 2013 (UTC)

Hi @Edokter: now that this gadget is being served to all new editors, and they're seeing it almost immediately after signup due to the welcome notification, I wanted to run a quick remote usability test to see if there are any glaring successes or problems we should investigate further. (The WMF has a usertesting.com account which we use for this sort of thing regularly.) I ran just two initial tests to make sure our test instructions made sense, and here are the relevant clips regarding the popup...

• Clip A
• Clip B

As you can tell, the good news is that the popup helped them see that they had notifications, and after that they successfully clicked the number indicator and opened their notifications. The bad news is that they both were confused by the red in the popup, and thought it indicated there was an error or problem, rather than a neutral notification. I think this supports the suggestion I made earlier, which was removing the red background. Of course Fabrice's team or someone still needs to run a larger experiment to get an objective look at clickthrough rates for various options on the table, but this is why we also do usability tests: more clicks are not always better, if new editors end up confused or irritated.

If you'd be interested in running another pair of usability tests to see if users react differently, let me know. Last but not least: the testers pick us, we don't pick them individually, but the person in Clip A said he had several thousand edits prior to the test, which is slightly unusual for usertesting as a site. Steven Walling (WMF) • talk 05:38, 7 May 2013 (UTC)

That is good feedback. I may test another design, it will have some red, as it matches the badge. User A expected orange, which only experienced editor would. I may also file a bug report on how to improve mw.notify, as it is quite inflexible; there is an option to make it persistent, but I can't set a custom timeout. — Edokter (talk) — 09:51, 7 May 2013 (UTC)

Toned it down for now. I have to go out for a few hours, will complete it tonight. — Edokter (talk) — 10:18, 7 May 2013 (UTC)

The relevant bug on mw.notify is 40307. Agreed about orange, that's why I pointed out he was an experienced editor. Steven Walling (WMF) • talk 19:20, 7 May 2013 (UTC)

You have new messages

I didn't see the orange bar that says I have new messages. I did happen to see a 1 that I had never seen before to the right of my name at the top of the page. I clicked and was told I have new messages. I like the old way better.— Vchimpanzee · talk · contributions · 19:56, 7 May 2013 (UTC)

Well, if you really want the OBoD in all its glory back, I made a script when this was first released to restore it: you can install the script by going to your common.js page and pasting in the following line:

importScript("User:Writ Keeper/Scripts/orangeBar.js");

. Thanks to a change Kaldari made yesterday (and forgot to tell me about, coincidentally; I had to stumble across it myself earlier today), it should be nearly identical to the way it was before; it just takes slightly longer to appear than it used to. Writ Keeper ⚇♔ 20:32, 7 May 2013 (UTC)

Thank you. I got worried when I couldn't find my question but see others have been concerned about the same thing.— Vchimpanzee · talk · contributions · 20:47, 7 May 2013 (UTC)

And I just found what I needed in the Help Desk archives. If I had been a little more patient ...— Vchimpanzee · talk · contributions · 21:18, 7 May 2013 (UTC)

Notifications popups

Hi, where was the consensus for this, and most especially for enabling it by default? -— Isarra ༆ 18:08, 8 May 2013 (UTC)

It was done at the Foundation level, not through any community based process. Beeblebrox (talk) 18:48, 8 May 2013 (UTC)

No, I don't think it was; I think that was User:Edokter doin' his thing. Not sure where he claimed consensus, though it may (or may not) have been an IAR-type thing to bridge the gap before the WMF response. Writ Keeper ⚇♔ 18:51, 8 May 2013 (UTC)

That was Edokter (and me, and a couple other people maybe) just trying to get something together in light of all the concerns raised about newbies not noticing the small little box. I don't think IAR was ever specifically cited, but that was the general mindset. It's meant to be temporary, until the WMF comes out with something better. Ignatzmice•talk 18:54, 8 May 2013 (UTC)

(ec) If anything, I claimed a complete lack of consensus between the community and the WMF. Recognizing the immediate concern raised by the community, I stepped in and made the popup as a temporary measure. Haven't had complaints so far. — Edokter (talk) — 18:56, 8 May 2013 (UTC)

Consider this a complaint, then. Such things require consensus, and your change attempts to resolve poor design in one regard - a lack of noticeability or indication what the number actually is - with even worse in another - a very annoying popup that does not stay gone when dismissed, that still does not do anything to say what the notification is (or even what a 'notification' is in general), and that has little affordance to connect it to what it is even referring to due to its visual and spacial separation from the p-personal toolbar. This is not a good change, and indeed a good example of why we operate on a consensus- and review-based model, so please revert this at once if it is still live. -— Isarra ༆ 22:14, 8 May 2013 (UTC)

Why should it stay gone when dismissed? The whole point is to let editors (new editors especially) know that something important has happened. If people don't want it, there's a link there do the documentation page explaining how to turn it off. Ignatzmice•talk 22:28, 8 May 2013 (UTC)

Actually, there was an overwhelming consensus to reinstate the orange bar, but that proved problematic in that the proposed solution was untested. So I created one that was based on an existing framework thus not prone to bugs. I also reiterate that it is a temporary solution. I just came from IRC on a collaborative meeting where we agreed on a new notification system that will hopefully meet the approval of the community. Until then, the red popup only takes one click in your prefereces to disable. — Edokter (talk) — 22:30, 8 May 2013 (UTC)

Hey Edokter. I appreciate the work you've been doing trying to design something to meet community requests, but considering that there are now gadgets being tested to meet this need, can we consider turning off the red popup as default? (I don't mean removing it entirely.) The new gadgets by Kaldari still need design work I'm sure, but they already do more advanced behavior like only appear on Talk messages, so they work as orange bar replacements for editors that want them. I am also asking because in about an hour, my team is launching a controlled test of a new version of Wikipedia:GettingStarted, and the red popup is likely to have a negative effect on our test results. This is because it fires immediately after you sign up, due to the welcome notification. Steven Walling (WMF) • talk 19:02, 9 May 2013 (UTC)

Hi Steven. That does create the situation that prompted the red popup in the first place, namely that new editors don't immediately see that they have messages. Would it be OK to have the F2 gadget replace it as a default gadget? — Edokter (talk) — 19:08, 9 May 2013 (UTC)

Can't wait for a reply now. I'll check back in 45 minutes. — Edokter (talk) — 19:28, 9 May 2013 (UTC)

As a side question: is there a list of vars like mw.echo.overlay.notificationCount that allow my to filter talk page messages? — Edokter (talk) — 19:17, 9 May 2013 (UTC)

@Edokter: I think the F2 gadget as default would be fine, as an interim solution. Steven Walling (WMF) • talk 19:37, 9 May 2013 (UTC)

I don't know about ones that tie into Echo or mention a notification count, but if you're just looking for whether there's a new talk page message or not, Kaldari put in a new variable called "wgUserNewMsgRevisionId" that has the revid of the earliest unread edit on one's talk page, or null if there aren't any, which is cleared once a user goes to their talk page. I've already switched my OBoD script over to using it (which, as a consequence, no longer uses any API calls and is now quite simple and compatible with any browser/skin, I believe). Writ Keeper ⚇♔ 19:42, 9 May 2013 (UTC)

Okay, until we figure this out and pending some other small todos my team has, we're just going to hold off on launching our A/B test of the new GettingStarted. That will give us a week to let everyone figure out what's the state of notifications. Steven Walling (WMF) • talk 20:14, 9 May 2013 (UTC)

@Steven (WMF): I'm OK with switching to F2. That will give us a chance to measure the community reaction. — Edokter (talk) — 20:24, 9 May 2013 (UTC)

If folks want to un-default the Notifications gadgets and make F2 on by default, I'm fine with that. I went ahead and disabled the animation for now, since it's a bit janky and still needs some work. The gadget definition is 'toolbaralert2|toolbaralert2.js|toolbaralert2.css'. MediaWiki:Gadget-toolbaralert2 will need to be updated. Kaldari (talk) 20:52, 9 May 2013 (UTC)

Also, I should mention that it currently doesn't localize, i.e. it only displays the alert in English. Kaldari (talk) 20:55, 9 May 2013 (UTC)

It has been done. I don't know if the gadget should be moved to the Appearance section. Feel free to do so. — Edokter (talk) — 21:35, 9 May 2013 (UTC)

• Hi Isarra, Writ Keeper, Steven Walling (WMF) and other contributors to this discussion: I am happy to report that Kaldari has now made the F2 gadget the new default for the message indicator feature. It should now appear in your top navbar after your name when you have new messages. If you would like to turn it off, simply go to Preferences > Gadgets and scroll down to the "Appearances" section, then uncheck 'Alert me when I receive messages on my talk page. (new)'. Special thanks to Edokter and Ignatz for helping us create this new solution, in collaboration with other IRC participants (see our IRC chat report) -- and for replacing the earlier RBOD with this joint F2 gadget solution. And many thanks to the rest of you for all your help in improving this feature! Fabrice Florin (WMF) (talk) 22:53, 9 May 2013 (UTC)

Notifications flag

How do I get rid of that Notifications tag at the top of my screen? I've disabled everything I know to within my preferences, but it's still showing up. My talk page is on my watch list, so I don't need that thing. Thank you! ←Baseball Bugs ^{What's up, Doc?} carrots→ 21:07, 8 May 2013 (UTC)

You can't. --Onorem♠Dil 21:08, 8 May 2013 (UTC)

Wikipedia talk:Notifications is the place to discuss it I guess, but it's not going away. --Onorem♠Dil 21:09, 8 May 2013 (UTC)

Was this done in place of that orange message bar that was there for years? ←Baseball Bugs ^{What's up, Doc?} carrots→ 21:16, 8 May 2013 (UTC)

• See User:Writ Keeper/Scripts/orangeBar. Also, yes, it was. Ignatzmice•talk 21:20, 8 May 2013 (UTC)

This should get rid of the little number (e.g. "(0)") at the top of the page (not fully tested):

li#pt-notifications { display: none; }

However, do not use the above unless you have some other talk notification gadget or user script enabled! You need an immediate notification of talk page messages of some sort; it looks really bad if you continue editing after someone sends you a message asking you to stop and discuss. (I doubt you look at your watchlist between every edit.) – PartTimeGnome (talk | contribs) 20:04, 9 May 2013 (UTC)

Edit-window bar blocks notification popup

Tracked in Phabricator
Task T50078

Sorry to contribute to the fun but when I click the notifications tag in edit mode, the top bar of the edit window (with "Advanced", "Special characters", "Help" and "Cite") blocks part of the notifications popup (which, FWIW, I like now :-)). My skin is Vector. Thanks and all the best, Miniapolis 21:05, 9 May 2013 (UTC)

Are you using the Visual Editor? If so, this is bug 48078. – PartTimeGnome (talk | contribs) 21:15, 9 May 2013 (UTC)

Add 'move-subpages' permission to the filemover group

I already suggested this under Proposals, but got no response -- perhaps this is the right place instead. It's a pretty trivial request, and I can't imagine there would be opposition. Perhaps the reason no one responded there was that it is too trivial, and really, implementing it is just a technical issue. The following is a slightly modified version of what I wrote there.

Filemovers have been vetted to ensure that they can be trusted to move images, for example. I would think they should also be trusted to move subpages. Use case: Right now I am trying to help clean up categories and pages in the GLAM area, and we are renaming several project pages to use the full name of the institutions. This is extremely tedious because some of them have many subpages, and they have to be moved one-by-one.

I was just granted 'filemover' rights, but that doesn't allow me the ability to move subpages.

Klortho (talk) 14:06, 2 May 2013 (UTC)

• Support I see no reason not to. -- King of ♥ ♦ ♣ ♠ 07:49, 3 May 2013 (UTC)
• Comment This initially seems like a fine idea, even though the two rights aren't terribly connected. I'm just a tiny bit worried that this would lead to filemover being assigned solely for the move-subpages right, though. Perhaps add this to rollback instead? I think the use of this right should be (generally) easy to undo, as with rolled back edits, so that would seem like a better fit to me. – Philosopher ^{Let us reason together.} 19:41, 3 May 2013 (UTC)

Either one would be fine. Please tell me -- where does this go from here? I've never suggested a change here before, and I'd like to know what is the procedure for actually getting it done. Do we need a quorum of "support" votes? Klortho (talk) 03:01, 4 May 2013 (UTC)

To be clear, I would support either version, but greatly prefer the rollback one. As for next steps, it'll wait a few days. If nothing happens, you can bump a (preferably) uninvolved experienced user or admin who posts on this board to judge the existing consensus and try to get it moved to the next step. I think the next step is filing a bug on Bugzilla, but I'm not quite sure. Hope this helps. – Philosopher ^{Let us reason together.} 02:14, 6 May 2013 (UTC)

• Oppose "filemover" has a clear purpose: moving files. Why screw that up by making it "file and subpage mover"?

  But the proposal in the comments to add it to "rollback" is even worse, as that has absolutely nothing to do with file moving, and I strongly oppose that. If you want to create a "slightly more trusted than the average user, but not enough to go through RfA" group, propose that (after reading WP:PEREN#Hierarchical structures) instead of trying to add rights piecemeal to rollback. Anomie⚔ 20:28, 6 May 2013 (UTC)

  • Really, we should just have a rights group with some of the obscurer admin rights that occasionally come in handy. Suppressredirect, move-subpages, markbotedits, unwatchedpages, maybe tboverride. Basically that category of rights that have too high an abuse risk to be bundled with anything else, but require nowhere near as much trust as the core admin tools. — PinkAmpers&^{(Je vous invite à me parler)} 23:11, 6 May 2013 (UTC)
    • That could be useful - a "trusted user" group, perhaps. However, that seems like a prime target for wiki-drama when asking which rights to bundle into it and who to give it to, so maybe it's not a great idea? – Philosopher ^{Let us reason together.} 04:28, 8 May 2013 (UTC)
• Oppose Moving subpages is a different thing. Create a "subpagemover" user group instead if this is needed. --Stefan2 (talk) 21:22, 6 May 2013 (UTC)
  • Creating a usergroup for something that trivial seems quite silly to me - there's no need to keep expanding the number of usergroups we have for each function that is made more widely available. When proposing that it be included with rolback, I wasn't considering function but level of generic trust, which should be the relevant criterion, I think. But perhaps ... what did you think of PinkAmpersand's idea? – Philosopher ^{Let us reason together.} 04:31, 8 May 2013 (UTC)
• Sigh It really shouldn't be this difficult for me to get this right. I have no interest in applying for adminship, but this particular right would let me work more effectively, which would benefit WP. Move-subpages, IMO, should be the default behavior: how often do you want to move a page, but leave all the subpages in place? So, I don't think there's a huge potential for abuse. Most people who get filemover probably wouldn't even notice that they had this new right. So, what is the problem, other than the aesthetic issue that it doesn't "fit" with filemover? Philosopher wrote: "I'm just a tiny bit worried that this would lead to filemover being assigned solely for the move-subpages right, though." -- Yes, that's what happened with me: in hopes of this proposal being accepted, I requested filemover, and got it. So what? I'm technically competent and trustworthy, and I'm not going to abuse filemover, and it might come in handy in the future. On the other hand, I agree with Philosopher that creating a "subpagemover" group for this is silly -- the whole point of having groups is to bundle rights, to keep things from getting too complicated.

Okay, I'll propose PinkAmpersand's idea of "trusted user". Klortho (talk) 12:49, 8 May 2013 (UTC)

• I think just adding this to rollback makes much more sense. This is not a common enough issue to warrant an entire new userright and it doesn't make a great deal of sense to merge it with filemover. Beeblebrox (talk) 18:44, 8 May 2013 (UTC)
• Support Beeblebrox' idea. This userright is meant for mainspace, as is rollbacker, and unlike filemover, and we don't need a new usergroup. Nyttend (talk) 23:17, 9 May 2013 (UTC)

Computer problems

1. The 'radio' buttons on all 'Revision history' pages always revert to the top two, after looking at any version - which makes a systematic trawl from the older edits to the most recent article page rather difficult. I've got a new computer with Windows 8 and Internet Explorer 10; the old one with Windows XP and IE 7 did not have this problem.

2. "Internet Explorer is not working correctly" is constantly being displayed. Sometimes it means that I lose eveything I have done. Again, I did not have this situation with the old computer. Anyone got any idea why and what I might do about it?

RASAM (talk) 21:30, 6 May 2013 (UTC)

As a start, do you experience the same problems with a different browser? --AKlapper (WMF) (talk) 08:49, 7 May 2013 (UTC)

They do mention this didn't happen using IE 7 on Windows XP, so I'd guess not.

Regarding the first point, you might find the "← Previous edit" and "Next edit →" links at the top of diff pages useful for tracing through history one edit at a time. Also, on a history page you can click the "prev" link next to an edit to quickly compare it with the previous edit without touching the radio buttons. (I appreciate that these tips aren't so useful if you want to compare edits in groups rather than one at a time.) – PartTimeGnome (talk | contribs) 20:23, 9 May 2013 (UTC)

File move strangeness

Can anyone figure out what's going on with File:Detail from Alfons Maria Mucha's "The Slavs in Their Original Homeland".jpg? I moved it to that title from File:Deatail from Alfons Maria Mucha's "The Slavs in Their Original Homeland".jpg, but now the image information page also displays as a redirect. I just can't figure it out. Thanks.--ukexpat (talk) 15:36, 7 May 2013 (UTC)

The redirect from the original location to the new location, was also on the new pointing to itself. - X201 (talk) 16:08, 7 May 2013 (UTC)

File:Detail from Alfons Maria Mucha's "The Slavs in Their Original Homeland".jpg (edit | talk | history | links | watch | logs)

Somehow you managed to do the move twice: first, moving the file page from the bad name to the good name, and then moving the redirect page on top of the file page. This needs an admin, I think, to recover the lost file page with its permission templates and such like. -- John of Reading (talk) 17:03, 7 May 2013 (UTC)

Jeez, so I did. I have no idea how or why I did that. Would a friendly admin in the neighbourhood please stop by and try to fix? Thanks.--ukexpat (talk) 17:42, 7 May 2013 (UTC)

I'm not entirely clear what the problem is. Is it that a redirect is listed at File:Detail from Alfons Maria Mucha's "The Slavs in Their Original Homeland".jpg#File usage? I don't think that's necessarily wrong: I recently moved File:Lewisham train crash 1857.jpg to File:St Johns train crash 1898.jpg and the old name is listed as a redirect at File:St Johns train crash 1898.jpg#File usage. --Redrose64 (talk) 18:34, 7 May 2013 (UTC)

It's more subtle than that. After the first move, at 13:04:34, there was a redirect at the wrong name and a proper file page at the right name. Eleven seconds later, the second move copied the redirect over the file page. The file was originally uploaded in November 2012, but the file page created then has been lost. Is it possible for an admin to recover it? -- John of Reading (talk) 19:13, 7 May 2013 (UTC)

There's very little in the logs for the old name, and nothing in the logs for the new name. There's nothing at the deletion archive, for either file. I rather think that it's beyond the powers of a WP:ADMIN, you need a specialist with direct SQL access. --Redrose64 (talk) 20:36, 7 May 2013 (UTC)

Can anyone recover the lost page from a database dump? This needs someone with a database dump and some Perl skills; I have the dump but not the skills. -- John of Reading (talk) 07:47, 8 May 2013 (UTC)

But that sounds like a bunch of work; wouldn't it be easier to put together a new page? We still have the image with its watermark, so we know its origin; this page says that it was published in 1912, so it's PD-US, and its Czech copyright status will depend on his 1939 death date. Meanwhile, a little confirmation on the unavailability of this data for admins: Special:Contributions/Chocolatemedia (the uploader) shows nothing around the time of upload except this, to add the new image to Mucha's article. Special:DeletedContributions/Chocolatemedia shows six edits, all of which are to Aslyoga, "an organiation that provides free yoga classes for the Deaf and Hard of Hearing communities". Nyttend (talk) 22:02, 9 May 2013 (UTC)

Namespace numbers

Wikipedia data structure

Namespaces
Subject namespaces		Talk namespaces
0	(Main/Article)	Talk	1
2	User	User talk	3
4	Wikipedia	Wikipedia talk	5
6	File	File talk	7
8	MediaWiki	MediaWiki talk	9
10	Template	Template talk	11
12	Help	Help talk	13
14	Category	Category talk	15
100	Portal	Portal talk	101
118	Draft	Draft talk	119
710	TimedText	TimedText talk	711
828	Module	Module talk	829
Former namespaces
108	Book	Book talk	109
442	Course	Course talk	443
444	Institution	Institution talk	445
446	Education Program	Education Program talk	447
2300	Gadget	Gadget talk	2301
2302	Gadget definition	Gadget definition talk	2303
2600	Topic		2601
Virtual namespaces
-1	Special
-2	Media
Current list (API call)

Who came up with these numbers? I guess the first few were created in chronological order, and the virtual namespaces in negative chronological order, but who decided that modules should be 828? -- Ypnypn (talk) 00:21, 8 May 2013 (UTC)

Its partially because it includes all Wiki's and their namespaces including some that have been deleted and some that were created and never launched. For example Wiktionary and Wikispecies both have some that Wikipedia doesn't have. I hope that helps. Kumioko (talk) 00:34, 8 May 2013 (UTC)

So there are a total of at least 414 namespaces, across all wikis across all time? -- Ypnypn (talk) 01:23, 8 May 2013 (UTC)

mw:Namespace_registration refers, but its list is far from comprehensive. LeadSongDog come howl! 04:34, 8 May 2013 (UTC)

150 is clearly what we as a community are lacking. ~ Amory (u • t • c) 06:09, 8 May 2013 (UTC)

Extensions are able to pick any namespace ID they'd like that's not already in use. I believe 828/829 wasn't chosen for any particular reason other than it's unlikely to conflict with any other numbers chosen. In practice, these numbers should matter very little to most users. ^demon^{[omg plz]} 12:55, 8 May 2013 (UTC)

I'm guessing that they're numbered in the first place to transcend language difficulties — if they program the system to treat namespace 3 in one way, namespace 75 in another way, etc., they don't need to worry about the names of the namespaces, while they'd need to worry about it if they told the system to treat namespace Benutzer Diskussion: in one way and namespace Anexo: in another. Nyttend (talk) 13:29, 9 May 2013 (UTC)

Well yes, that's how namespaces work internally. ^demon^{[omg plz]} 15:36, 9 May 2013 (UTC)

Errors...

It seems MediaWiki is choking this morning - been getting a few of the "Error" pages, and CSS styles(?) are pretty much borked on pages right now - all the text is displaying with absolutely zero formatting... - The Bushranger _{One ping only} 13:41, 8 May 2013 (UTC)

At least, in Monobook - Vector seems to be OK, and oddly, most of the edit window seems to be OK, but the actual pages are wholly unformatted. - The Bushranger _{One ping only} 13:45, 8 May 2013 (UTC)

• (edit conflict) Yeah, I've been getting that too. Sometimes the page doesn't load; sometimes the page loads w/o CSS. JavaScript (e.g. Twinkle, suggestions in the searchbar, right-click to edit sections) also seems to be gone (and I haven't messed around with my common.js that recently...) P.S. I'm in Monobook too. Ignatzmice•talk 13:48, 8 May 2013 (UTC)

I am seeing it too -- the hamsters need a rest or replacement.--ukexpat (talk) 13:51, 8 May 2013 (UTC)

Ditto. Using Vector, some pages take upwards of 30 seconds to load (although eventually of all them do). All Wikimedia sites are causing the same problem (at least WP, Meta, and MW). Sometimes the CSS doesn't load. I remember a similar thing happening a few months ago. -- Ypnypn (talk) 13:54, 8 May 2013 (UTC)

Seems to be fixed now... Ignatzmice•talk 14:05, 8 May 2013 (UTC)

This sounds a lot like the outage on Wednesday, 13:30 - 14:00 UTC, due to a memcached server going offline. "As usual this caused all kinds of cascading failures on other clusters such as Squid/Varnish. When not overloaded, these clusters would only serve cached pages at that point." Should be fixed now. Sorry for the problems. --AKlapper (WMF) (talk) 01:32, 10 May 2013 (UTC)

Gadgets in the Slovene Wikipedia

Tracked in Phabricator
Task T50278

Hi, my apologies for bringing this up at the English Wikipedia, however this page seems to be the best place to get an answer. In the last days I've spotted that all the gadgets have disappeared in the Slovene Wikipedia (sl:) (HotCat doesn't work, PopUps don't work etc., the 'Gadgets' tab in the Preferences has disappeared). Any idea what's the reason for this and what to do? --Eleassar ^{my talk} 10:38, 9 May 2013 (UTC)

This is weird. sl:Special:Gadgets is supposed to show a list of all gadgets available on the wiki, but it is blank. sl:MediaWiki:Gadgets-definition has not been modified since March. Based on this, it looks as if there is some internal problem with the Gadgets extension on slwiki. I suggest you either file a bug in bugzilla:, get on IRC #wikimedia-tech ^connect, or contact Duesentrieb (email). — This, that and the other (talk) 11:10, 9 May 2013 (UTC)

Ok, thank you. --Eleassar ^{my talk} 11:55, 9 May 2013 (UTC)

A null edit to MediaWiki:Gadgets-definition on slwiki will fix the problem. See also bugzilla:37228. Issue has been known about for a while, and was caused due to the partial outage that occured yesterday. Reedy (talk) 12:27, 9 May 2013 (UTC)

It has worked. Thank you. --Eleassar ^{my talk} 12:47, 9 May 2013 (UTC)

Preferred format for pronunciations

I'm doing Commons:Commons:Bots/Requests/Smallbot_9 where ~6.5k pronunciations will be uploaded to the commons. What is the preferred format for pronunciations: .ogg or .webm?Smallman12q (talk) 17:46, 9 May 2013 (UTC)

.ogg --Edgars2007 ^{(talk/contribs)} 17:49, 9 May 2013 (UTC)

User questioning if they've been hacked

Please advise if I should report this to anyone. On May 6, 2013, I made This revert, and posted a standard Warning template on the IP address talk page. Today, I received Spurious posting? inquiry on my talk page. I think the IP is used by many users, but thought I should mention this if there's anything that needs to be addressed. — Maile (talk) 18:17, 9 May 2013 (UTC)

It looks as if the user just didn’t notice that she is logged out.—Emil J. 18:28, 9 May 2013 (UTC)

That would be my guess. Besides which, the user doesn't seem to know exactly what their user account name is. — Maile (talk) 18:30, 9 May 2013 (UTC)

Relinking Google for SSL https

List of pages: wp:Google https links. -Wikid77 14:21, 8 May 2013 (UTC)

Note: #Confirmed stats.grok pageviews omit https. -Wikid77 10:38, 10 May 2013 (UTC)

There are still articles linked in Google with secure-server SSL prefix "https:" (rather than "http:"), and the pageviews in stats.grok.se have been down to 75% lower for some of those pages. To unlink and relink article "Parabola" now I have, step 1, renamed it as typical "Parabola (mathematics)" (which Google has indexed with "http:" prefix), while "Parabola" remains a redirect. The plan is to rename "Parabola (mathematics)" back again, and see if that resets and relinks the original title "Parabola" as simple http protocol. Then we can check the pageviews to see if they rise back near 2,500 pageviews/day from the recent 600/day. Questions:

• Should we wait a whole week before renaming back to "Parabola" to relink?
• What has caused some articles to relink as secure https in Google?
• Even if relinked as http protocol, will Google later return to https links?

If we can get confirmation, where "Parabola" can be relinked and return to nearly 4x times higher pageviews, then I would conclude that Google's secure-server https links are hindering users from reading pages. Meanwhile, I am considering to rename back within a few hours, unless someone knows it typically takes several days to reset an https link in Google. -Wikid77 (talk) 11:05, 25 April 2013 (UTC)

What the... ? I've undone this move (which broke the hatnote, by the way): we do not randomly move pages around based on our own personal theories as to how Google works, and I can't even comprehend the problem statement. If you want to experiment, you've got a sandbox to do so. Chris Cunningham (user:thumperward) (talk) 12:20, 25 April 2013 (UTC)

Other editors have been discussing Google's https article-links for some weeks now, but it is a very complex problem which will be difficult for many people to comprehend without weeks of study. Fortunately, this is a wiki, and there are other people to help handle complex issues. -Wikid77 13:22, 25 April 2013 (UTC)

• Double-renaming did not relink article as http protocol: After a period of 2 hours, the article was re-renamed from new http-protocol title "Parabola (mathematics)" back to "Parabola" but that did not immediately reset the Google https link as being http protocol, as would be expected if the old name were deleted, then the rename performed later. -Wikid77 13:22, 25 April 2013 (UTC)

• • File a bugreport in bugzilla with wmf site-requests. They have some contacts in Google. Will probably take long, but it can help. I doubt however that it really does matter a lot in traffic if Google found it primarily by https or by http. I wonder however if perhaps there is an issue with non-canonical duplicates of this page somewhere, which might affect it's page rank. If there are canonical issues, then that could also explain why google becomes confused and moves it to https as primary protocol. —TheDJ (talk • contribs) 14:02, 25 April 2013 (UTC)

I would be interested in knowing why we think (or more specificly how it happens) that google pointing people to https has a negative correlation with page views? Bawolff (talk) 16:31, 25 April 2013 (UTC)

• Several major articles with Google https links have lower pageviews: There is very strong evidence of articles with Google https links having relatively low pageviews in April 2013:

• Page https "Parabola" (stats-201304, 600/day formerly 2,500/day) lower than rare "Hyperbola" (stats-201304 1,300/day)
• Page https "Gone with the Wind (film)" (stats-201304, 1,860/day formerly 4,900/day) lower than "Lawrence of Arabia (film)" (stats-201304 2,800/day)

Because the term "parabola" is 5x more common in Google hits, the pageviews should be higher, not 50% lower, than for the rare term "hyperbola". Similarly, the article "Gone with the Wind (film)" (as an American cultural icon from 1939) should have higher pageviews than "Lawrence of Arabia (film)" (1962), rather than 33% lower. In general, highly popular articles should have much higher pageviews, compared to similar but less-popular topics. The renaming of page "Parabola" as "Parabola (mathematics)" was intended to delink "Parabola" in Google, to allow deletion of the redirect, and then re-renaming back to "Parabola". However, the deletion might have required a multi-day period with no redirect named "parabola" (as wp:SALTed to prevent re-creation), and that would likely have caused more debates. All of this is compounded by re-indexing a page in Google without revealing the complex trade secrets of how Google's PageRank algorithm operates. Hence, the re-indexing of https-link pages in Google might be better if handled as a discrete wp:OFFICE action, without a lot of public discussion about Google's company-proprietary PageRank methods. -Wikid77 (talk) 03:50, 27 April 2013 (UTC)

the article "Gone with the Wind (film)" (as an American cultural icon from 1939) should have higher pageviews than "Lawrence of Arabia (film)" (1962), rather than 33% lower. Sorry man, but that's completely bogus. — Scott • talk 12:26, 27 April 2013 (UTC)

May be so many people talking about it is an indication few people need to check out the article on parabola, but many need to check out the term hyperbola to find out what it is. Nil Einne (talk) 09:38, 1 May 2013 (UTC)

When many people talk about a topic, the article pageviews usually run higher, not 4x lower. Compare the higher pageviews for "Parabola" in prior months: prior year stats-201204 (1,927/day), prior month stats-201303 (2,490/day), versus April 2013 stats-201304 (620/day). -Wikid77 13:04, 6 May 2013 (UTC)

Now Hyperbola https drops pageviews 66%

In March, article "Parabola" had an https-link in Google, to drop pageviews 75%, while "Hyperbola" remained with "http:" link and held unchanged pageview levels. However, on 27 April 2013, the pageviews of "Hyperbola" dropped an average 66% each day, and now it is linked in Google by https secure-server protocol (just like Parabola). That is clear evidence that the https-links in Google are deterring people from reading those article pages, probably due to browsers which ask multiple security certificate alerts before the page can be viewed. The pageview stats from April 2013 averaged a 75% drop in pageviews for "Parabola" while the stats from May 2013 show a 66% drop in "Hyperbola" pageviews, when comparing:

• April Parabola stats-201304, 620/day, versus March, Parabola stats-201303, as formerly 2,490/day.
• May Hyperbola stats-201305, 405/day, versus March, Hyperbola stats-201303, as formerly 1,192/day.

There were no edits to Hyperbola (edit | talk | history | protect | delete | links | watch | logs | views) between April 14-29, so the switch from "http:" prefix, to https link circa 27 April 2013, is not directly tied to an edit during the 12 prior days. -Wikid77 13:04, 6 May, 05:42, 7 May 2013 (UTC)

I'm just now following this discussion about the https pageview drop. How do we know we are deterring human views instead of perhaps eliminating a large number of automated views from unknown sources? In our work on the WP:TOP25 charts, we have learned that certain articles have high ongoing viewcounts that do not appear to be caused by human views, e.g., Cat anatomy - it is regularly among the most viewed weekly articles in the raw stats produced at WP:5000, but its far too popular to be caused by human views, so we have excluded it from the WP:TOP25 chart. I am interested to know if there is evidence (anecdotal or otherwise) that we are deterring human views.--Milowent • ^hasspoken 14:32, 6 May 2013 (UTC)

It's all my friends who keep looking at Cat anatomy. Quite a turn-on! Thincat (talk) 14:51, 6 May 2013 (UTC)

• Lowered views of GWTW film poster and lower weekends indicate human viewers: Although it is an interesting possibility that 3,000 bots (or automated views) were reading "Gone with the Wind (film)" (GWTW) from Google every day, and have stopped due to the Google https-prefix link, the related 33%-reduced pageviews of the film poster strongly indicate the reduction is with human viewers. While pageviews of the GWTW film article have fallen over 62% per day, from average 4,900 in February/March to 1880/day in April/May, the pageviews of the GWTW film poster have also fallen, meanwhile, 33% from 112 to 75 pageviews per day (see: GWTW poster pageviews-201303, 201304). So, even if 3,000 bots per day were formerly reading the film article, I do not believe those bots were viewing the film-poster description page as 33% of prior pageviews. Instead, the reduced readers are humans, also inhibited to view the poster less (when not seeing the film article).
  Then, in comparing weekend pageviews, the article "Parabola" maintained its same 2-to-1, weekday-versus-weekend pageview ratio, when the pageviews dropped from range 3,150-1,650 on Sundays in March 2013, to 780-400 on Sundays in April 2013 (see: pageviews-201303, 201304). It is unlikely that automated pageviews would "take the weekend off" at the same rate as human viewers. Instead, a broad sample of human readers would be as likely to have reduced viewing on weekdays and lower weekends, at the same proportional rates, when all pageviews dropped over 60% lower. For those reasons, I strongly believe the reduced pageviews represent thousands of human readers who do not view pages, as often, with secure-server, https-prefix links. -Wikid77 (talk) 05:42, 7 May 2013 (UTC)

• Now 300 major articles have Google https: I created "wp:Google https links" as an essay/list of Wikipedia pages with Google "https:" protocol links. The results show more than 300 major articles now requiring secure-server access if clicked from Google Search. Perhaps the https links are related to fixes on the mobile website, where pages, with domain "en.m.wikipedia.org" were set to link as "canonical" but with https prefix to enwiki. In many cases, the daily pageviews have dropped nearly 60%-75% during March/April 2013. The https links include many common terms and major articles in various fields of study:

• Common articles with https Google links: Email, Water, Oxygen, Carbon dioxide, Ecology, iPod, iTunes, Amazon Kindle, Fluorescent lamp, Electric light, Scientific method, Rock of Gibraltar, Determinism, etc.
• Film articles with https: Filmmaking, Film noir, Gone with the Wind (film), Film genre, Venice Film Festival, etc.
• Medical articles with https: Cancer, Schizophrenia, DNA, in vitro, Cancer stem cell, Acute (medicine), Preventive medicine, Squamous-cell carcinoma, Psychosomatic medicine, Carcinogenesis, etc.

As noted, over 300 of the articles cover major topics in each field of study. Although many users can still gain access to pages by "https:" protocol (or retype as typical "http:" prefix), the reduced pageviews, for each major article, have skewed the measurements of reader interest in each topic, giving the false impression that those major articles no longer have a strong base of viewers. Several major articles even give the illusion of leaving the Top 1,000 most-viewed ("Cancer" or "Oxygen" or "DNA" in March), or "Nikola Tesla" or "Mark Twain" or "Shakira" or "Lady Gaga" in late April 2013, as if their daily pageviews were actually below 6,200 per day, rather than 7,000-12,000 or higher. -Wikid77 (talk) 14:21, 8 May 2013 (UTC)

Page view stats declining from 22 April

See: "wp:Village_pump_(technical)/Archive_110#Sudden drop in pageviews" and

see: "wp:Village_pump_(technical)#Relinking Google for SSL https". -Wikid77 17:59, 7 May 2013 (UTC)

Hi. I asked at VPM but no one could shed any light. The Page view stats at Depression (mood) began declining precipitously on about 22 April in a weird way. Other, but not all, articles have been similarly affected. Any Ideas? --Anthonyhcole (talk · contribs · email) 03:49, 1 May 2013 (UTC)

Possibly this is related to the fact that the mobile website was causing duplicate content in Google. bugzilla:35233 —TheDJ (talk • contribs) 19:14, 1 May 2013 (UTC)

I didn't understand that. I'd like to know if the actual traffic to this and other pages has dropped by two-thirds overnight, or of this is an artifact of our data-collection. If the number of visitors to some of our more important health pages such as Schizophrenia, Cancer and Depression has actually dropped to a third overnight, we should get to the bottom of it, don't you think? Does anyone else think this needs to be clarified?

It matters to me because I'm trying to get professional medical societies to review some of our medical articles, and that task will be made easier if I can show them reliable statistics.

If no one here knows what's going on, can you tell me who would/might know? --Anthonyhcole (talk · contribs · email) 06:11, 5 May 2013 (UTC)

• Google https links to Schizophrenia, Cancer and Depression (mood): As discussed about other pages since early April 2013, those articles (such as "Depression (mood)") have also changed in Google to have secure-server, https-prefix links, rather than typical "http:" prefix, as happened to several other articles in March 2013 ("Parabola" or "List of best-selling books" etc.). Although developers think the https-protocol requests have been properly logged, for pageview counts, some people think the pageview counts are still not correct (as perhaps undercounting the https transactions in those log files), while other people think that https security certificate alerts, in some browsers (such as MSIE not Firefox), have deterred people from viewing those articles, once their browsers ask about the secure-server access, and hence think the 66%-lower pageviews indicate actual reduction in readers viewing those pages. The reduced pageviews are shown by both stats.grok.se and the German Wiki-tech displays of pageview counts (such as for one-month "&Zeitraum=1M"). -Wikid77 (talk) 17:59, 7 May 2013 (UTC)

Heya, this is Diederik from the Analytics Team @ WMF. We are looking into this issue and I will report back once I have a satisfactory explanation. Drdee (talk) 20:29, 9 May 2013 (UTC)

Thanks, I have confirmed the https pages are omitted by stats.grok (see below). -Wikid77 10:38, 10 May 2013 (UTC)

Hi,

As promised, I would come back with an explatation regarding the pageview drop for https traffic. As mentioned before, we turned off the https/ip6 webrequest logging stream to webstatscollector back in March because we inaccurately concluded that we were double counting those pageviews. But in fact, by switching of that stream, we stopped counting of all https/ip6 traffic. Our apologies for this mistake.

We took the following steps to correct the situation:

1) As of May 14, 18:44 we re-enabled the https/ip6 stream to webstatscollector (see https://wikitech.wikimedia.org/wiki/Server_admin_log). Within the next hour or two, the pageview counts for articles that Google has indexed using the https protocol, should rebound to similar levels as they were before we switched off this stream.

2) One of the reasons we switched of this stream, was that the https/ip6 stream has a non-functional implementation of adding sequence numbers. Because these sequence numbers are off, our packetloss monitoring is incorrect. We fixed this situation by submitting patchset: https://gerrit.wikimedia.org/r/#/c/63220/

Please let me know if there are still lower than expected pageview counts in a couple of hours from now. Drdee (talk) 18:59, 14 May 2013 (UTC)

• Thanks for fixing the https/ip6 stream to webstatscollector: That was a fast fix, as I had expected to wait another 2 weeks to fit some update schedule. I can appreciate trying to handle the other packetloss problems at the same time, and the https pageviews would not have mattered so much if there were not other software problems which caused the http/https protocols to store 3 separate copies of pages/images in the Google Search index. Meanwhile, we had admins being rude, even snarky, about attempts re-index Google around the https-related bugs. So, not only were you quick to fix the stream to webstatscollector, you were also very polite about it at the same time, as they say, "a gentleman and a scholar" both. Thanks again. -Wikid77 (talk) 05:21, 15 May 2013 (UTC)

Confirmed stats.grok pageviews omit https

As feared, due to some major articles showing 75% drop in pageviews with Google https-protocol, I have confirmed the stats.grok.se omits the https-protocol pageviews. In tests run during 12 hours on 9 May 2013, over 60 https-prefix pageviews of a Space Shuttle mission-patch image file were ignored, while "http:" views were counted:

• http://stats-classic.grok.se/en/201305/File:Sts-99-patch.png - (logged 5 not ~65 https-protocol pageviews)
• http://stats-classic.grok.se/en/201305/File:Sts-98-patch.png - (logged 70 http pageviews)

Both mission-patch images ("http" for STS-98 and "https" for STS-99) were viewed within minutes of each other, repeatedly over 65 times, during the day-long testing of pageviews. The stats.grok.se website has counted only the "http" pageviews and omitted all https-protocol pageviews. Based on prior pageviews of major articles (see list: wp:Google https links), it is suspected that other pageview tools also omit the https-protocol page requests in the counts. -Wikid77 10:38, 10 May 2013 (UTC)

Huh

Why do people care about page views? I thought we were here to write an encyclopedia. — Scott • talk 17:37, 11 May 2013 (UTC)

• Comparing pageview counts helps to prioritize which articles to revise first: Although it would be great if all articles could be updated instantly, the reality, after more than 11 years, is that many major topic areas of Wikipedia are still maintained, or expanded, by "skeleton crews" of a relatively few people who cannot update all articles fast enough. Hence, by comparing pageview counts, then the editors can see which articles are read almost to the minute, versus some articles read only a few times per week. By updating the frequently-read pages sooner, and waiting to update the less-read pages, then even a handful of people can provide current, or recent, data at almost "magical" levels of response. So, when a news event occurs, or a scientific expedition publishes a report, then editors can use pageview counts to see which articles to update sooner. For example, with the "Costa Concordia disaster" in Italy, the pageviews revealed how, on some days, over 2x more readers were viewing the ship article ("Costa Concordia") rather than the "~disaster" subarticle, and so the ship article became a priority in posting some summarized data, as read by more people than the subarticle. Similarly, the pageview counts can reveal when a crime suspect's article is read many times more often than the original crime-event article, and so effort can be prioritized to update the suspect's article sooner, for summarized updates, rather than divert hours to update the crime-event article, while also writing or updating other articles in a timely manner. As a rule, "If it aint read often, then update other pages first". Eventually, some editors can become quite adept at prioritizing which pages to update sooner, and add crucial information, and allow more free time outside Wikipedia. It can be quite a relief when realizing some "important" articles are just not read enough, or templates used enough, to update today, or even this week. Relax, do something else, and reconsider later. It can wait. -Wikid77 (talk) 23:31, 11 May 2013 (UTC)

Thanks for the detailed reply. I work for the long term, so this approach is foreign to me. — Scott • talk 12:32, 12 May 2013 (UTC)

Internet Explorer

Tracked in Phabricator
Task T42998

Can anyone verify that IE is actually prompting users when visiting the https version of Wikipedia, and if so, what the exact warnings are? I'm unable to reproduce this using a variety of different settings. — RockMFR 01:56, 13 May 2013 (UTC)

Verified that they do exist, uploading and posting screenie with how I got it now. Tazerdadog (talk) 03:27, 14 May 2013 (UTC)

Scary message is here:

I got this by literally typing in the address shown in the address bar of the screenshot. (https://www.wikipedia.com/Hyperbola) Note the .com, this was an accidental oversight, but wikipedia is a .org site. I can provide more details, but I don't know where to look... Tazerdadog (talk) 03:41, 14 May 2013 (UTC)

That was caused by typing in .com instead of .org. Wikipedia's security certificate is only valid for the .org domain, so an attempt to access the .com address through HTTPS will always return a security certificate warning. Firefox gives me a similar warning if I click your link. Solution: type in .org instead. jcgoble3 (talk) 03:45, 14 May 2013 (UTC)

Ok, it fixes it for me if that change is made, so that is almost certainly not the problem. Tazerdadog (talk) 03:52, 14 May 2013 (UTC)

This is a known issue with the wikipedia.com domain (which only does redirects btw). I linked the bug tracking issue. —TheDJ (talk • contribs) 06:15, 14 May 2013 (UTC)

Some people assume that all internet URLs end with ".com", possibly the same ones that assume that they all begin with "www." Just last week, I had two different people complaining that they couldn't get to the desired website - one was inserting "www." where none existed, and the other couldn't believe that ".gov.uk" was a valid TLD. --Redrose64 (talk) 12:31, 14 May 2013 (UTC)

Talk to Google

I believe his is altogether the wrong way of going around things. We should ask Google to have a facility to refer to the pages on the domain using http: always like they have the facility to always refer to them with www. or not according to the web owners preference. This would get rid of anything like this and probably help people elsewhere too. Sites can always use canonical for ones which don't follow the standard. I think our messing trying to flush the caches by renaming is a waste of our time.

Thinking about it more Google would really like such a facility I believe as it would improve their statistics. They would probably also like the option of sites saying they can support https even though they marked everything as normally http so logged in users could have an option to automatically use https as often as possible when using google. Dmcq (talk) 14:15, 14 May 2013 (UTC)

Did I read correctly above that the mobile views were accessing articles with canonical set to https urls? What on earth is the sense in that? Dmcq (talk) 14:50, 14 May 2013 (UTC)

Anyway I do support having rel=canonical on all our pages to get the domain and http right whatever about being able to tell Google about the general situation. Might even catch out some of the mirrors before they figure out they should change that when copying Wikipedia :) Dmcq (talk) 16:01, 14 May 2013 (UTC)

In fact we should probably use HTTP headers instead of sticking the link into the head of the html. This will ensure even the images get a canonical url. Pity this won't catch out the mirrors but that's life. Dmcq (talk) 16:18, 14 May 2013 (UTC) Dmcq (talk) 16:18, 14 May 2013 (UTC)