Wikipedia:Village pump (technical)

The technical section of the village pump is used to discuss technical issues about Wikipedia. This page is not for new feature requests. Bugs and feature requests should be made at the BugZilla or the Village pump proposals page because there is no guarantee developers will read this page. Problems with user scripts should not be reported here, but rather to their developers (unless the bug needs immediate attention).

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. Questions about MediaWiki in general should be posted at the MediaWiki support desk.

view · edit

Frequently asked questions (see also: Wikipedia:Technical FAQ)

Click "[show]" next to each point to see more details.

If something looks wrong, purge the server's cache, then bypass your browser's cache.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.

No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See task 3864. There is an `accesskey` property on it (default to `accesskey="f"` in English). Logged-in users can enable the "Focus the cursor in the search bar on loading the Main Page" gadget in their preferences.

No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.

If you have problems making your fancy signature work, check Help:How to fix your signature.

If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.

If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.

For server or network status, please see Wikimedia Status. If you cannot reach Wikipedia services, see Reporting a connectivity issue.

« Archives, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212

Tool for editing and viewing wiki markup, offline?

I use Linux (Ubuntu 8.04, Hardy Heron). I would like to be able to copy and save the plain text versions of Wikipedia articles (with the wiki markup) and edit and view versions of the articles at home. I want to practice editing tables, and so forth, without having to use the Sandbox. How can I do this? Thanks very much! Whatever404 (talk)

Set up a local MediaWiki: mw:Installation. Amalthea 12:47, 20 April 2009 (UTC)[reply]

That sounds like fun- but can I spoil the party and ask, how. And if that question is too naive, how closely does the Turnkey Ubuntu version emulate our classes and settings, which are needed for table work? It does sound like a useful exercise. --ClemRutter (talk) 13:32, 20 April 2009 (UTC)[reply]

If you mean "classes" as in <table class="wikitable" …, you'll just want to make sure you copy MediaWiki:Common.css, MediaWiki:Monobook.css, etc. to the same title on your wiki. — CharlotteWebb 19:14, 21 April 2009 (UTC)[reply]

Thanks, everyone. Charlotte: if I install the items you mentioned, does that mean that the wiki environment on my computer will function in the same way as Wikipedia? Whatever404 (talk) 00:39, 24 April 2009 (UTC)[reply]

Just adding my thanks for a simple answer to a question I had not yet asked --ClemRutter (talk) 12:25, 1 May 2009 (UTC)[reply]

Whatever404: I see that you have no subpages in your user space yet. The usual way we test things is to create test pages under our own user page. Here is a link to your first personal test page: /Test1. Numbering your test pages usually is a good idea, since they tend to become more over time. I created my /Test57 some day ago... And I hope you have discovered the Show preview button when you edit pages?

--David Göthberg (talk) 16:01, 24 April 2009 (UTC)[reply]

Thank you, David, but I was interested to edit tables in ways that would probably only be useful to me, and would probably not be used in articles. Some of these edits are related to medical information that I would rather not make public. I guess it would be a more appropriate thing to ask the Wikimedia developers but I didn't really know where to go, to do that. Thanks. Whatever404 (talk) 16:53, 24 April 2009 (UTC)[reply]

I'll add that there are times I'd love a lightweight, offline app that I can use to test out tables, templates, and other wikitext structures without editing directly into wikipedia. it never actually occurred to me to set up a local wiki, and it sounds a bit like overkill. but...

--Ludwigs2 17:22, 24 April 2009 (UTC)[reply]

You could maintain the text & markup in whatever editor you like, copy and paste it into an edit window, and use the "Show preview" button to view the results. Ntsimp (talk) 14:39, 27 April 2009 (UTC)[reply]

yes, I know, but that doesn't actually qualify as 'offline'. --Ludwigs2 15:26, 27 April 2009 (UTC)[reply]

I hate underscores in links

Is there any script that can automatically strip internal links of underscores when saving the page? –xeno ^talk 22:36, 23 April 2009 (UTC)[reply]

I'm sure that AWB can do this, but not exactly what you want. I use User:Cameltrader/Advisor.js for a lot of cleanup— perhaps this feature could be added. ---— Gadget850 (Ed) ^talk 23:40, 23 April 2009 (UTC)[reply]

Sorry I didn't make myself clear, I meant maybe some kind of greasy monkies script that would strip the underscores on the fly as I'm saving the page... –xeno ^talk 23:44, 23 April 2009 (UTC)[reply]

I don't think it's a good idea to do it automatically, right now on this page there are at least 2 signatures with user_talk, and in articles some links may legitimately contain underscore as well (say, as a link to FILE_ID.DIZ). —AlexSm 15:53, 24 April 2009 (UTC)[reply]

Yea, you're probably right. I was hoping that there was some script that could intercept when I pasted a link with underscores and strip them. Probably too much effort for too little gain =) –xeno ^talk 15:59, 24 April 2009 (UTC)[reply]

It wouldn't realy break anyting though, as far as MediaWiki is concerned spaces and underscores are treated the same way in internal links (or rater all spaces are converted to underscores in the resulting HTML markup) so FILE_ID.DIZ and FILE ID.DIZ are both links to the exact same article. --Sherool (talk) 17:46, 24 April 2009 (UTC)[reply]

Well, I didn't say "break", just that it's not nice to change it automatically: in articles it might be intentional, and on talk pages it would creates unnecessary changes, resulting in more complex diffs. —AlexSm 16:27, 25 April 2009 (UTC)[reply]

Xeno, if you're willing to select some text and then press a button, you might want to try my script user:js/urldecoder with extra option urlDecoderIntLinks=true, although it does much more than just replacing _ to spaces. —AlexSm 16:27, 25 April 2009 (UTC)[reply]

I think this is exactly what I am looking for. Thank you. –xeno ^talk 14:32, 29 April 2009 (UTC)[reply]

You are a god among men. cheers, –xeno ^talk 14:37, 29 April 2009 (UTC)[reply]

backup

Does anyone know if wikipedia is backed up to a electromagnetic pulse resistant repositary? Andrewjlockley (talk) 23:16, 23 April 2009 (UTC)[reply]

No. A well placed EMP or asteroid strike would involve a degree of unrecoverable losses. A large fraction of the data would eventually be recoverable via various mirrors and dumps and such, but doing that would be a side effect of our distributed information culture and not really part of a data protection plan per se. Dragons flight (talk) 23:34, 23 April 2009 (UTC)[reply]

An EMP attack is pretty likely on a 100year timescale. There are many datacentres which offer pretty cheap EMP-resistant storage. At the very least, an antipodeal backup should be provided to ensure that the strike doesn't destroy all of the sites. Databarracks is one such UK provider. Andrewjlockley (talk) 09:09, 24 April 2009 (UTC)[reply]

Ha. Yes an eight-year-old organization should plan for a "100 year" event that has never happened in Florida (the main server farm) and pretty much can't be expected to happen save for nuclear war and the end of civilization. Backups are good, distributed backups are better, but let's not get silly about this. Dragons flight (talk) 09:32, 24 April 2009 (UTC)[reply]

Don't forget we also have a nice backup on the toolserver in Germany, so the disaster would have to take out Florida and Germany to be successful. MBisanz ^talk 09:46, 24 April 2009 (UTC)[reply]

Pssst: none of enwiki's text actually goes to Germany. Dragons flight (talk) 10:18, 24 April 2009 (UTC)[reply]

Neither do big chunks of the user table, or any deleted/oversighted revisions, or any media AFAIK. The toolserver is probably not a very useful backup mechanism: we'd have all 280 million revision entries, but no associated text! Happy‑melon 10:39, 24 April 2009 (UTC)[reply]

There's a full copy of everything at the Toolserver/Amsterdam-data-center except for images; toolserver users just don't have access to everything for privacy reasons. Toolserver users have access to views of the normal database tables, the text is on clematis. Mr.Z-man 22:51, 24 April 2009 (UTC)[reply]

No, but this was fairly close. SharkD (talk) 04:11, 25 April 2009 (UTC)[reply]

They're looking to build a complete backup datacentre to run the site (not just store it) somewhere in the northeastern US. I don't know if that will incorporate database slaves that would retain the site content in the event of the unrecoverable loss of the database servers in Florida. But in any case a huge amount of our 'current' content is stored in the memcaches of the Apache servers, which would definitely be duplicated 'up north', and on the squid servers in Florida, Amsterdam and Seoul; those are roughly evenly-spaced around the world. Someone would just have to write an 'unparser' to convert from HTML back to wikitext. And of course, anyone who's downloaded a copy of the database dump is storing a complete copy of the most important part of our data; that could easily be restored if necessary. If an EMP were to hit right now, we'd probably just be set back to August 2008, when the last full enwiki dump was completed, or to March 2009 with the last current-version dump (and a lot of messy licensing issues on who to credit going from one to the other). I'd say that between copies of the dumps, what data is replicated to the toolserver, and whatever's lying around on Brion's USB key (:D) we could probably recover pretty much everything up to a month or so ago; the most likely things to be lost are user preferences/watchlists, user passwords (which could be a bit of a bummer) and deleted pages. The oversight table would almost certainly be lost (bah humbug :D). But as DragonFlight says, thanks to the information age, it would be a tall order to actually permanently lose much 'public' content, even in an end-of-the-world scenario. Happy‑melon 10:39, 24 April 2009 (UTC)[reply]

The Seoul cluster, which was always the smallest of the three, is being decommissioned (or already was). It was part of a space sharing agreement with Yahoo that is ending / ended. Also, the last full history text dump of enwiki was more than 2 years ago, and not in 2008. We've had aborted dumps, partial dumps, and corrupt dumps since then. Yes we could get off the ground with the current versions dump (from March) and stub-meta-history for author lists and copyright purposes (last completed in October), but if an act of God destroyed Florida there would probably be a lot of article history that we never recover. Dragons flight (talk) 11:02, 24 April 2009 (UTC)[reply]

Ah, was it 2006 not 2008? Interesting also about the Seoul cluster; I wasn't aware of that. They're still listed on wikitech, so I guess the curtain's still up for the moment.

Actually, with the current versions dump from John Smith's computer, plus the revision and logging tables from the toolserver, we'd have all the info we needed for copyright purposes: a text to work from, and a complete list of its authors. So it would just set us back to March, and we'd have to delink everything before that magic 'first version' in page histories. It would be wierd, but not end-of-the-world. Happy‑melon 11:29, 24 April 2009 (UTC)[reply]

I think it's entirely reasonable to protect WP from the consequences of a nuclear/EMP war. I don't think distributing the data would be sufficient. I think that a nuclear-hardened backup should be specifically created. Andrewjlockley (talk) 12:17, 24 April 2009 (UTC)[reply]

If you want to put up the time and money to create such a backup, then you are certainly welcome to copy any and all materials from m:Database dumps to any repository you wish. As for Wikimedia working on such a project, I think it would be a waste of resources. Dragons flight (talk) 12:46, 24 April 2009 (UTC)[reply]

Besides, would there be anybody left to read WP after a nuclear war? Let alone restore it? Yintaɳ 21:46, 24 April 2009 (UTC)[reply]

We have most of our servers in a hurricane area and at sea level, and we're worried about an EMP? Mr.Z-man 22:50, 24 April 2009 (UTC)[reply]

What's the size of the backup? Bearing in mind that there are already easily accessible nuclear-hardened stores, what's the argument for not using them? Andrewjlockley (talk) 02:14, 25 April 2009 (UTC)[reply]

can you download the whole think to DVD's? they shouldn't be affected by an emp. Is anyone doing this already?

Dye-based media such as CDs and DVDs have a limited shelf-life, especially when exposed to light. Not sure whether magneto-optical devices such as minidisks (which have a longer shelf-life) are also susceptible to EMP. I believe Sony is developing/has developed a high-capacity storage medium based on this technology, but intended for institutional needs. SharkD (talk) 04:32, 28 April 2009 (UTC)[reply]

Concerns about Wikimedia's disaster recovery infrastructure should be directed to people who actually know what they're talking about, rather than random Wikipedians who read this page. You should also do your homework before asking anywhere. A lot of the information being thrown around here is third-hand "I heard it somewhere" kind of information. — Werdna • talk 13:50, 29 April 2009 (UTC)[reply]

To where should they be directed? Andrewjlockley (talk) 23:16, 29 April 2009 (UTC)[reply]

link to copyright vote doesn't work on secure server

On the secure server, clicking the "Vote now!" link in the notice board (for "Please participate in a vote to determine the future copyright terms of Wikimedia projects (vote ends May 3, 2009)") leads to a page with the text "Wiki does not exist" rather than the expected page about the vote. The current link is Vote Now! (broken), but should be Vote Now! (correct) (for English, anyway).

code broken on Rastafari movement

If I try to go to Rastafari movement I get this: <link rel="stylesheet" href="/w/index.php?title=MediaWiki:Print.c at the top of the page with the usual background but if I click Rastafari which redirects to Rastafari movement to works ok...

just though i should mention it. rdunn PLIB 14:35, 28 April 2009 (UTC)[reply]

I see no problem. Probably a caching issue. Try purging the page. — Edokter • Talk • 14:48, 28 April 2009 (UTC)[reply]

I tried that already but it didnt work. I just kicked the computer and it loaded correctly afterward... rdunn PLIB 14:51, 28 April 2009 (UTC)[reply]

The lesson here, boys and girls, is that violence is always the answer! EVula // talk // ☯ // 15:05, 28 April 2009 (UTC)[reply]

That's the true engineer's approach. If you can't fix it, hit it with a hammer! – ukexpat (talk) 15:06, 30 April 2009 (UTC)[reply]

auto signatures?

maybe I'm missing something obvious here, but wouldn't it be possible to have the software automatically add your signature? I mean literally - when you hit the 'Save Page' button, the software compares the old version to the new, checks to see if the last characters of the newly added bit (barring whitespace) are ~~~~ and if not, adds them. you'd have to exclude changes marked as minor, and changes that occur in the middle of a paragraph, but with a little regexp it's certainly doable.

of course, I don't know what kind of impact that would have on server load, or how that would compare to the current load induced by sigbots... --Ludwigs2 16:27, 28 April 2009 (UTC)[reply]

Technically it would be trivial. Socially, it would be annoying as hell, when you get signatures popping up in places you don't want them. How can the software tell between an edit that's unsigned because you forgot, and an edit that's unsigned because you don't want it to be signed?? Happy‑melon 17:19, 28 April 2009 (UTC)[reply]

well, I was just assuming that the only times you don't want an edit to be signed are easy to identify programatically (minor edits, fixes, section headers, etc.). As a rubric, anytime you add one or more full paragraphs to the body of a section (in the sense of text that begins on a fresh line and ends on a fresh line, and excludes headers) you want a signature; anytime you modify an extant paragraph (or add something that's only a section header) you don't. maybe there's odd cases (like adding a list item to someone else's example list) that need to be accounted for, but I bet we could account for 99% of talk and administrative edits using that rule. plus, we could always add a checkbox to suppress signatures for abnormal cases. I mean, the sigbots have rules for when they do and don't add signatures, and they seem to work ok. this just does the same thing with the advantage of immediacy (the software has access to both the current and and revised versions in a way that sigbots coming in after the fact don't). --Ludwigs2 18:00, 28 April 2009 (UTC)[reply]

This seems like an awful lot of effort for something that, ultimately, isn't the end of the world; the occasional unsigned comment is caught pretty quickly, and there are plenty of times where signing isn't needed. Typing "~" four times just isn't that hard. EVula // talk // ☯ // 03:49, 29 April 2009 (UTC)[reply]

I'm a fan of automating what can be automated. four characters may not seem like a lot, but multiply it by a couple hundred thousand edits per day and it adds up to a lot of collective wasted time. but yeah, it's not a huge deal. --Ludwigs2 04:18, 29 April 2009 (UTC)[reply]

Well, there's a rumor that we might shift the discussion pages to something more akin to LiquidThreads, which would negate the entire need for an automated signing system. But personally, I think automating it would cause far, far more trouble than its worth (and would take more collective time to fix false-positives than collective time spent typing four tildes). EVula // talk // ☯ // 05:47, 29 April 2009 (UTC)[reply]

yeah, that's a distinct possibility. good enough for me, unless someone else thinks it's a good idea and wants to chime in about it.

--Ludwigs2 06:08, 29 April 2009 (UTC)[reply]

I've been assigned LiquidThreads as my next task, so you can probably expect it over the next few months, probably towards the end of the year. — Werdna • talk 15:46, 1 May 2009 (UTC)[reply]

<pre-emptive champagne>Woohoo!!!</champagne>... :D Happy‑melon 11:06, 2 May 2009 (UTC)[reply]

catchpa on archive

Why do I have to enter the catcha 5 times in order to archive a talk page??? 76.66.202.139 (talk) 07:32, 29 April 2009 (UTC)[reply]

Might be helpful if we knew which page you were looking at. Sorry to hear about your troubles; while I do strongly support the ability of anons to edit, I also generally think it's in the best interest of any regular contributor to register an account to avoid various issues, including this sort of thing. – Luna Santin (talk) 21:39, 29 April 2009 (UTC)[reply]

This might be a stupid question but ... did you get the captcha correct the first 4 times? Mr.Z-man 23:43, 29 April 2009 (UTC)[reply]

I think the removal of large amount of text from a page triggers it for ip users. You can create an account to avoid this. Chillum 23:46, 29 April 2009 (UTC)[reply]

Sounds like you hit the external link CAPTCHA. (That's still activated, right?) --MZMcBride (talk) 22:37, 2 May 2009 (UTC)[reply]

Vulnerability

The problem has been raised (in a very roundabout way, unfortunately) that short pages may be especially vulnerable to defamatory attacks, since material added to them is likely to show up on Google search result pages, and continue to do so (because of the Google cache) even after it has been deleted from Wikipedia. I don't entirely see this myself, but if anyone can add any enlightenment on the problem, please comment at WT:WikiProject Disambiguation/Vulnerability to attacks.--Kotniski (talk) 10:54, 29 April 2009 (UTC)[reply]

Autoconfirmation

Resolved

Just now Wikipedian2 (talk · contribs) asked why Twinkle tells him that his account were "too new". His account however is far older and has far more edits than are normally required. His account is also far past the the $wgTorAutoConfirmAge and $wgTorAutoConfirmCount setting from TorBlock. There also is no entry for the user in the Special:AbuseLog, so he wasn't deautoconfirmed through that (would that be a permanent deautoconfirmation, BTW?).
So I'm at a loss. Is there another way to lose autoconfirmation status that I'm missing? The user was apparently recently caught in an autoblock, and was ip-block exempted. Can that have anything to do with it?
Amalthea 11:46, 29 April 2009 (UTC)[reply]

~~Any abuse filter triggers?~~ nope. –xeno ^talk 20:27, 29 April 2009 (UTC)[reply]

On my preferences - it says I am autoconfirmed - but apparently not? Wikipedian2 (talk) 15:48, 1 May 2009 (UTC)[reply]

That makes it a twinkle bug, not a MediaWiki bug. — Werdna • talk 16:37, 1 May 2009 (UTC)[reply]

Precisely why I went to twinkle bugs, where my case was closed and I didn't come here immedietly. Wikipedian2 (talk) 17:15, 1 May 2009 (UTC)[reply]

Hmm. I'd be very surprised if there was a bug in that trivial logic, but I'll try to work it from there. Sorry for the noise. Amalthea 10:26, 2 May 2009 (UTC)[reply]

Bug in file history

The page http://en.wikipedia.org/wiki/File:H1N1_map.svg shows a bug in the file history.

The url of the links to each version in the date/time column are out of sync with the displayed date/time and with the displayed images. e.g. the date 13:33, 29 April 2009 links to a file with 1353 in the url. The 13:05, 29 April 2009 date links to a file with 1333 in the url and so on down the list. Charvest (talk) 14:08, 29 April 2009 (UTC)[reply]

Those are actually linking to the correct files. The way we currently name archive file versions is a little odd: it's got the timestamp of when the file was replaced rather than its original upload date. --brion (talk) 16:11, 29 April 2009 (UTC)[reply]

That's an insane system. Thanks for the reply. Charvest (talk) 16:37, 29 April 2009 (UTC)[reply]

Yes, yes it is. :) --brion (talk) 14:35, 30 April 2009 (UTC)[reply]

Mysterious malfuctioning redirect

re: Epoch (geology) (edit talk links history)

Can someone figure out why the redirect http://en.wikipedia.org/wiki/Epoch_(geology) fails to forward one to the proper page... and fix it, please. // 24.62.190.234 (talk) 20:25, 29 April 2009 (UTC)[reply]

Works for me , what's it doing for you? –xeno ^talk 20:26, 29 April 2009 (UTC)[reply]

It redirects to Series (stratigraphy). If you consider that this is the wrong page, then you need to let us know what is proper. ---— Gadget850 (Ed) ^talk 20:40, 29 April 2009 (UTC)[reply]

Image toggling?

Does the English Wikipedia currently have a method for toggling between images? I'd like to add a button that toggles between two images of identical size for an article at FAC, the Euclidean algorithm. The idea is to allow the reader to shut off a GIF animation by replacing it with a still image of the same size. An FAC reviewer suggested that the animation, although helpful for understanding the algorithm, distracted them from reading the article. I do not want to convert the GIF into an OGG movie. The frames of the animation are available as SVGs. Thanks for your suggestions, Proteins (talk) 13:29, 30 April 2009 (UTC)[reply]

File:Daniel Roth.jpg

This is showing in Firefox but not in IE. I am also told it does not work in Safari. Why might this be? Stifle (talk) 13:43, 30 April 2009 (UTC)[reply]

Yeah, broken JPEG image (inverted b/w image). I've seen stuff like that before. I have no idea how you can fix it though. —TheDJ (talk • contribs) 15:19, 30 April 2009 (UTC)[reply]

More than 20 groups please

Hey, an editor has spoken to me of his difficulty in creating a template for the singles of Prince because the template does not allow anymore than 20 groups. Can this please be increased, it seems harmless enough. Thanks. — R ² 16:14, 30 April 2009 (UTC)[reply]

There's some info on this at Template talk:Navbox#More than 20 groups. --Amalthea 16:46, 30 April 2009 (UTC)[reply]

Sorting navbox columns?

Is there a way to sort navbox columns in a way similar to a sortable wikitable? There is a an ongoing debate regarding the table at Template:2009 US swine flu outbreak table, some people feel that it should be sorted alpha by state, while others argue to sort by number of cases. If this could be made sortable it would solve the problem.

I know some editors here were able to help out yesterday on the table for the main 2009 swine flu outbreak article, so I'm hoping someone can give us a hand on the US article as well. Thanks! Wine Guy Talk 17:05, 30 April 2009 (UTC)[reply]

Many thanks to Ludwigs2 for your help!! Wine Guy Talk 19:45, 30 April 2009 (UTC)[reply]

Comment requested

Comment is requested at Wikipedia:Bots/Requests for approval/ListasBot 5, regarding whether or not there should be a bot that makes non-visible changes to {{WikiProjectBanners}} and {{WikiProjectBannerShell}} templates. The fixes would be made so that the KingbotK plugin for AWB would be able to properly handle the pages. Matt (talk) 17:42, 30 April 2009 (UTC)[reply]

Unified login

This may have been asked before many times, so apologies if it has. Are there any plans to properly unify the login, so your edits are moved over & also, so if you want to, you can use the same userpage? Thanks. dottydotdot (talk) 19:59, 30 April 2009 (UTC)[reply]

What do you mean by "your edits are moved over"?? Happy‑melon 22:09, 30 April 2009 (UTC)[reply]

So that if you have edited Wikipedia say 500 times, then those edits are mirrored on a different site? dottydotdot (talk) 07:40, 1 May 2009 (UTC)[reply]

But why would they need to be mirrored? The content that was edited is not duplicated to that other site, unless it is imported. Happy‑melon 07:51, 1 May 2009 (UTC)[reply]

I do believe that some work has been done on creating global preferences. So that you would only need to set your signature and timezone once. —TheDJ (talk • contribs) 09:48, 1 May 2009 (UTC)[reply]

So that somehow, it would show that you had 500 edits on Wikipedia & you're not completely new. I was just wondering if there were any plans to make it a bit more "unified" but I can see the want to keep them separate. Cheers dottydotdot (talk) 10:16, 1 May 2009 (UTC)[reply]

Showing that you have made 500 edits when you haven't would be wrong. There are several tools people can use to see how new someone is more reliably than that anyways. — Mike.lifeguard | ^@en.wb 19:18, 2 May 2009 (UTC)[reply]

The proposal to mirror edits across projects is consistent with the notion of a "unified identity" which is quite different from the notion of "unified login". "Unified login" is a user interface convenience, but "unified identity" changes in basic ways how the editor role is implemented. --User:Ceyockey (talk to me) 11:30, 3 May 2009 (UTC)[reply]

Preview page expiring

I've been having problems with edit preview pages expiring under Internet Explorer 7.0.6001.18000. If I navigate away to check a detail and return IE says that the page is expired. Sometimes even it claims not to bable to find the page when I finish an edit. This problem, has only appeared in the last couple of days.--Peter cohen (talk) 10:40, 1 May 2009 (UTC)[reply]

Searching for an account with a suffix

Recently, Mammamia9905 has been blocked as a sockpuppet of Petergriffin9901. BaltoPat9902 was blocked earlier. Naturally, I'm interested in finding usernames that end in 9903 and 9904 so I can see what they've been up to. Can anyone think of a reasonable search method?—Kww(talk) 12:10, 1 May 2009 (UTC)[reply]

Not except for an exhaustive search. — Werdna • talk 13:28, 1 May 2009 (UTC)[reply]

Sure, easy enough with grep and a list of usernames. The problem of course is that there are about 9 million usernames (9,577,482 in fact) on en.wiki, so any search of them not that's not optimized is a bit nasty. For your specific query, the raw data (all usernames on en.wiki) is available here: tools:~mzmcbride/all-users-2009-05-01.txt.gz (37 MB). I grepped the file for ".*990[0-9]$":

Usernames that end in 990[0-9]

Err, have fun! :-) --MZMcBride (talk) 19:59, 1 May 2009 (UTC)[reply]

An embarrassment of riches, but not as bad as you might think. Relatively few are actually Name9902 and Name9903. It'll take a few hours, but I'll manage it some afternoon when I'm bored. Or maybe just submit a massive checkuser request to keep those guys on their toes.—Kww(talk) 22:05, 1 May 2009 (UTC)[reply]

Change tagging activated

Hi all,

After working out some minor performance issues, we've activated change tagging on Wikimedia. You can see the available tags, and filter by them at Special:Tags. Individual tags can be styled using the mw-tag-$tag class applied on changes lists (tag names are sanitised for this purpose).

Let me know if there are any problems! — Werdna • talk 13:33, 1 May 2009 (UTC)[reply]

Looks interesting but... how do you tag pages? Is there documentation for the extension/addition? –Drilnoth (T • C • L) 14:43, 1 May 2009 (UTC)[reply]

Tagging is done by the software. Presently, this is done mostly by the Abuse Filter – the actions taken by a filter can include applying a tag to the edit in question. — Werdna • talk 15:40, 1 May 2009 (UTC)[reply]

Okay; thanks. Sounds like another great feature! –Drilnoth (T • C • L) 15:49, 1 May 2009 (UTC)[reply]

Great feature indeed, though I'm a tad worried for the younger editors. The interface is getting more cluttered by the day.... —TheDJ (talk • contribs) 10:06, 2 May 2009 (UTC)[reply]

Any chance of wrapping tags in an identifiable class? I know that the whole li entry is given a class that identifies the tag (what happens if more than one tag is applied, BTW? Multiple classes?), but being able to style the "(references removed)" phrase itself would be useful. The same class should apply to all tags, as well, otherwise you have an open-ended set of classes to look for. Shall I open a bug? Happy‑melon 11:03, 2 May 2009 (UTC)[reply]

Yes, open a bug (if someone hasn't already beaten you to it). --MZMcBride (talk) 22:22, 2 May 2009 (UTC)[reply]

Template:Bug Happy‑melon 17:19, 3 May 2009 (UTC)[reply]

History tab at the top

This discussion has been closed. Please do not modify it.

The following discussion has been closed. Please do not modify it.

Precedent: Articles at the top have tabs with labels such as "discussion" and "history". The discussion tab leads to the talk page. The use of the word "discussion" is good because it is easy to understand and not jargon. Once someone clicks it, it shows the talk page (not named "discussion").

Suggestion: There is a "history" tab. Like the above, a history page is fine. However, the tab should not say "history" because this is jargon. Instead, it should say "authors".

Use of "authors" would give credit to editors, much as we give editors credit in the form of barnstars. It's just a nice thing to do to encourage writing, give recognition, and reduce jargon. User F203 (talk) 18:23, 1 May 2009 (UTC)[reply]

But that would be incorrect, implying that the page is a list of editors who have contributed to the current page text. But the history page also includes vandals, edits that were reverted, edits that have been totally overwritten, random drivebys by bots, protecting admins who have made no actual contribution to the content, etc etc. All authors are listed on the history page, but not everyone on the history page is an author. Happy‑melon 18:37, 1 May 2009 (UTC)[reply]

Every editor listed was an author is some form, even if what they wrote was reverted. But this is the wrong forum per instructions at the top. It also doesn't have to say "authors". It could say "editors". Even admins who protect it get some credit because they shaped the article by protecting it. Look at this suggestion as positive! User F203 (talk) 19:28, 1 May 2009 (UTC)[reply]

It's been called history for the longest time, I doubt we will change it now. Too ingrained in the gestalt. –xeno ^talk 19:31, 1 May 2009 (UTC)[reply]

I remember seeing another tab change about a year ago so change is possible. This is a positive change! Anyway, according to the instructions, this is the wrong forum so please end the discussion here. User F203 (talk) 19:37, 1 May 2009 (UTC)[reply]

Special:BlockList : hide proxy blocks

Would it be possible to hide blocks by ProccseeBot from the blocklist? –xeno ^talk 18:36, 1 May 2009 (UTC)[reply]

Some JavaScript could do it pretty easily (a modified version of User:MZMcBride/hideentries.js, I imagine). To do it in MediaWiki core would be a bit more tricky. There's no bot flag in the logging table. You'd have to create an option to exclude a specific user using a URL parameter or something. Not really clean. --MZMcBride (talk) 19:42, 1 May 2009 (UTC)[reply]

Hmmm... It's not too important to be honest, I just noticed that ProcseeBot can sometimes monopolize the log. =) –xeno ^talk 19:45, 1 May 2009 (UTC)[reply]

We've all been there. ;-) --MZMcBride (talk) 20:01, 1 May 2009 (UTC)[reply]

How long does it take the search engine to catch up with changes to articles?

^^^^ –xeno ^talk 19:25, 1 May 2009 (UTC)[reply]

According to rainman when I just asked him on IRC, he said a day. --MZMcBride (talk) 19:44, 1 May 2009 (UTC)[reply]

thanks =) –xeno ^talk 19:45, 1 May 2009 (UTC)[reply]

Interface redesign mock-ups

Trevor Parscal has been creating mock-ups of a redesigned, more efficient interface. Some of the preliminary sketches are viewable here. Comments, suggestions, etc. welcome here or on the talk page there. Cheers. --MZMcBride (talk) 19:39, 1 May 2009 (UTC)[reply]

I'm a fan of those mock-ups. I personally think monobook is overused and outdated, and I think it would do good to give it a little face-lift that will improve usability as well as aesthetics. cmelbye (^t/_c) 04:53, 2 May 2009 (UTC)[reply]

It looks like an improvement for newer users: it's simple. I like the aesthetic, but I'd probably customize it for myself: first, I hate drop-down menus (which, by the way, often fail on touch-only devices like the iPod Touch or iPhone, which don't have "hover" options), and second, I'm used to the general scheme of our current interface. It might be improved by a little tweaking using things we already have: for example, there should probably be a way to tell which tabs are active, like the orange border/bolding currently used through the "active" CSS class. Aside from that, my concern would be mainly in the implementation side of things: what technical features would it use to achieve the design features? Would these technical features degrade gracefully? Would they allow the sort of customization I'd desire? Would old customizations (user scripts) be broken? {{Nihiltres|talk|log}} 06:45, 2 May 2009 (UTC)[reply]

I don't think there is any getting around using dropdown menu's in the end. We just have too many options per page. —TheDJ (talk • contribs) 11:14, 2 May 2009 (UTC)[reply]

The way to get around using drop-down menus is not to use them. Pretend they don't exist (which may be the case for some users). The only place they are used now in the edit page is for inserting characters. The mockups are nice, but they need to work for the vast majority of users if they're going to be the default. Delicious carbuncle (talk) 16:43, 2 May 2009 (UTC)[reply]

My point is rather that we need to make sure that those menus are effective. If a redesign means, for example, that everyone browsing on an iPhone or tablet (which can't hover on elements) can't access the menus, we have a problem; if a redesign means that a pile of user scripts break really nastily, we have a problem; if a redesign prevents users from customizing the interface at all … you see what I mean? My concern is that, although the redesign might help quite a number of viewers, we don't want to break things for the rest of them, myself included. I'd love to see a live mock-up. {{Nihiltres|talk|log}} 20:16, 2 May 2009 (UTC)[reply]

urls containing ]

how do you include an url which contains [] as a reference? It seems to break {{cite web}} as well as a direct external link. dramatic (talk) 04:42, 2 May 2009 (UTC)[reply]

Either percent-encode the '[' and ']' characters, using '%5b' and '%5d' respectively, or use the magic word {{urlencode:[blah]}} . The latter might be a problem if "blah" contains certain special characters, such as "/", so it doesn't seem to be suitable for full urls. I've changed Russell Ward for you.-gadfium 06:18, 2 May 2009 (UTC)[reply]

Tangentially: No one answered my question earlier about whether it could be possible to allow percent-encoded square brackets in wiki-links, and hence in article URLs, like for example [[Benzo%5ba%5dpyrene]] ? --83.253.251.229 (talk) 17:25, 2 May 2009 (UTC)[reply]

[, |, and ] will never be allowed in page titles, ever. You can put them in the link display text, though: Foo[bar]baz|boop --MZMcBride (talk) 00:16, 3 May 2009 (UTC)[reply]

That's what I expected. Thanks for the answer, which I take to be authoritative until someone says otherwise ;-) --83.253.251.229 (talk) 18:04, 3 May 2009 (UTC)[reply]

References removed automatically appended to edit summaries

See the editing history of Duran Duran. Just curious, since when did MediaWiki do this? Someguy1221 (talk) 04:50, 2 May 2009 (UTC)[reply]

Since Special:Tags were enabled (see a few sections up). --MZMcBride (talk) 07:09, 2 May 2009 (UTC)[reply]

Can someone point to more information on these tags, how they are identified, and what they mean? Thanks. Delicious carbuncle (talk) 16:45, 2 May 2009 (UTC)[reply]

They're coming from AbuseFilter. For example, "references removed" comes from filter 61. Don't know if there's an easy way to see which tags come from which filter, though.... --MZMcBride (talk) 22:20, 2 May 2009 (UTC)[reply]

Ah! Thanks. I'll repost my comment from WP:AN - An obvious measure to reduce confusion would be to link the "tag" back to the filter which appends it. Was this change announced anywhere (other than the extremely vague note at WP:VPT)? It doesn't seem to be covered on WP:AF, unless I missed it. Expect more confusion. Delicious carbuncle (talk) 18:38, 3 May 2009 (UTC)[reply]

Agreed. I keep a pretty close eye on software changes and I was still "WTF?" when I saw it the first time. It needs a link or something. Or a tooltip or a CSS mark of some sort. --MZMcBride (talk) 18:40, 3 May 2009 (UTC)[reply]

I've taken it to Wikipedia talk:Abuse filter#Filter ID for us noobs? after being caught unawares; seems like a good place to have that discussion. 9Nak (talk) 19:11, 3 May 2009 (UTC)[reply]

Possible for bot to take wikipedia edit article history information to notify editors of AfDs?

Is it technically possible for a bot to take information off of wikipedia (or the below two webpages) and then contact editors with x amount of edits about an Article for deletion?:

What bot expert can I contact about this? Thank you. Ikip (talk) 14:37, 2 May 2009 (UTC)[reply]

Try WP:BOTREQ. --MZMcBride (talk) 22:18, 2 May 2009 (UTC)[reply]

Thanks again McBride. Ikip (talk) 00:10, 3 May 2009 (UTC)[reply]

Navbox issue

Does anyone know how to make the "v/d/e" links disappear from the navbox here? I don't see what's up with the template code at the moment. — Carl (CBM · talk) 19:36, 2 May 2009 (UTC)[reply]

An error has been made recently, which broke the paramater because it was renamed to Navbar instead of navbar. —TheDJ (talk • contribs) 19:48, 2 May 2009 (UTC)[reply]

Done issue fixed. —TheDJ (talk • contribs) 19:52, 2 May 2009 (UTC)[reply]

Apologies, I mixed up the Navbar template with the navbar parameter. — Edokter • Talk • 20:46, 2 May 2009 (UTC)[reply]

Cosmetic changes (Wikitext cleanup options of pywikipediabot)

On my talk page someone disagrees about some specific issues involved in the cosmetic changes-script (which is applied when my bot adds interwiki's). Because I am mostly active on nl.wikipedia, I am not sure if there is any policy on these topics on en.wikipedia. Can anyone help me out on this issue? Thanks in advance. Kind regards, --Maurits (talk) 21:59, 2 May 2009 (UTC)[reply]

Just speaking personally, I'm not a fan of extending "WP:" to "Wikipedia:", editors usually choose one or the other for a reason, even in the piped portion. I'm not sure what our current mind on changing Image: to File: when making other edits is. –xeno ^talk 22:07, 2 May 2009 (UTC)[reply]

In recent days, I examined several changes the script can do, I think the following settings are more or less safe to use on en.wikipedia and could probably be recommended for interwiki bots as default settings for English wikipedia:

        text = self.fixSelfInterwiki(text)
        text = self.standardizeInterwiki(text)
        #text = self.standardizeCategories(text)
        text = self.cleanUpLinks(text)
        text = self.standardizeheadernames(text)  # not in script
        text = self.removeunicodecontrols(text)   # not in script
        #text = self.cleanUpSectionHeaders(text)
        # #### text = self.putSpacesInLists(text)                 # already de-activated by default
        text = self.translateAndCapitalizeNamespaces(text)
        #text = self.removeDeprecatedTemplates(text)
        text = self.resolveHtmlEntities(text)
        #text = self.validXhtml(text)
        #text = self.removeUselessSpaces(text)
        #text = self.removeNonBreakingSpaceBeforePercent(text)
        #try:
        #    text = isbn.hyphenateIsbnNumbers(text)
        #except isbn.InvalidIsbnException, error:
        #    pass

(updated 01:01, 3 May 2009 (UTC))

It would be nice if there was way to let it apply the general fixes from AWB (see Wikipedia talk:AutoWikiBrowser#AWB general fixes module for python). I don't think "translateAndCapitalizeNamespaces" should be a problem in article namespace, as the only one it's likely to be applied is "Category:". -- User:Docu

The complaint likely resultd from this edit to the Wikipedia: space. –xeno ^talk 00:28, 3 May 2009 (UTC)[reply]

Given that sample, the option is probably more useful for languages with localized namespace names. It should probably be turned off too. -- User:Docu

The changes made are minor and, I agree, for some part quite irrelevant. However, I only apply them to pages when interwiki's are changes. The main problem, I think, was on the file: vs. image: question. From the discussion on Wikipedia:ANI#File: vs. Image: I conclude that Image: has indeed been deprecated by File: and therefore, the changes were correct.

For some part, I like the "safe version" proposed bij Docu. However, the hyphenation of IBSN-numbers -for example- are useful in my opinion, and that goes fore most options. Because the bot works on various languages, some changes might be less relevant on the en.wikipedia indeed. Switching them off would however switch them off for all languages. Note that the qualification is "more useful" against "less useful" and not "not useful" or something worse. Thanks for your reactions. Kind regards, --Maurits (talk) 15:16, 3 May 2009 (UTC)[reply]

For the namespace part, I think when it was written, shortcuts like "WP:" were redirects and not a $wgNamespaceAliases. As for the file/image question, I rather not enter it. Ideally one could just use just the "Capitalize" part. As for the ISBN settings, I'd have to look into it in detail.

It shouldn't be too complicated to modify the script that it doesn't apply all settings to all languages (e.g. there is already some code for nl and de). -- User:Docu

I'll check out the possibilities of "nationalization" (although we should be careful in applying it), indeed I noticed some nl-code at the bottom of cosmetic_changes.py. I excluded 'wp' and 'wikipedia' from translateAndCapitalizeNamespaces to solve the actual issue. Kind regards, --Maurits (talk) 17:14, 3 May 2009 (UTC)[reply]

My solution didn't work, perhaps someone with better Python skills can help me out. There is already a line that excludes main namespace:

           # skip main (article) namespace
           if thisNs and namespaces:
               text = wikipedia.replaceExcept(text, r'\[\[\s*(' + '|'.join(namespaces) + ') *:(?P<nameAndLabel>.*?)\]\]', r'[[' + thisNs + ':\g<nameAndLabel>]]', exceptions)

(the whole is like this):

   def translateAndCapitalizeNamespaces(self, text):
       """
       Makes sure that localized namespace names are used.
       """
       family = self.site.family
       # wiki links aren't parsed here.
       exceptions = ['nowiki', 'comment', 'math', 'pre']

       for nsNumber in family.namespaces:
           if not family.isDefinedNSLanguage(nsNumber, self.site.lang):
               # Skip undefined namespaces
               continue
           namespaces = list(family.namespace(self.site.lang, nsNumber, all = True))
           thisNs = namespaces.pop(0)

           # skip main (article) namespace
           if thisNs and namespaces:
               text = wikipedia.replaceExcept(text, r'\[\[\s*(' + '|'.join(namespaces) + ') *:(?P<nameAndLabel>.*?)\]\]', r'[[' + thisNs + ':\g<nameAndLabel>]]', exceptions)
       return text

Probably a variant of the main namespace exclusion is needed to exclude wikipedia namespace. I tried to list it as an exception, but the script didn't accept that. Thanks in advance, --Maurits (talk) 17:31, 3 May 2009 (UTC)[reply]

You could change the line above to if thisNs and namespaces and thisNS != "Wikipedia": in the sample above, but then you still have to deal with the "WT:" pseudo-namespace. At "Wikipedia:Namespace#Case_insensitivity", I listed a series of titles for the same pages. You can use it to test which ones get normalized to "Wikipedia" ('WP', 'Project', 'wikipedia', 'wP', 'project') and which ones don't. As this isn't ideal, I'd skip the function. -- User:Docu

Thanks for your advice. As there shouldn't be any links to talk pages in articles, I think that this solution will do. If it doesn't, I will skip it when appropriate. Kind regards, --Maurits (talk) 21:39, 3 May 2009 (UTC)[reply]

question about parserfunctions, nesting, and processor loads

I made a template - {{do list}} - to standardize text lists and keep line breaks from occurring in the middle of phrases, but another editor (User:Edokter) suggested it might be too heavy on the preprocessor. The issue boils down to whether it's more efficient to use nested conditionals (so that unused IFs never get evaluated at all) or to use sequential conditionals (in which all the IFs have to be evaluated, but nothing gets stacked in memory). eg, which of the following is better?

{{#if:{{{1|}}}
    |{{{1}}} {{#if:{{{2|}}}
        |{{{2}}} {{#if:{{{3|}}}
            |{{{3}}}... ... ... 
        }}
    }}
}}

{{#if:{{{1|}}}|{{{1}}} }}
{{#if:{{{2|}}}|{{{2}}} }}
{{#if:{{{3|}}}|{{{3}}} }} 
... ... ...

or (alternately) should I just scrap this template as being too expensive either way you look at it? --Ludwigs2 15:02, 3 May 2009 (UTC)[reply]

Neither option would be dangerously heavy on the preprocessor in the quantities used in that template, IMO. Nested conditionals get very difficult to read in line-wrapping editors, and it becomes easier to get brace counts wrong. In principle, yes, nested conditionals are fractionally more efficient (the preprocessor isn't building a stack, it's recursively expanding objects inline, so the memory usage is probably comparable), but not overwhelmingly so. Happy‑melon 17:07, 3 May 2009 (UTC)[reply]

Problem with the infobox

Hi, I'm contributing to the article, Saidabad. There is a problem with the infobox formatting. I'm not able to fix it. Any help is much appreciated. Randhir 18:48, 3 May 2009 (UTC)

The Zone and Ward variables take the value entered for Metro and wrap it like "List of <metro> corporation wards" for the template. Thus you can't wikilink the metro entry as it will break the template. Nanonic (talk) 19:06, 3 May 2009 (UTC)[reply]

Valid change tags

If you go to "Special:SpecialPages", there is a cryptic link to "Valid change tags", which can also be reached from the Special:Contributions page. One of the more unusual "valid change tags" is "nonsense movies". Not "nonesense books", "nonsense article", "nonsense Pokemon", but only "nonsense movies". Should that be instead, "nonsense article", or "nonsense moves"? Excuse me, it is "Nonsense movies?". What's up? Or should I say, "What's up, Doc?" 199.125.109.77 (talk) 23:23, 3 May 2009 (UTC)[reply]