Jump to content

Wikipedia:Village pump (technical)

From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by 76.65.131.217 (talk) at 06:40, 19 October 2013 (→We need a watch list for categorization). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

 Policy Technical Proposals Idea lab WMF Miscellaneous 
Shortcuts
The technical section of the village pump is used to discuss technical issues about Wikipedia. Bugs and feature requests should be made at Bugzilla (see how to report a bug). Bugs with security implications should be reported to security@wikimedia.org or filed under the "Security" product in Bugzilla.

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.


VE avoided by 85% of new usernames

I want to note how, in the final days on the top menu, the VisualEditor was avoided by 85% of the new usernames being tracked for choice of text editor. See: wp:VEDASH for the VE dashboard graph which showed the average low 15% usage among new usernames, who mainly preferred to use the wikitext source editor for 85% of edits, even though VE was still on the edit-tabs at that time. Also see below: "#VE opt-in usage near 0%". -Wikid77 15:57, 9 October 2013 (UTC)[reply]

VE opt-in usage near 0%

After VE was removed from WP's top menu on 24 September 2013, to become an easy opt-in feature in Special:Preferences, then the VisualEditor was avoided by 99.7% of users being tracked for choice of text editor. See: wp:VEDASH for the VE dashboard graph which showed the average usage (after 25 September 2013) remained well below 1% of all edits, often ranking as 1⁄5% (0.002, or 2 edits per thousand). As many experienced software developers have emphasized: WYSIWYG interfaces can be very tedious to use, and many power users quickly switch to text-based editing of pages, as faster to perform the work at hand. That is why we computer scientists developed hypertext markup languages, as copy/paste text languages, to allow diff-links between revisions, with new features by a macro scripting language (for templates), and to also allow multi-word search in markup keywords (although most browsers still "find string" rather than "hunt words" in multiple spots). It can be much faster to keep wp:checklists of intended text changes, to focus on each step to edit, and then re-proofread the final page to checkmark each step as successfully done. By comparison, point-and-click steps are not obvious in a diff listing. Also see above: "#VE avoided by 85% of new usernames". The actual usage levels of VE needed to be emphasized, for future consideration. -Wikid77 15:57, 9 October 2013 (UTC)[reply]

Most of those 99.7% of users still don't know that they have to manually go to their preferences and re-enable VisualEditor. It's not a "choice" if you don't know that the option exists. From the users' perspective (not counting the <1% of people involved in Kww's default-state RFC), VisualEditor just silently disappeared two weeks ago, with no indication that they are even allowed to opt-in if they wanted to. Whatamidoing (WMF) (talk) 17:43, 10 October 2013 (UTC)[reply]
I'm forced to agree that this isn't a very telling statistic, and describing it as people "avoiding" VE is a gross misinterpretation. Although on the flip side, if VE had been remotely popular, we would still be seeing more opt-ins as people went looking for preferences, and we'd also be seeing some comments from people wondering where it went. Unless I'm not looking in the right places, I've seen no mention of it thus far. equazcion� 17:50, 10 Oct 2013 (UTC)
We've seen a handful of comments like this one, which appeared within hours at the Mediawiki feedback page. Not everyone asks their technical questions here, and of course if your question has just been asked and answered on the same page, then most people won't ask it again. Whatamidoing (WMF) (talk) 18:34, 10 October 2013 (UTC)[reply]
Well, I didn't only look here (as in this page), but on help desk and teahouse as well, but thanks for pointing out the VE feedback forum at MediaWiki. Although I'm not sure who would know to look there if they merely noticed VE had disappeared and weren't previously involved in testing/feedback in some way. People who just liked using it would probably be asking where it went in some Wikipedia venue, I would think. equazcion� 19:07, 10 Oct 2013 (UTC)
Whatamidoing, is it going to be WMF's official policy to portray the RFC and the subsequent change in VE's status as something that did not represent community consensus? I really am beginning to find these efforts to portray our RFC process as a problem tedious. If the WMF can come up with a more accurate way to sample consensus, I'm eager to listen.—Kww(talk) 00:43, 11 October 2013 (UTC)[reply]
I was planning to ignore this (both because I'm tired of complaining about Whatamidoing's communications, and because I wanted to try and keep this discussion above the level to which Whatamidoing would like to bring it down), but since you mention it, I find the little backhanded snipe here ("not counting the <1% of people involved in Kww's default-state RFC") childish and disgusting. VE must be very close to her heart for her to feel hurt enough to lash out as she's been doing, and I try to sympathize, but I think everyone feverish about this at the WMF could stand to take half a chill before they say more of the wrong things to more people here. equazcion� 02:13, 11 Oct 2013 (UTC)
Whether the RFC process' oversampling of experienced editors and metapedians (i.e., people like you and me) is a "problem" depends on the question you're asking. I can't think of a better system for figuring out how a complex policy should be applied to an article: someone on the WP:TOP5000 list is much more likely to be able to deal with a NOR or NPOV issue than someone who has only made one tiny edit.
However, the RFC process is pretty obviously not an effective method of getting responses from brand-new editors or from prospective editors. So I think the answer to your question is, who's in your community? The RFC process is an excellent method of determining the views of the highly active editors (the three or four thousand people like you and me who make more than a hundred edits each month). If that's "the community", then you very likely have community consensus represented on that page. If your idea of "the community" includes the tens of thousands of editors who made just five or ten edits in a month, then that RFC does not seem to include their views (which might or might not agree with the RFC's outcome; nobody really knows).
Equazcion, I'm not sniping about the RFC for being too small; it seems to be about the third largest ever. I'm only assuming that everyone who participated in the RFC, and later noticed that VisualEditor disappeared, is smart enough to make the connection between the two. So 99.7% of users haven't opted in, and since the number of people opting in (~0.3%) is smaller than the number of people who participated in the RFC (~0.9%), we can assume that at least some of those 99.7% know exactly what happened (and also any others who missed the RFC but saw the VPT and AN discussions). But that's still a vast majority of editors, including tens of thousands of new editors, who just don't know, and therefore can't be "choosing" in either direction. Whatamidoing (WMF) (talk) 18:26, 14 October 2013 (UTC)[reply]
Perhaps those "tens of thousands of new editors" were never given a link to WP:How to edit a page? Which in its very first sentence mentions VE, and in its second sentence describes how to chose VE. That any number of new editors opt-in shows that they can choose; that they don't opt-in suggests that they have chosen. ~ J. Johnson (JJ) (talk) 22:53, 15 October 2013 (UTC)[reply]

Proposal to Reduce the API limits to 1 edit/30 sec. for logged out users

We seem to be getting attacked by SpamBots a lot recently, or bots inadvertently get logged out during it's runs. Or we have incidents like User:RotlinkBot editing from IP farms, that can't be range blocked. Either way, legitimate bots shouldn't be editing from IPs and the SpamBots tend to come from IPs. I propose the API limits for editing while logged out should be set to 1 edit/30 sec. That way, the potential damage is manageable. Please note that the API is different from editing Wikipedia directly. It will not effect the IP editors on Wikipedia. It will only effect automated tasks, aka bots, that are using an IP instead of a username. Any input on this?—cyberpower ChatOnline 21:20, 2 October 2013 (UTC)[reply]

Yes... if an RfC is to be held here, I shall unwatch this page. --Redrose64 (talk) 21:41, 2 October 2013 (UTC)[reply]

One per minute seems a little too strict; maybe one per thirty seconds? -- Ypnypn (talk) 21:43, 2 October 2013 (UTC)[reply]

Ok. I have changed it to 30 seconds.—cyberpower ChatOnline 23:36, 2 October 2013 (UTC)[reply]
  • Support one per thirty seconds, which is about as fast as a human could reasonably edit. Robert McClenon (talk) 21:58, 2 October 2013 (UTC)[reply]
  • Comment only a bot could reasonably edit faster than that, and a bot that isn't logged in is a bot that is malfunctioning. The User:RotlinkBot issue is a special case that may involve malice or malware. Robert McClenon (talk) 22:00, 2 October 2013 (UTC)[reply]
  • Question Tools such as WP:TW use mw:API for editing pages. For example, if you wish to nominate a file for deletion, WP:TW makes three API edits within a short period of time. Are IP editors able to use these tools somehow? --Stefan2 (talk) 22:27, 2 October 2013 (UTC)[reply]
    No. IP editors can't use Twinkle. Twinkle will not be phased by this change.—cyberpower ChatOnline 23:34, 2 October 2013 (UTC)[reply]
    An IP editor could use Twinkle by using it with something such as Greasemonkey. I don't know if any of our long-term IP editors do so, but it's certainly possible. Anomie⚔ 01:31, 3 October 2013 (UTC)[reply]
    Thinking about Stefan2's question, we probably want to extend the notice to Twinkle users, WP:AFCH users and more to give them a heads up that this is coming Hasteur (talk) 23:59, 2 October 2013 (UTC)[reply]
  • Oppose Way too short. It's easy to see that last spelling error just as you hit Save page ... to make IPs wait 30 seconds is inappropriate. What's the current limit? NE Ent 00:05, 3 October 2013 (UTC)[reply]
    The API is different from editing Wikipedia directly. The API is used for bots and external programs. They can edit normally on Wikipedia itself. — Preceding unsigned comment added by Cyberpower678 (talk • contribs) 00:10, 3 October 2013 (UTC)[reply]
    Oh, that's very different... never mind (see Emily Litella if you're too young to understand) (of course getting a contributor who don't know what they're talking about is what you get for forum shopping to WP:AN) NE Ent 01:13, 3 October 2013 (UTC)[reply]
    If it's true this affect VE editing, as stated below, my original opposition was correct. NE Ent 18:35, 14 October 2013 (UTC)[reply]
  • Comment Can anyone provide me examples of why we allow anonymous writes via the API at all?—Kww(talk) 00:22, 3 October 2013 (UTC)[reply]
    I can't. Hence this proposal, but their may be external editing programs that might use the API, so that would be one example, but other than that, IP edits shouldn't be happening through the API.—cyberpower ChatOnline 00:26, 3 October 2013 (UTC)[reply]
    WP:Teahouse's scripts are usable by anons and use the API to make edits. Anomie⚔ 01:35, 3 October 2013 (UTC)[reply]
    Why do we let anonymous users edit at all? Oh yeah, its a foundation principle. Whether people edit via the API or HTML interface, it should make no difference. What matters is the substance of the edit. Legoktm (talk) 02:26, 3 October 2013 (UTC)[reply]
  • Support disabling entire write API for anons unless someone can think of a good reason why they need it. -- King of ♥ ♦ ♣ ♠ 00:35, 3 October 2013 (UTC)[reply]
    Note that things such as submitting AFTv5 feedback or using the VisualEditor are included in the write API. Anomie⚔ 01:39, 3 October 2013 (UTC)[reply]
    Hm. Would an IP need to make multiple VisualEditor edits within a short period of time? --Stefan2 (talk) 12:01, 3 October 2013 (UTC)[reply]
  • Support Yeah, with King on that. I can't see any reason to allow anonymous API edits at all. I had no idea that was currently allowed. Spammers will probably still find ways, but there's no reason to make it this easy for them. But short of disallowing anon API edits altogether, yes, a rate limit should be imposed. equazcion� | 00:52, 3 Oct 2013 (UTC)
  • Support If I understand mw:API:Main page correctly, this shouldn't affect real people in any way (I don't understand how you would even be able to make an API edit at all as a human), but it should be able to slow down botspam, so it's seemingly a good first step. I'd suggest that we investigate Equazcion's ban-API-edits-entirely proposal. Nyttend (talk) 01:14, 3 October 2013 (UTC)[reply]
    I wouldn't recommend that. There may be legitimate reasons to have to IP edit through the API, but since it's almost never going to happen, limiting it to 1 sounds like the reasonable first step.—cyberpower ChatOnline 01:21, 3 October 2013 (UTC)[reply]
    You can use Special:ApiSandbox, make a POST request through your browser, etc. Legoktm (talk) 02:29, 3 October 2013 (UTC)[reply]
  • Comment So far, this proposal is long on rhetoric and low on facts. There is an assertion that spam bots are using the API while logged out. Links? What's to stop the spambot from screen-scraping the UI edit form, which many probably already do? There is talk of User:RotlinkBot, but that's a registered account and if it continued its unapproved actions after being blocked I don't see any links showing that either. And if it really had an "IP farm", couldn't it cycle through the farm to make N edits every 30 seconds (one per IP)? And what's to stop a spambot from spamming from various registered accounts until the checkusers block its IP? Anomie⚔ 01:31, 3 October 2013 (UTC)[reply]
    • As I said, spam bots would probably manage to continue making spam edits, but currently we're almost encouraging it by making it remarkably easy. I don't see much if any legitimate reason to allow anonymous API edits whatsoever. For the RotlinkBot history, see Wikipedia:Archive.is RFC. equazcion� | 01:44, 3 Oct 2013 (UTC)
    • I just noticed your mention above that the article feedback tool uses the API. If disallowing anonymous API edits would disable article feedback for anonymous users, that would indeed be a problem. A 30-second rate limit shouldn't interfere with that though. equazcion� | 01:55, 3 Oct 2013 (UTC)
  • @Cyberpower678: It appears there are a number of people who are not familiar with the API. Perhaps you could add a statement at the top of the RfC similar to this in order to help people understand the proposal better.

The API (or Application Programming Interface) is the way software interacts with Wikipedia. The API is for bots and automated tools, not for humans. Click here to see the API.

Briefly explaining the API at the top of the RfC might save a lot of time for those who are not familiar with it. FWIW, I think 1 edit per minute would be fine for IP editors. Best. 64.40.54.196 (talk) 02:20, 3 October 2013 (UTC)[reply]
  • No. Ratelimiting API edits is not the right way to fight spam. If you do that, they'll just screenscrape. Bots logging out is a failure of the bot to use assert=edit properly. Legoktm (talk) 02:23, 3 October 2013 (UTC)[reply]
    Screenscraping is more difficult than the API. It's at least one step to fight spambots.—cyberpower ChatOnline 02:29, 3 October 2013 (UTC)[reply]
    Not really. It would take about 5 minutes to write an edit function that screenscrapes. Really it would take no time at all, Pywikibot-compat ships with one. Legoktm (talk) 03:26, 3 October 2013 (UTC)[reply]
    Yeah I'm not really getting Lego's logic there. The API is built for automated edits, so that legit automated tools don't need to screen scrape. Why provide the same ease to likely illegitimate ones? Better to make it harder for them, at least. If they do resort to screen scraping, that'll still end up being slower than API edits, and will be far less reliable at making successful edits. equazcion� | 02:36, 3 Oct 2013 (UTC)
    You just said that any anonymous edit is an illegitimate edit. Are you sure you meant that? Legoktm (talk) 03:26, 3 October 2013 (UTC)[reply]
    Are you sure I said that? equazcion� | 03:47, 3 Oct 2013 (UTC)
    Gr, I read too fast, sorry. Regardless, do you have any evidence that proves that anonymous edits via the API are likely to be bad? Legoktm (talk) 18:54, 3 October 2013 (UTC)[reply]
  • Comment: I don't think this is the right way to go about this, but if it does happen, I think a limit of 10 per 5 minutes (or some other multiple of 1 per 30 seconds) should be used instead, to allow an occasional short burst. Jackmcbarn (talk) 02:31, 3 October 2013 (UTC)[reply]
Thanks, Werieth and cyberpower. This what I was asking about since before I registered this account, I edited for years as an unregistered IP account. I guess my follow-up question would be what does it mean to edit through the API and not the regular Wikipedia editor...but I have the feeling that will be a technical answer and I'm guessing the bottom line is that regular editors don't edit through the API. I guess you can confirm or deny this? Thanks again! Liz Read! Talk! 17:01, 3 October 2013 (UTC)[reply]
@Liz: It means using anything else than the default edit form for editing, be it Twinkle, VisualEditor, Popups, or any other from the dozens of tools which can edit pages on your behalf. The only notable exception I can think of is wikEd, which merely enhances the default form. Matma Rex talk 17:17, 3 October 2013 (UTC)[reply]
Well, Matma Rex, I think you need to have a registered account to use Twinkle but not for VisualEditor. Since VE was intended to make Wikipedia more "user-friendly" for casual users who might use an IP account, this sounds like it could negatively affect them. I know that, unfortunately, sometimes I make an edit and immediately notice a typo and need to correct it. Thirty seconds doesn't sound long but when you think of it as two edits per minute, it does slow you down when you're making small edits.
I wish there was a way to distinguish unregistered accounts from spambots. How about only 1 edit every 5 seconds? Liz Read! Talk! 17:47, 3 October 2013 (UTC)[reply]
  • So the fact that this will effectively block all IP editors (who are forced to use VE which edits via the API on their behalf according to the core developer above) emphasizes that this proposal is a request to require all IP editors to create an account to edit on Wikipedia against the spirit of pillar 3. Technical 13 (talk) 18:02, 3 October 2013 (UTC)[reply]
  • I meant 5P#3 and have corrected it above. Technical 13 (talk) 15:51, 3 October 2013 (UTC)[reply]
    The same argument as in my first reply about "Wikipedia is free content" still applies. The "free" in free content refers to the reader's right to modify, redistribute, sell our content under various conditions (e.g., you need to give attribution, at least an URL). It does not refer to everyone's right to edit the actual content on Wikimedia sites, but if they disagree they can fork. Using that logic, you could also say that blocking users goes against our principles, as they can no longer edit, or that open proxies should be able to edit. Besides, (and this is the main point) they are not proposing to completely eliminate IP edits, but to throttle anonymous edits through the API, which bots and scripts (notably VE) use. I hope you can understand the different between distributing modified or original copies with attribution and actually editing and submitting the changes to the original source. If you want to argue the case for your opinion using basic Wikimedia principles, I would personally go with founding principle #2. Maybe I'm wrong though, in which case I would love to know the original meaning of editing in "Wikipedia is free content". πr2 (t • c) 18:07, 3 October 2013 (UTC)[reply]
  • WP:5P#3 specifically says, "Wikipedia is free content that anyone can edit, use, modify, and distribute". Setting this ratelimit would prevent IPs from editing Wikipedia (except to use it as a social networking site because they will only be able to freely edit talk spaces). Technical 13 (talk) 19:00, 3 October 2013 (UTC)[reply]
    • T13, don't you think that's just a little too much hyperbole? This doesn't prevent IP addresses from editing articles (and in fact, I'm not sure where you're seeing a connection to talk pages at all), except for a few edge cases. I think those edge cases, combined with the very small benefit, makes this proposal not worth it, but let's not go around saying things like outright preventing IPs from editing Wikipedia. Writ Keeper ⚇♔ 19:22, 3 October 2013 (UTC)[reply]
  • Writ Keeper, if IPs edit with VisualEditor, and VE edits via the API, and the API is ratelimited to only allow 1 edit every 30-60 seconds, then that prevents IP editors from editing in a convenient manner being able to catch and quickly fix their typos or whatnots as is exampled below. If IPs can't edit conveniently, then they are forced to register and account or not edit. As an example use case, there are 0 Category:G13 eligible AfC submissions waiting to be assessed. The way that I personally review these is to open 5-10-20 of them in separate tabs and let them load. As soon as the first one is loaded, I zip though all of them and click "review" to open the AFCH script. Once I get to the end of the list doing that, I zip though all of them again clicking "tag for G13" for all of the ones that are blank or obviously nothing. Doing this takes about 5 seconds to tag 10-15 pages as G13. If I wasn't logged in, I wouldn't be able to review G13 eligible drafts because I would be locked out by this proposal. Now, I admit this is an edge case, but there are lots of various edge cases like this. Technical 13 (talk) 19:35, 3 October 2013 (UTC)[reply]
    • No you wouldn't, because IPs don't currently use VE. They might in the future, in which case this proposal would be much more significant (though it still wouldn't *absolutely* deny them, which is what you said), and there are other things that IPs would use the API for legitimately, but right now, *any* use of the API by a logged-out editor remains an edge case, not a main case. This proposal is not unreasonable. Writ Keeper ⚇♔ 19:43, 3 October 2013 (UTC)[reply]
  • Support. This is highly unlikely to stifle any good edits, but more than likely to stifle many bad ones. bd2412 T 16:06, 3 October 2013 (UTC)[reply]
  • Oppose. Not a bad idea on paper, but the possibility of limiting good edits through things like VE, the Teahouse scripts, etc. is not worth making things very slightly less easy for spambots. Writ Keeper ⚇♔ 18:03, 3 October 2013 (UTC)[reply]
  • Oppose if VE edits through the API, as stated above. I do not use VE, but I often make multiple edits per minute, manually, not using any bots or tools. Here is an example of 11 edits in 5 minutes from my contribution history, which contains many such examples:
    • 22:46, 2013 September 27 (diff | hist) . . (+8)‎ . . m WRGC (AM) ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:46, 2013 September 27 (diff | hist) . . (+6)‎ . . m WPVM-LP ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:45, 2013 September 27 (diff | hist) . . (+10)‎ . . m WPHY-CD ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:45, 2013 September 27 (diff | hist) . . (+10)‎ . . m WPGC-FM ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:44, 2013 September 27 (diff | hist) . . (+6)‎ . . m WLBT ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:44, 2013 September 27 (diff | hist) . . (+5)‎ . . m WLAB ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:44, 2013 September 27 (diff | hist) . . (-2)‎ . . m WKND (album) ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:43, 2013 September 27 (diff | hist) . . (-35)‎ . . m WET Web Tester ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:43, 2013 September 27 (diff | hist) . . (+6)‎ . . m WDBD ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:42, 2013 September 27 (diff | hist) . . (+8)‎ . . m WCQS ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
    • 22:42, 2013 September 27 (diff | hist) . . (-4)‎ . . m WBAL-TV ‎ (Fixing "Pages with citations using unnamed parameters" error.) (current)
I'm fixing small errors with legitimate edits. If these edits were blocked just because I used VE, I would not be able to fix these errors as efficiently. – Jonesey95 (talk) 18:37, 3 October 2013 (UTC)[reply]
@Jonesey95: It doesn't apply to registered editors, aka you.—cyberpower ChatOnline 14:18, 4 October 2013 (UTC)[reply]
  • Support - IP editors wanting to work faster can............. register. They should............... register. Why we still do not insist that everyone ................. register is beyond me. Carrite (talk) 06:00, 4 October 2013 (UTC)[reply]
  • Oppose After seeing more discussion, particularly from those who support this proposal, I see absolutely no reason to do this. The stated goal is to "stop" spambots, but this is only an extremely minor and temporary inconvenience for the spambots, if the assumption that these spambots actually use the API is even correct. This is far from justifying the furthering of the treatment of IP editors as second-class editors. And I'm very disappointed that there are editors who support this because it makes things worse for IP editors.
    Also, Speedy Close because there is no way this proposal will have any of the positive benefits it claims but will certainly give us the drawbacks that have been identified (e.g. breaking WP:Teahouse for IP editors). There's no need to continue to waste everyone's time with it or to continue with the anti-IP sentiments some are raising. Anomie⚔ 11:31, 4 October 2013 (UTC)[reply]
  • Oppose It should be obvious that anonymous users shouldn't run bots. However, anonymous users need to be able to use tools such as mw:VE. Also, it is trivial for spambots to edit using the standard edit form. Download the standard edit form, extract some information from it and post your spam to the server. Disabling the API wouldn't really make things more difficult for spambots. --Stefan2 (talk) 20:35, 4 October 2013 (UTC)[reply]
  • Comment To confirm, yes, VisualEditor does indeed use the API, as do many other tools (and long-term I imagine the wikitext editor will too; it's an insane architecture right now). This means this would negatively impact anonymous users of VisualEditor and similar tools, like the suggested longer-term set of curatorial bits and bobs around categorisation (a "souped-up" HotCat gadget, possibly), language links (extending Wikidata's current gadget), etc.. I would strongly advise against wanting this configuration change. Jdforrester (WMF) (talk) 01:26, 5 October 2013 (UTC)[reply]
    • I'd hope the wikitext editor would not exclusively use the API, or you'd be locking out non-JavaScript-supporting clients from editing at all. Degraded experience is fine, "go away" isn't. Anomie⚔ 11:19, 6 October 2013 (UTC)[reply]
  • Comment This is pointless. I do a lot of spambot blocking, and occasional reverting. But what I see is edits spaced out over minutes, so I do not think this wll help stop spambots. I would suggest putting in an edit filter first to try to catch this abnormal behaviour before a permanent code change. For bots out of control that can happen when logged in, so I see the gain is small in that respect. Graeme Bartlett (talk) 04:14, 5 October 2013 (UTC)[reply]
  • Oppose per the above. Humans, whether directly or indirectly, do edit Wikipedia through the API, and many people make more than two edits per minute at times.  Hazard SJ  03:27, 6 October 2013 (UTC)[reply]
  • Oppose. I don't see any indication that spambots actually use the API, and it's reasonable to assume that most probably don't. I took a look at Wikipedia:Archive.is RFC, as Equazcion suggested, and it appears that in most cases the edits were spaced out by minutes or hours. It's difficult to see what defense this proposal would provide against its intended target. —Emufarmers(T/C) 05:37, 7 October 2013 (UTC)[reply]
  • Oppose. When doing an operation such as moving content from one article to another, I have both articles open in two separate browser tabs. I perform the operation, then save the edits near-simultaneously. If this change would prevent an IP editor doing useful work such as this (for instance if they were using the VE) it would be a bad idea. Perhaps a different threshold that would clearly identify a bot (6 edits in 2 minutes maybe?). --LukeSurl t c 12:47, 7 October 2013 (UTC)[reply]
  • Oppose. API usage is legitimate when editing via some external tool, browser or script, including VE and some toolserver/labs stuff. Wikipedia website is just one most popular interface, but that doesn't mean there aren't or shouldn't be others. Yes, spambots most likely use API, but so do legitimate IPs. And this wouldn't solve the issue, just inconvenience spammers to edit slower or use website POSTs. Saying IPs should register because they use some external editor instead of the website is counter to our philosophy. —  HELLKNOWZ  ▎TALK 10:31, 10 October 2013 (UTC)[reply]
  • Support. Actually, I don't see any reason why we should allow anonymous API edits at all. --NaBUru38 (talk) 17:15, 12 October 2013 (UTC)[reply]

Something suddenly wrong with ref numbering/lettering -- and more (on IE10, at least)

The referencing machinery is so complex I don't know how to begin investigating what's wrong. Suddenly (within the last 12-24 hours?), well... here's an example:

Article text.<ref group=upper-alpha>Ref text</ref>
{{reflist|group=upper-alpha}}

should produce

Article text.[A]
A. ^ Ref text

Instead, it's producing

Article text.[A]
1. ^ Ref text

Here's actual, live code so you can see what's happening for you.

Article text.[A]

  1. ^ Ref text

If the first thing on the immediately previous line is A then things are OK for you; if it's 1 then you're getting the malfunction I'm getting. If it's anything else then the universe has gone mad.

The code's set up in User:EEng/sandbox if you want to try it yourself. EEng (talk) 23:09, 5 October 2013 (UTC)[reply]

Looks fine for me. Werieth (talk) 23:11, 5 October 2013 (UTC)[reply]
Yeah looks good to me too. A, a, and 1 are displaying in your sandbox page, respectively. equazcion� | 23:12, 5 Oct 2013 (UTC)
Wait, wait, let me guess... You're using some browser other than IE. 'Cause I'm on IE10.9.9200.16686 (for the avoidance of doubt). Just tried it on Chrome and it works correctly. So, any fellow IE sufferers getting the same problem? (Please, no gloating from Chrome/Mozilla/Safari types.) EEng (talk) 23:35, 5 October 2013 (UTC)[reply]
But, but, but Internet Exploder© is the best browser ever invented! Werieth (talk) 23:45, 5 October 2013 (UTC)[reply]
I said NO GLOATING! EEng (talk) —Preceding undated comment added 23:49, 5 October 2013 (UTC)[reply]
I wasnt gloating IE6 is the best browser ever created, and I still use it. Werieth (talk) 23:50, 5 October 2013 (UTC)[reply]
It displays correctly for me in IE10, but I'm using an older version (10.0.9200.16540). equazcion� | 23:40, 5 Oct 2013 (UTC)
I don't suppose that for the greater good you'd like to upgrade to .16686 just to see? Why does IE need 16000 versions anyway???? OK, someone out there knows what's going on with this, I'm sure. We await your wise counsel. EEng (talk) 23:47, 5 October 2013 (UTC)[reply]
It's called Agile software development. You know, like VisualEditor.... Risker (talk) 03:24, 6 October 2013 (UTC) Not gloating, I use IE all the time, just not this version.[reply]
I would but they tell me it may require a reboot, and I've got a hot streak of 78 days of uninterrupted uptime going on my computer. That and I never use IE except to check the problems other people report with it, which I don't experience because I use browsers that work and don't require reboots to update =] equazcion� | 23:50, 5 Oct 2013 (UTC)
I said NO GLOATING! But where exactly are you looking for this upgrade? And is the newest version offered beyond 16686, by any chance? EEng (talk) 23:55, 5 October 2013 (UTC)[reply]
Wait. Just answered my own question. "You've got the latest Internet Explorer for Windows 7, but you can be one of the first to try Internet Explorer 11 Release Preview." Well, gang, what do you think? Shall I stay in the frying pan, or jump into the fire? It says, "Be the first to try Internet Explorer 11 Release Preview." Who can resist that? I'd be the first! How proudly my friends and loved ones will be as they accept the Bold Self-Sacrifice Medal, on my behalf, from Mr. Gates himself! EEng (talk) 23:57, 5 October 2013 (UTC)[reply]
If I'm not mistaken, you can only install 11 if you have Windows 8, but (seems I'm mistaken, a preview release for 7 is out) yeah if you can I would. PS. this seems to be saying the latest version is a lower number than the one I have installed. I dunno. Anyway you should really invest in a real browser, they don't cost that much more :) </end gloat (for now)> equazcion� | 00:02, 6 Oct 2013 (UTC)
Yeah, well, your mother wears army boots. Anyone else seeing this? See test example I've inserted at the end of my original post. EEng (talk) 00:19, 6 October 2013 (UTC)[reply]

Wait, there's more going on than that! There's also extra vertical whitespace between each article title and the horiz line below it, and other formatting oddities -- all look good in Chrome. I've rebooted, cleared IE cache, the works. Here's what's weird -- when first reloading a page, the spacing between title and horiz line is correct, then at the last moment more space opens up between them -- it's about the same timing as when, on watchlist, all the little dropdown gadgets and stuff appear at the last moment. Hmmm. Now really -- is it just me? EEng (talk) 02:56, 6 October 2013 (UTC)[reply]

See this other bug which has a similar-sounding delay. Related? EEng (talk) 13:24, 6 October 2013 (UTC)[reply]

C'mon, really? Am I the only one? Anyone??? EEng (talk) 00:54, 9 October 2013 (UTC)[reply]
I installed the IE11 preview and I'm still not seeing your malfunction there, FYI. equazcion� | 04:44, 9 Oct 2013 (UTC)
Looks fine in a seldom-used 10.0.9200.16721. Of course, if you really have a version 10.9.9200.xxxxx it may have all sorts of top-secret features. NebY (talk) 16:35, 13 October 2013 (UTC)[reply]

Getting rid of "thank"

Tracked in Phabricator
Task T57497

How do I get rid of the "(thank)" option that has suddenly shown up when displaying diffs? Beyond My Ken (talk) 21:42, 8 October 2013 (UTC)[reply]

See Wikipedia:Notifications/Thanks#How to turn off this feature. PrimeHunter (talk) 22:18, 8 October 2013 (UTC)[reply]
Thanks, but the advice there was to check the "exclude from future experiements" button, which I already have checked - so something must have changed. Beyond My Ken (talk) 23:44, 8 October 2013 (UTC)[reply]
You're right. It worked last time I tried but not now. I guess WMF has decided it's no longer a feature experiment and shouldn't be disabled by "Exclude me from feature experiments". This in Special:MyPage/common.css should remove the link but not the surrounding parentheses:
.mw-thanks-thank-link {display: none;}
PrimeHunter (talk) 00:00, 9 October 2013 (UTC)[reply]
@PrimeHunter: Thank you, that worked. Beyond My Ken (talk) 01:08, 9 October 2013 (UTC)[reply]
Well, it seemed to work perfectly at first, but now it's leaving behind a pair of parentheses "()" and a "|" on history pages where "Thanks" used to be. I can live with that, but I do wish WMF would stop fucking around with this stuff and leave the interface alone, or at least provide opt-outs for any shiny new toys that old farts like me don't want to deal with. (The old interface was good enough for 120,000 edits, I don't need new bells and whistles, thank you.) Beyond My Ken (talk) 03:53, 9 October 2013 (UTC)[reply]
Weird. That should work. Try clearing your cache etc.? On the backend, this preference is dependent on the extension which includes some odds-and-ends related to Vector. That extension is getting phased out, though the preference will almost certainly stay as far as I know. I'll try to check if it's being futzed with currently. In any case, I think you can also use personal CSS to hide the feature, though so far I have only been able to do what PrimeHunter did. Ping: Okeyes (WMF) and Kaldari Steven Walling (WMF) • talk 00:10, 9 October 2013 (UTC)[reply]
I'm not sure why the preference would not work any more, but it's possible it was changed. Probably best to file a bug on this. Kaldari (talk) 00:26, 9 October 2013 (UTC)[reply]
I think the whole Thank thing is kinda hokey but really, are you such a curmudgeon that just being confronted by the opportunity to send a thank-you is too much to bear? I mean, there are starving children in China browsing via dialup and they would give anything to have a nice Thank button they could click, and here you are just throwing buttons away like garbage. Count your blessings. EEng (talk) 00:50, 9 October 2013 (UTC)[reply]
"Undo" was the outside option on that line, now "Thanks" is, so when I go to undo an edit, I hit thanks instead.

Please forward email addresses of unthanked Asian children and I will thank them for... something or other. Beyond My Ken (talk) 01:03, 9 October 2013 (UTC)[reply]

No, no, it's that they lack buttons they can click to do the thanking. They get 1/10 cent per click, you know. EEng (talk) 02:30, 9 October 2013 (UTC)[reply]
Maybe we should send them some of those cricket toys that click when you press them? You know, the ones they used for recognition signals in the D-Day invasion? Beyond My Ken (talk) 03:53, 9 October 2013 (UTC)[reply]
Exactly. I was going to send you a thanks for posting this, but I didn't want you to think I meant to undo your post. :) --Onorem (talk) 01:15, 9 October 2013 (UTC)[reply]
One of the options (I think it's Rollback -- can't remember for sure) doesn't even say "Are you sure?" -- just goes ahead and does whatever it does, so now I'm afraid to click any of them. EEng (talk) 02:30, 9 October 2013 (UTC) P.S. BTW, what does the "Unlink" item on the Twinkle menu do -- I'm really afraid to try that one![reply]
I never understood the unlink feature. The Twinkle documentation basically makes it sound like it's supposed to remove all incoming links to the current page. You can click it and see a list of links it's supposed to remove before actually confirming the operation. What's puzzling is that you'd think the link list would match What Links Here, but it usually doesn't. There's not much explanation around on how the two differ and I gave up trying to find out a while ago. Maybe Ill take a look through the code sometime for kicks. equazcion� | 02:44, 9 Oct 2013 (UTC)
I just tested it randomly a couple of times and it does now seem to match What Links Here (for the namespaces defined in Twinkle preferences). Maybe when I last checked there was a problem with the code, or I was just mistaken. Anyway there's your answer, it seems. equazcion� | 03:03, 9 Oct 2013 (UTC)
Way above my pay grade for sure. Thanks for investigating. Now can someone please tell me if I'm the only one with the mysterious IE malfunction? I can stand it a while, I guess, but it must be very confusing for people reading these articles. EEng (talk) 03:20, 9 October 2013 (UTC)[reply]
Yep, it's back: " | bedanken)" at the end of every line in a history. Thanking people must no longer be an experiment, and they didn't see fit to provide us with a preference to turn it off.—Kww(talk) 05:17, 9 October 2013 (UTC)[reply]
Please assume good faith. I just said above that, as far as I know, no one intended to turn off the preference to hide the thank button. It might have happened as part of a backend housecleaning, and if it did, it should be undone. We may someday ask that people use personal CSS or a userscript to hide the button, once some time has passed and it's clear how widely it's been adopted. We have not done such an analysis, and the opt-out preference should still be respected. If it isn't, it's a bug. Steven Walling (WMF) • talk 05:22, 9 October 2013 (UTC)[reply]
I actually wasn't assuming bad faith, Steven. I assumed that features eventually stop being experimental. For some, I expect you to produce a preference to turn them off, and others, I expect you to just incorporate as a permanent part of the interface. While I think the "thank" feature is the silliest thing I've ever encountered, I don't expect to have a preference for every line item in a history display. I object to the idea that a 500 line history has to have 500 individual thank-you buttons, but I that seems to be a done deal.—Kww(talk) 05:48, 9 October 2013 (UTC)[reply]
That's entirely reasonable. Thanks for the explanation. BTW: I think you're not the only one who finds the button in History to make less sense. We should maybe reconsider that placement. Diffs make perfect sense, and maybe the watchlist too (since most changes there are unique). But a history page takes more inspection usually. Steven Walling (WMF) • talk 05:53, 9 October 2013 (UTC)[reply]
If distinguishing the undo link from the thank link is an issue (it was for me), the alternative is to make them visually distinct. I set some loud colors on .mw-history-undo since it's an important function. — Scott • talk 09:47, 9 October 2013 (UTC)[reply]

randomly ordering Special lists

On some pages in the Special namespace, usernames should appear in random order, re-randomized with each refreshing, rather than alphabetically. I thought of this for Special:ListUsers/checkuser, because I shouldn't select as if I was canvassing for a friend and the person at the top of the alphabet shouldn't get overburdened with contacts for the kinds of contacts that should be distributed more evenly. Perhaps randomness should even be the default, with alphabetical being the optional alternative. Nick Levinson (talk) 15:54, 9 October 2013 (UTC)[reply]

It is possible to write Lua script modules which can read any content page (but not wp:Special_pages) and then redisplay the data, such as in randomized order, so that the same entries/usernames do not get "top billing" at the start of a list every time. It is also possible to write a Lua-based template for edit-preview, which could accept a block of text as copy/paste from a page, and then re-display the data in any format, almost instantly. -Wikid77 (talk) 20:32, 9 October 2013 (UTC)[reply]
@Wikid77: You should really check your facts before you say whether technical things are possible or not. I just checked, and it is impossible to access the content of Special:ListUsers/checkuser from a Lua module on Wikipedia. It is possible to get some data about the title, such as the title text and the namespace name, but any attempts to get the page content just return nil. — Mr. Stradivarius on tour ♪ talk ♪ 04:33, 11 October 2013 (UTC)[reply]
Thank you for clarifying (which I added into my message above), as why I had mentioned to focus on reformatting of "copy/paste" text from a Special page. Indeed, this is not UNIX, where users could merely save output from stdout into any file or application ("grep > myfile") for further processing. If the developers need something valuable to do, perhaps they could provide a Scribunto interface to allow a Lua script to read from the Special pages. -Wikid77 (talk) 05:19, 11 October 2013 (UTC)[reply]
I think its very unlikely for that to happen to non-transcludable special pages. Bawolff (talk) 12:00, 13 October 2013 (UTC)[reply]

spam

How is gods name does wikipedia stop spam? I have installed blacklists, captchas, etc on my own wikis and I am still besieged with spam.

Its disheartening that smaller but substantial hosting sites, which are similar to wikia, such as http://www.shoutwiki.com have no effective way to fight spam and to continue to allow anonymous editing, they must manually block spam bots.

This spam scourge indirectly effects wikipedia editing, because any smaller mediawiki site blocks anonymous editors. These potential editors never are introduced to the ease and wonder of editing a wiki.

Is there any working extension which can rid our sites of the relentless spam bots?

Does wikipedia have hosting, which allows me to get the umbrella protection from spam bots, but the freedom to modify the wiki the way I want? (unlike wikia), with root access?

Has the foundation ever considered hosting wikis?

These spam bots are relentless. Igottheconch (talk) 06:31, 11 October 2013 (UTC)[reply]

Have you seen mw:Manual:Combating spam? PrimeHunter (talk) 09:41, 11 October 2013 (UTC)[reply]
The short answer to How does Wikipedia stop spam? is: with difficulty, and only by continuous, dedicated, tedious work by an army of new page patrollers, recent changes patrollers and administrators flagging, deleting and blocking. They are helped by edit filters and increasingly intelligent bots, but basically it comes down to a lot of volunteer effort. JohnCD (talk) 10:32, 11 October 2013 (UTC)[reply]
  • Meanwhile, Bill Gates promised spam would be a "thing of the past" within five years, at the World Economic Forum in Switzerland in 2004, many years after predicting, "The Internet is a passing fad" in 1995. Should have been a comedian, because who would think he was serious. Anyway, we need to plan better tools to help our volunteers resist spam attacks. However, it might require creating wp:Micropages, as articles which are limited in size by edit-filters, with little room to add spam links. -Wikid77 (talk) 15:35, 11 October 2013 (UTC)[reply]
The question of whether there should be an officially supported MediaWiki hosting service was recently raised on the wikitech-l mailing list, see here and replies. That was just a kite-flying exercise, and IMHO there is no chance of such a service being introduced any time soon. Many spambots can be defeated with the Abuse Filter extension (searching Special:Version here for "abuse filter" shows that it is installed here; try that on your wiki). The polite name used here is "edit filter", but the extension is called the "abuse filter". Setting up filters is easy after you have done a few. You would probably need help to get started. Johnuniq (talk) 09:38, 12 October 2013 (UTC)[reply]

Site registered on Wikipedia's blacklist? - Yahoo! groups

I am trying to add an external link (uservoice.com/forums/209451) to the Yahoo! Groups article. But I get a error message saying that the site is registered on Wikipedia's blacklist. The website in question is an official feedback forum used by Yahoo! which contains thousands of entries. Just wondering why it is on a black list here? Thanks in advance, XOttawahitech (talk) 15:54, 11 October 2013 (UTC)[reply]

It was added to the global blacklist. You may be interested in previous discussions about this site. πr2 (t • c) 17:56, 11 October 2013 (UTC)[reply]
@Ottawahitech: You should probably request it on MediaWiki talk:Spam-whitelist. :-) πr2 (t • c) 00:18, 12 October 2013 (UTC)[reply]
@PiRSquared17:: Not the friendliest place on Wikipedia, but I managed to post a request. Fingers crossed. XOttawahitech (talk) 02:19, 15 October 2013 (UTC)[reply]

Help formatting web refs

I made this edit but just now realized it looks hideous. I tried to cite different individual sections of the same web-page, which meant I couldn't bundle them all under one ref name, but now it just looks like a bunch of separate refs to the same thing. Any advice on how to deal with this? Hijiri 88 (聖やや) 16:53, 11 October 2013 (UTC)[reply]

Use author-name cites or paragraph superscripts: A common technique is to abbreviate the extra cites to just have author name, page or paragraph, as literal text in a reftag, or add the year of publication if the author had multiple sources:

<ref>Smith. p. 23. para. 4</ref>
<ref>Smith 1998. p. 23. para. 4</ref>

However, another method is to have only one reftag cite for the source, but then add a paragraph superscript at each point where the reftag is used "[13] :¶4" by appending "<sup>:&para;4</sup>" as the superscript:

<ref name=Smith/><sup>:&para;4</sup>

The code "&para;" is the HTML entity to show the standard paragraph symbol "¶" which many readers will immediately realize means "paragraph number". Or for section, put ":sect.4" etc. There are templates which can show a paragraph superscript or just Template:Sup, but are easy to forget, and to make matters worse Template:Para is for "parameters" (should be "param") as hijacking the common abbreviation for paragraph "para" when "parm" is more common for "parameter", and so just remember "&para;" to show the paragraph symbol as a common-sense way to indicate one of several paragraphs on a webpage. -Wikid77 09:41, 12 October 2013 (UTC)[reply]

Show all types of Units

I really would appreciate if I can see units in all the system. For weight I would not like to use Google to convert it from KG to Lb. It would be great if you can work it out. — Preceding unsigned comment added by 98.250.95.86 (talk) 04:01, 12 October 2013 (UTC)[reply]

  • Template:Convert handles most conversions: Use the common Template:Convert to convert kg to pounds:
{{convert|23|kg|lb}}           → 23 kilograms (51 lb)
{{convert|23|kg|lb|abbr=on}} → 23 kg (51 lb)
For all units, see: Template:Convert/list_of_units, as a table which shows each unit-code to specify for each unit name in the list. Template {convert} is reasonably fast, about 50 conversions per second, and so it will not slow down the edit-preview or reformatting of most articles. -Wikid77 (talk) 10:38, 12 October 2013 (UTC)[reply]

Timing out issue on California State Route 1

Could someone check out the California State Route 1 page? It frequently seems to either time out, or take a very long time to save the page -- even if you try to save from the wiki source code instead of the visualeditor. I suspect it is the sheer amount of template use on the page, especially the use of templates to generate the table on the major intersections section. Thanks. Zzyzx11 (talk) 07:28, 12 October 2013 (UTC)[reply]

Update on this: a discussion has started on the article's talk page, and there are plans to convert these templates to the more efficient Lua based system. Thanks. Zzyzx11 (talk) 07:42, 12 October 2013 (UTC)[reply]
I get the same issue with Highway 401, where I get a "server down for maintenance" error when I try to save an edit. The edit is saved, but the screen still shows and it takes a mighty long time. The nesting of hundreds of templates surely plays a role in this. - Floydian τ ¢ 09:01, 12 October 2013 (UTC)[reply]
  • Faster road-junction templates: I have analyzed speed problems in the slow Template:CAint and Template:Jct, which shows road signs+names in each table of road intersections/exits. Both templates are slow, where {Jct} formats only about 5 junctions per second, and a table of 60 would burn over 12 seconds, plus 10 seconds for {CAint}. For {Jct}, there are 44 major types of roads in the U.S. plus over 1,730 subtemplates for the custom-road signs and abbreviation names, such as "FM" for Texas Farm to Market roads. I think the fastest improvement would be to separate the 50 U.S. states as 50 rapid templates, such as "Template:JctTX" to format over 100 Texas road signs/names per second and streamline that to use the name "Texas" for "TX" without lookup of TX (4 times) among the 55 U.S. states and territory abbreviations, to put "Texas" in the road-sign image, road-name wikilink, city1 or city2 wikilinked town names for Texas. Then, the template for each U.S. state could list the state's 50 or 80 current road-type abbreviations, such as Texas "FM" or "US-Alt" or "Spur" or "SR" (State Route, State Road) or "KTA" (Kansas Turnpike), etc. It is too confusing to treat the whole group of 1,730 custom-road signs (Special:PrefixIndex/Template:Infobox_road/) as a giant pool of road types. -Wikid77 (talk) 19:59, 16 October 2013 (UTC)[reply]

Diff two pages

It is possible to manually construct a link that compares one page with another: [//en.wikipedia.org/w/index.php?diff=111111111&oldid=222222222]. Is there a way to automate that? What I want is to have a link that shows the difference between the current version of a module and the current version of the module's sandbox. I don't see a way for a module to get the required oldid values to generate the diff link—or is there? Johnuniq (talk) 09:53, 12 October 2013 (UTC)[reply]

You can use Special:ComparePages for this. Example -- John of Reading (talk) 10:19, 12 October 2013 (UTC)[reply]

Image size and ratio for portraits in lists

Is there a preferred or mandated image size and ratio for portraits in lists? For example the list of field marshals at Field marshal (United Kingdom) are all of differing ratios and some are whole body images while other are head and shoulders images and yet others are pretty much focused on the head. I would suggest that an image size and ratio (possibly the golden ratio?) be mandated while a head and shoulders portrait should be preferred. This would of course require portrait versions, cropped from available images, to be created but that would not be an insurmountable difficulty. Greenshed (talk) 10:13, 12 October 2013 (UTC)[reply]

Wikipedia:RefToolbar 1.0 does not autofill from ISBN, DOI, URL, etc.

It used to and now it doesn't. It's been a month or more since it worked for me.-- Brainy J ~✿~ (talk) 17:25, 12 October 2013 (UTC)[reply]

It's because an admin needs to change var defaultRefTagURL = 'http://reftag.appspot.com/'; to var defaultRefTagURL = '//reftag.appspot.com/'; – long story short, there's a URL protocol difference (more info). Theopolisme (talk) 18:12, 12 October 2013 (UTC)[reply]
Thank you for bringing this up! Based on the lack of complaints, I thought it just stopped working for me. --NeilN talk to me 18:19, 12 October 2013 (UTC)[reply]
Wait a sec, even if that fix is implemented, I don't think ISBN lookup will work. It appears that at least the relevant script is no longer available on the server (although the others -- DOI, URL, etc. -- appear to be there). Theopolisme (talk) 18:22, 12 October 2013 (UTC)[reply]

Requested a fix at MediaWiki_talk:RefToolbarLegacy.js#Protocol_relative_link. Theopolisme (talk) 15:14, 13 October 2013 (UTC)[reply]

 Done --Redrose64 (talk) 19:01, 13 October 2013 (UTC)[reply]

Template question

This includes some line breaks that would show up in transclusions, and are shown here just for ease of reading

In my sandbox, I passed parameter value "foo" to parameter "team, and it produces "foo women's basketball"

However, if I enter "foo Lady", it generates "foo Lady basketball" i.e dropping "women". This is desirable, but I don't understand how it happens. Clearly, something detects the word Lady" and suppresses "women" but I looked at the code and do not see it.


The goal is to make it work for 2013–14 Southern Utah T–Birds basketball team, where the team name indicates the gender (but the word "Lady" is not present), and the infobox link should be:

not

Update: I asked here, because Technical 13 claimed to be studying, but gave an answer here:

It's because of the the code on Template:Infobox NCAA team season/team that uses a module to ReGex replace (.*)Lady blah blah (and it is case sensitive).

So that I can learn something about templates, can someone point me to this? I don't see it.--SPhilbrick(Talk) 02:11, 13 October 2013 (UTC)[reply]

{{Infobox NCAA team season/team}} :
|(.* Lady)(.*) women's|%1%2|plain=false}}
--- and right after that ---
|(.* Cowgirl)(.*) women's|%1%2|plain=false}}
It's around the 18th line down (line 22 in the image) just before the documentation starts. These two code segments are groups of parameters for the first and second "#invoke:String" statements at the beginning of the template (#invoke:String refers to running Module:String). For simplicity's sake, it looks something like this:
{{#Invoke:String|replace| {{#Invoke:String|replace| -- One replacement occurs, and then a second replacement is run on its result.
---- the bulk of the template code, which once done, its result put through the replacements: ----
|(.* Cowgirl)(.*) women's|%1%2|plain=false}} ---- Parameters for the "inner" #invoke:String replacement that occurs first
|(.* Lady)(.*) women's|%1%2|plain=false ---- Parameters for the very top #invoke:String replacement, that occurs second
I hope this is somewhat clear but feel free to ask things. See the screenshot to the right of the code formatted and syntax-colored, hopefully for a bit more clarity. equazcion� 02:33, 13 Oct 2013 (UTC)
You might also find the syntax highlighter gadget useful. —Remember the dot (talk) 03:55, 13 October 2013 (UTC)[reply]
Thanks. The odd thing is, I thought I did a search for the word "Lady" but I must be wrong, as it is right there.I see, this is a subtemplate--SPhilbrick(Talk) 12:47, 13 October 2013 (UTC)[reply]

Equal hight divs

Consider this.

The main eyesore is the inequal height of the 'columns'. I have now spent weeks trying to find a solution to equalize the heights in CSS. I have explored flexboxes and table-like display porperties. My final conclusion: It can not be done.

What's in the way are the headers; I want them outside the divs, but that breaks any box model I've tried. Is there any brilliant web designer who has a simple solution that does not involve inordinate ammounts of CSS, hacks, tables, javascript and still be compatible with all browsers? — Edokter (talk) — 12:00, 13 October 2013 (UTC)[reply]

Equal columns are the bane of all designers' existence, and as far as I'm aware, there's no non-javascript cross-browser solution in existence that wouldn't be considered a hack. If I were doing this I'd do it with javascript, and have the non-equal columns to fall back on for browsers without javascript. It's not really too big of an eyesore, and the vast majority of people will see the javascript version. equazcion� 12:39, 13 Oct 2013 (UTC)
And the HTML purists are still wondering why we use tables for presentation... I know it's trivial to do in jQuery, but it really should not need to be that way. — Edokter (talk) — 15:15, 13 October 2013 (UTC)[reply]
I say just go with a table, since it's easy, clean, and works everywhere. Jackmcbarn (talk) 19:23, 13 October 2013 (UTC)[reply]
I would lose the flexibility of the entire framework; the sole reason I wanted to move away from tables. — Edokter (talk) — 19:37, 13 October 2013 (UTC)[reply]
What if you used a new table for every row? Then the layout is still just as flexible (and it's still no more complex). Jackmcbarn (talk) 19:40, 13 October 2013 (UTC)[reply]
The current style is not possible using pure tables either. — Edokter (talk) — 20:53, 13 October 2013 (UTC)[reply]
What about the way I have it now (for "In the news" and "On this day")? Jackmcbarn (talk) 21:23, 13 October 2013 (UTC)[reply]
It works for display, but it breaks the flow of the headers with regard to the content in screen readers, so it is not really a solution. — Edokter (talk) — 22:50, 13 October 2013 (UTC)[reply]
I'm out of ideas then. Sorry I couldn't be of more help. Jackmcbarn (talk) 00:38, 14 October 2013 (UTC)[reply]

Just doing some cursory googling searching: is Equal Height Columns with Cross-Browser CSS and No Hacks (demo) of use? Theopolisme (talk) 20:55, 14 October 2013 (UTC)[reply]

I saw that, but I don't think it'll work for our purposes. I'd furthermore still consider that method a hack, even though the author doesn't think it is, and it uses some copious amounts of CSS that Edokter doesn't want. It involves layering several different colored containers on top of each other and shifting their positions so they cascade with each other and look like columns. I don't think that'll work for the individual "cell"-like elements currently being considered in our main-page mockup, at least not without many more containers and CSS, and the curved corners would further complicate that method (I'm not sure if it would even really be possible to maintain that appearance in the end using this method). I'm also not sure of the footer links in the mockup could be forced down that way, as the method in that article seems to just extend color down behind content divs that are still smaller, rather than allowing placement of elements based on their bottom positions. equazcion� 21:08, 14 Oct 2013 (UTC)
Maybe it is a good idea to ask on mail:wikitech-l? Helder 21:54, 14 October 2013 (UTC)[reply]

MLA style citation - inconsistent with the example in Wikipedia:Citing_Wikipedia

I just noticed that if I use the "Cite this page" tool, the MLA citation looks like this (citing Sokoban):

Wikipedia contributors. "Sokoban." Wikipedia, The Free Encyclopedia. Wikipedia, The Free Encyclopedia, 21 Sep. 2013. Web. 13 Oct. 2013.

Comparing with the example given in Wikipedia:Citing_Wikipedia, there are two obvious deviations:

  • According to Wikipedia:Citing_Wikipedia, the publisher is "Wikimedia Foundation, Inc.". According to the citing tool, the publisher is "Wikipedia, The Free Encyclopedia", an exact reduplication of the name of the website.

I'm totally confused. Which should I follow? — Preceding unsigned comment added by Xiaq (talk • contribs) 16:12, 13 October 2013 (UTC)[reply]

Request deletion of user Javascript page

What's the proper way to request the deletion of a user Javascript page, such as User:GoingBatty/script/Sources.js? {{db-u1}} doesn't seem to work in this case. Thanks! GoingBatty (talk) 23:10, 13 October 2013 (UTC)[reply]

Don't know what the right place is since templates won't work here, but I've deleted it for you. ^demon[omg plz] 23:17, 13 October 2013 (UTC)[reply]
For the future, you can just ask an admin to do it, or, you could put this at the top of the page: //[[Category:Candidates for speedy deletion by user]] - Please delete this page because ... . That should add the page to the proper csd list and provide the readable message for the reviewing admin. equazcion� 23:21, 13 Oct 2013 (UTC)
Thanks to both of you! GoingBatty (talk) 23:58, 13 October 2013 (UTC)[reply]
This question comes up a lot. {{db-u1}} and the like all work just fine in javascript; they just don't give any indication of doing so. Is there anywhere good that we could clarify this? Jackmcbarn (talk) 00:41, 14 October 2013 (UTC)[reply]
I added it to the template documentation. equazcion� 01:08, 14 Oct 2013 (UTC)
There has to be a way to force the template to visibly work on said JavaScript pages on top of just adding the page to the category. I'm thinking that it would require a little snippet of code in MediaWiki:Common.js that would detect that it is the only thing on the page and display the proper Db box for people to see. If I worked up something like this, is there any administrators that would be willing to implement it or do we need to have a big RfC to get something like that added to the site common.js? Technical 13 (talk) 02:33, 14 October 2013 (UTC)[reply]
Template use in Javascript and CSS files is scary. If you type in {{db-f1}} or {{delete}}, then the page correctly appears in Category:Candidates for speedy deletion, but you don't see this, so the page might end up being deleted even if you just want to add some Javascript code which contains the text {{delete}} for some other reason. I often see user scripts in various maintenance categories because of this. Would it be possible to show the categories somehow? --Stefan2 (talk) 21:25, 14 October 2013 (UTC)[reply]
The categories do show up, although I think the CSD categories are "hidden", so you'd need to have hidden category display enabled in preferences. It should be fairly obvious to the reviewing admin whether the template transclusion was activated by accident from within code, versus being placed at the top of the page on its own line (CSD-tagged pages don't just disappear -- an admin has to go look at them and make a determination). equazcion� 21:32, 14 Oct 2013 (UTC)
Category:Wikipedia files with unknown source includes several JS and CSS pages. Go to any of those pages and you will find no list of categories at the bottom. --Stefan2 (talk) 21:42, 14 October 2013 (UTC)[reply]

That's interesting. I had tested {{db-u1}} in my own javascript page and saw the categories during preview, so I assumed they'd show up after saving -- but it seems they don't (I just tested previewing again, along with saving this time). This will probably be a problem at least sporadically. equazcion� 21:52, 14 Oct 2013 (UTC)

"See TfD"

I don't (much) mind the nomination for deletion of templates that I've used, or the labeling of these templates as possibly ripe for deletion. However, the way that <See TfD> is mass-autosplattered seems a bit problematic. When I look at my (yes, yes) article Rob Hornstra, I see that it now has chunks autoconverted into bold italic. Is there a known bug in the implementation of the mass-autosplattering, or was my own previous formatting of the article, uh, sub-optimal, thereby inviting this or a similar disaster?

(In order to make this more comprehensible and to avoid further fuck-ups and compound confusion, I'm resisting the temptation to fiddle with the article for now. I'll let the TfD run its course. Though if somebody here who knows what they're doing wishes to jump in, go ahead. Incidentally, I do realize that a number of the external links need attention.)

Apologies if this is a matter that has already come up somewhere. -- Hoary (talk) 01:46, 14 October 2013 (UTC)[reply]

I've fixed the bold-italic bug with this edit to the TfD template. The problem was that the apostrophes in the article were interacting with the apostrophes in the template, making the bolding and italics appear in the wrong place. As for whether the tagging itself is a good idea, that's probably better discussed at WT:TFD. — Mr. Stradivarius ♪ talk ♪ 02:32, 14 October 2013 (UTC)[reply]

Aha, neat fix! Well done. And thank you. -- Hoary (talk) 09:48, 14 October 2013 (UTC)[reply]

defaultsummaries.js edit request

Could someone who knows JavaScript review the edit request at MediaWiki talk:Gadget-defaultsummaries.js? This needs to be reviewed by someone who knows what they're doing before it goes live. — Mr. Stradivarius ♪ talk ♪ 02:07, 14 October 2013 (UTC)[reply]

And could someone take a look at the request on MediaWiki talk:RefToolbar.js and have a look over the code please? — Martin (MSGJ · talk) 10:04, 14 October 2013 (UTC)[reply]
Why do people keep disabling the editprotected tags? It makes it harder for people who actually know some JS to approve the requests since they can't find them Legoktm (talk) 17:01, 14 October 2013 (UTC)[reply]

In API, action=opensearch returns only 15 items instead of 100 result asked for

Tracked in Phabricator
Task T57696

In the API sandbox (https://en.wikipedia.org/wiki/Special:ApiSandbox#action=opensearch&format=json&search=R&limit=100&namespace=0) I am querying for search results for "R" as follows; '/w/api.php?action=opensearch&format=json&search=R&limit=100&namespace=0' but it returns only the following 15 results. 

  "R",
  [
      "Race and ethnicity in the United States Census",
      "Russia",
      "Romanization",
      "Rock music",
      "Romania",
      "Republican Party (United States)",
      "Roman Catholic Church",
      "Rome",
      "Record producer",
      "Republic of Ireland",
      "Royal Navy",
      "Radio broadcasting",
      "Rolling Stone",
      "Rugby union",
      "Rhythm and blues"
  ]

I am querying for 100 search results but it is still only returning 15 results always. Can anyone please tell as to what I am doing wrong or the search results are limited to 15 results only. I am sure for a letter R there should be many articles. - Harin4wiki (talk • contribs) 09:38, 14 October 2013 (UTC)[reply]

Searching (including using the API) is resource intensive and hence far more limited than other API queries (15 could well be the non-bot limit). What are you trying to do? Could it be done a different way? For example, do you just want to search article titles or whole articles? - Jarry1250 [Vacation needed] 10:26, 14 October 2013 (UTC)[reply]
I've actually been having the same problem for a while. I figured I must've be doing something wrong, since I'm not so familiar with the API's search actions. An actual error gets returned with the results when you set the search limit to a number over 100, but if the limit is between 15 and 100, it simply cuts off at 15 with no error. Kinda strange. equazcion� 10:33, 14 Oct 2013 (UTC)
That does sound like a bug. - Jarry1250 [Vacation needed] 10:59, 14 October 2013 (UTC)[reply]
Note the bug would be in the search backend, not the API module. I just checked the code for the API module and all it does is pass the limit to PrefixSearch which passes it to the search backend. Anomie⚔ 13:24, 14 October 2013 (UTC)[reply]

I am only searching for the articles that are starting from the characters which the user keys in. So I want a list of articles which start from the characters the user has keyed in and then I can get the entire article once the user clicks on a specific article. But getting only 15 search results is too limited. I need to get atleast 50 to 100 search results. - Harin4wiki talk

Okay, in which case you should use https://en.wikipedia.org/w/api.php?action=query&list=allpages&apprefix=R&aplimit=100 or similar. Search is way more powerful than you require. - Jarry1250 [Vacation needed] 10:59, 14 October 2013 (UTC)[reply]

Thanks a lot all of you for all the help. Jarry1250, the API call you have suggested makes sense and I may go ahead and use that to get the desired functionality. Ofcourse I will tweak it for getting only the Main Page + various other combinations. I just checked on the API sandbox and that call has various parameters and it makes sense for me to use that. Once again, Thanks to all of you. - Harin4wiki talk —Preceding undated comment added 17:54, 16 October 2013 (UTC)[reply]

Toolserver problem accessing Autoblock checker

Basically, it won't. I get a 403 telling me something's expired. I don't even know if this is the right place to ask... Peridon (talk) 16:43, 14 October 2013 (UTC)[reply]

Toolserver user accounts expire if the user does not renew them every 6 months. The autoblock checker (FYI, it helps to provide a link to the problem) is run by User:Nakon. They seem to be only intermittently active, their last edit was 2 weeks ago. Mr.Z-man 16:55, 14 October 2013 (UTC)[reply]
Ta. I've left a message. Peridon (talk) 17:06, 14 October 2013 (UTC)[reply]

Removing "thanks", part II

Since bugzilla:55648 seems to not be happening -- and, it seems, more "thank" links may show up on more pages in the near future -- I wrote this script as a working mockup (fully functional, you can install it now to alter thank link behavior accordingly), and would like to propose that it or something similar become the way thank links are handled by default in the future. There is no need to constantly display a "thank" link on every list item on all pages where they could apply. Let's just dynamically show them via Javascript when an editor wants to use them (where javascript is available; the non-javascript default would be to show them all).

Note: I personally rather like the thank feature and use it with some frequency, but merely think we could be smarter about its implementation. Plastering constant thank links everywhere seems haphazard. equazcion� 20:48, 14 Oct 2013 (UTC)

I like this suggestion. The alternatives would be to make the link-text smaller, eg. ( undo | thank ), or a different color. –Quiddity (talk) 21:20, 14 October 2013 (UTC)[reply]

Polish Wikipedia irregularity

When I visited pl:Bombka, the page was shaking.—Wavelength (talk) 21:13, 14 October 2013 (UTC)[reply]

What does shaking mean? Which browser and which screen resolution is this about? --AKlapper (WMF) (talk) 23:48, 14 October 2013 (UTC)[reply]
The characters were moving back and forth on the page, too fast for me to ascertain the directions, but the distances were probably less than the width of an average character. The browser is Safari. The screen resolution is 1680 by 1050. When I resized the window, the problem disappeared, and I was unable to bring it back by resizing the window again.
—Wavelength (talk) 00:18, 15 October 2013 (UTC)[reply]
Was your house also shaking? It may have been an earthquake or something. equazcion� 00:07, 15 Oct 2013 (UTC)
It sounds like something that I encountered a while ago. It was a bug in older versions of Safari, where it is continuously adding and removing the scrollbar, due to some element being on the edge of the requirement for requiring the window to have scrollbars. It only happens within one specific pixel size and with one version of Safari if I remember correctly. —TheDJ (talk • contribs) 08:55, 15 October 2013 (UTC)[reply]
This sometimes happens to me when viewing Bing Maps (guess who owns that site?) in Firefox. It's always a left-right shake of about 1em and only when the mouse cursor is over the map. --Redrose64 (talk) 11:50, 15 October 2013 (UTC)[reply]

Unable to view logs

Been having a problem the last two days with being able to view logs--when I click the "logs" link on my contribs page, I end up getting a gateway timeout. What's going on? HangingCurveSwing for the fence 22:17, 14 October 2013 (UTC)[reply]

They seem to be working for me, but we recently did have timeout issues on log pages: Wikipedia:Village_pump_(technical)/Archive_118#timeouts. Do you have a specific page that gives the timeout for you? equazcion� 22:25, 14 Oct 2013 (UTC)
Yep, my own log page. HangingCurveSwing for the fence 22:30, 14 October 2013 (UTC)[reply]
Confirmed, got 504 timeout on that page (after a pretty long delay). I've tried several other pages that were fine though, including Special:Logs/Equazcion, Special:Logs/Kww, Special:Logs/Drmies, Special:Logs/Technical 13. equazcion� 22:49, 14 Oct 2013 (UTC)

Glitch allowing IP to create talk pages for deleted articles

An IP has found a way to exploit a glitch in the system that is allowing them to repeatedly recreate this talk page Talk:Moorti Persaud even though the article page (and the associated talk pages) have been deleted. -- TRPoD aka The Red Pen of Doom 01:52, 15 October 2013 (UTC)[reply]

IPs can create any talk pages they want, by design. I don't see a glitch here. Jackmcbarn (talk) 02:09, 15 October 2013 (UTC)[reply]
Yep, it's always been this way. If it's becoming a problem in this particular case, though, you can always ask for the page to be salted at WP:RFPP. — Mr. Stradivarius on tour ♪ talk ♪ 07:56, 15 October 2013 (UTC)[reply]
GBfan's protected it. Peridon (talk) 10:02, 15 October 2013 (UTC)[reply]

I can see why they would be allowed to create talk pages for existing pages, but for non existing articles - why would that be by design? -- TRPoD aka The Red Pen of Doom 20:18, 15 October 2013 (UTC)[reply]

WP:AFC wouldn't work unless anons could create talk pages w/o subject page. --Redrose64 (talk) 20:54, 15 October 2013 (UTC)[reply]
Isn't that only in specific namespace(s), though? -— Isarra ༆ 16:17, 16 October 2013 (UTC)[reply]

titles containing colons

I'd like a redirect from the name Bi:gwurrung, analogous to Sto:lo. However, the colon is a problem, and {{colon}} or &#58; do not help. Is this because the initial bi- is a language code? This is the spelling used in a public linguistic database we're working from, so it's a good idea to have a rd. Any workaround? — kwami (talk) 04:38, 15 October 2013 (UTC)[reply]

Yes, it's because "bi" is a language prefix, and unfortunately no, there is no way to have a redirect at that title. You'd have to put it at a similar, allowed, title instead. See Wikipedia:Naming conventions (technical restrictions)#Colons for the details. — Mr. Stradivarius on tour ♪ talk ♪ 07:52, 15 October 2013 (UTC)[reply]
Thanks. I do have similar titles as rd's, but since I'm only looking for rd support, I don't need to mess w the article. It shouldn't be much of a problem.
Too bad &#58; won't work the way &#47; will for /Gwi. — kwami (talk) 07:56, 15 October 2013 (UTC)[reply]
The difference is that the page /Gwi language can actually exist locally and the problem is just in the linking syntax, while for Bi:gwurrung the title itself is not locally possible. This is the same issue as how it's not possible for a page such as "Book: A Novel" to exist in the article namespace. Anomie⚔ 11:40, 15 October 2013 (UTC)[reply]
As an aside both bi∶gwurrung, bi꞉gwurrung and bi։gwurrung are valid titles. (I'm not sure how wikipedia feels about such unicode hacks). Bawolff (talk) 14:51, 16 October 2013 (UTC)[reply]
While those are neat hacks, they're impossible for most users to type so we should probably avoid them. Anomie⚔ 13:06, 17 October 2013 (UTC)[reply]

Wikipedia:Requests for comment/Template editor user right needs to be closed

This proposal for a new user right ended a few days ago and needs to be closed by some uninvolved administrator, if one would kindly take it on. Please use {{closing}} if/when you do decide to, so other admins know it's being taken care of. Thanks :) equazcion� 11:21, 15 Oct 2013 (UTC)

The proper place to make such requests is at WP:AN/RFC... and I see that it's already listed in the queue of about two dozen others waiting for closure. --Redrose64 (talk) 11:54, 15 October 2013 (UTC)[reply]

Redirect disables back button, continued

Further to Wikipedia:Village pump (technical)/Archive 117#Redirect disables back button - I've had this happen on a regular link. It happens when I have bookmarked a http: page - let's say VPT - and on going to that bookmark, Wikipedia initially takes me to the bookmarked page but within a fraction of a second, it automatically takes me to the https: equivalent. I know that I'm on a https: because there's a little black padlock at the left of the address bar. No amount of back-buttoning will take me back to where I started because as soon as I get back to the http: page it instantly forwards me to the https: page. Fortunately, in Firefox, it is possible to right-click the back button in order to select from a list and so skip back to a point before the bookmark was visited. IE 9 has a down-arrow link to the right of the back-button which does the same job, but equivalent features are not available in all browsers. --Redrose64 (talk) 12:12, 15 October 2013 (UTC)[reply]

I've tried to reproduce this both during the original discussion and now, but can't seem to. I've made bookmarks to the HTTP address of VPT in Firefox 24, Chrome 30, Opera 17, and IE 11, and attempting to go to them while both logged in and out does not cause any problem single-clicking the back button to get to the preceding page. equazcion� 12:22, 15 Oct 2013 (UTC)
After this happened, I looked in my cookie box and found a whole bunch of "forceHTTPS" cookies, fifteen or more from different WMF domains. I zapped them, tried again - and the problem recurred - I then found that ten or so of the "forceHTTPS" cookies had been recreated. --Redrose64 (talk) 12:51, 15 October 2013 (UTC)[reply]
Yes, that's what it is supposed to do. Unless you uncheck the option to browse securely from your preferences, you WILL be redirected. And if you don't go back more than one step, then yes your back button will do this. This is explainable and 'normal' behavior I would say. Are there really browsers that don't give you an option to go back 'further' than the directly previous page ? I'd be surprised by that. Anyways. this is not uncommon, it happens to me quite often on other websites with redirects as well. —TheDJ (talk • contribs) 15:01, 15 October 2013 (UTC)[reply]
When I was first starting in HTML - back in the days of 3.2 - I was cautioned against using client-side redirects of the form <META HTTP-EQUIV="REFRESH" CONTENT="0;http://www.example.com/"> because "that will break the back button on most browsers". It's the same problem - there is no useful time delay before the redirect kicks in.
The five major desktop browsers have the facility for selecting from a list of pages in the tab's history, but the means to access it varies, so it might not be well-known. A right-click on the back button works in Chrome and Firefox; a held-down left-click on the back button works in Chrome, Firefox, Opera and Safari. Neither technique works in IE, but that has a separate down-arrow icon to the right of the forward button. --Redrose64 (talk) 16:54, 15 October 2013 (UTC)[reply]

Interwikis

Wikimedia UK recently migrated their website to a new address, from uk.wikimedia.org to wikimedia.org.uk. As a result the details at Special:Interwiki are now out of date. I can't change this myself, so is there anyone who can updates the details, or point me in the direction of someone who can help? Is the list exclusive to Wikipedia or is there a central list which needs updating? Richard Nevell (WMUK) (talk) 12:40, 15 October 2013 (UTC)[reply]

I raised this at the WMUK Water Cooler, but it got archived - this is the final version of the thread. Please note that wikimedia.org.uk redirects to wiki.wikimedia.org.uk --Redrose64 (talk) 12:55, 15 October 2013 (UTC)[reply]
Thanks for clearing that up Redrose. I archived the thread because it looked closed but I'd forgotten it had been discussed. Richard Nevell (WMUK) (talk) 13:50, 15 October 2013 (UTC)[reply]

New Page Patrol script not working?

Hi all.
TheJosh's "New Page Patroller" script stopped working for me three-four days ago. No particular reason I can find. The script hasn't been updated in a while. (And yes, I have tried turning my computer off then on again, de-installing and re-installing the script, etc.) WinXP, latest Firefox. Not working for me in Chrome or IE either. Anyone else having this problem?
Peter "I am now aware that REGEX is not a domestic cleaning product, that would be AJAX, yeah?" in Australia aka --Shirt58 (talk) 12:35, 15 October 2013 (UTC)[reply]

@Shirt58:, should be fixed now. Legoktm (talk) 21:15, 15 October 2013 (UTC)[reply]

Tricky captchas

Please see Wikipedia:Help desk#Captcha insanity for a problem that an IP has been having with the captcha system. It's easy for those of us who never see them to not realize that people having trouble with the system are the ones who can't report it. —Anne Delong (talk) 14:27, 15 October 2013 (UTC)[reply]

Good point. Reminds me of a challenge I had recently; I wanted to post something to a discussion site, which was information interesting to the audience, so I was not looking for help, I was looking to help. I couldn't post without registering. I tried registering, and got an error. The message with the error said in the case of errors, contact the sysadmin. But you had to login to contact the sysadmin. Oh well, their loss. We should make sure we don't fall into the same trap.--SPhilbrick(Talk) 20:06, 15 October 2013 (UTC)[reply]

Switch of mbox templates and category handler to Lua

I've started a thread over at Template talk:Mbox about my implementation of all of the {{mbox}} family templates, plus the {{category handler}} template, as Lua modules. These templates have millions of transclusions, so I would appreciate comments and some more eyes on the code. Please let me know what you think over at the discussion page. — Mr. Stradivarius ♪ talk ♪ 15:15, 15 October 2013 (UTC)[reply]

Infobox image uses no longer listed at image pages

I randomly clicked File:ThrustSSC front.jpg on the Main Page to see what the hell it was. There was no article link at the image description, so I checked the "File usage" list, and surprisingly saw nothing from en-wiki there, despite having later found it in ThrustSSC's infobox. Checking a couple of other examples, like America the Beautiful and its infobox image, I found the same issue. My mistake, looks like that image does list the article usage correctly.

I was thinking this had to do with {{infobox}}'s recent move to use a Lua module rather than traditional wikicode, but seeing as this doesn't happen with all infoboxes, I'm not sure what caused this. equazcion� 16:51, 15 Oct 2013 (UTC)

I believe that the file uses is reported from caching, and updating the page(s) that use the image should update that list. It has been one-off cases in the past that the file use list is not always correct but after some time it fixes itself (hence caching issues come to mind). --MASEM (t) 16:59, 15 October 2013 (UTC)[reply]
The article has contained the image for at least some months now, and has had a number of edits today. I just tried a null edit as well anyway and that doesn't seem to update the list. equazcion� 17:03, 15 Oct 2013 (UTC)
(edit conflict) I made a null edit to ThrustSSC just in case, but that doesn't fix it. At File:ThrustSSC front.jpg#mw-imagepage-section-linkstoimage there is the usual list of pages, including Main Page, but not including ThrustSSC. --Redrose64 (talk) 17:14, 15 October 2013 (UTC)[reply]
I think the issue might be that the infobox has File:Cmglee Thrust SSC front.jpg, which redirects to File:ThrustSSC front.jpg. Chris857 (talk) 17:35, 15 October 2013 (UTC)[reply]

Poor peer-reviewing

Take a look at this:

http://www.cbc.ca/news/technology/bogus-science-paper-reveals-peer-review-s-flaws-1.2054004

For Wikipedia, with its reliance on citations of apparently-reliable material, this should be worrying.

DOwenWilliams (talk) 19:23, 15 October 2013 (UTC)[reply]

  • Science-topic editors using multiple older sources: There has been talk among the science-topic editors to require multiple, independent sources to avoid problems of unusual studies, which perhaps report fringe cases as not adequately representing the theories being presented. That technique, to be more conservative and read multiple sources before posting text into Wikipedia, would deter the posting of false data from a few rigged publications. However, it is interesting to see science journalist John Bohannon (re-)entering the realm of gonzo journalism to be the culprit who submitted the falsified document to so many science journals. Anyway, perhaps we need an essay to recommend and explain techniques for using multiple, independent sources along with wp:RS. Essay "wp:Find sources" is aimed at locating even one source to back the text. -Wikid77 (talk) 04:25, 17 October 2013 (UTC)[reply]

Flow Office Hours - updated

Hi. The Flow team will be holding an IRC office hours session in #wikimedia-office at 18:00 UTC on 17 November to talk about Flow as a whole. (with more office hours, in the future). The most up-to-date documentation pages are currently WP:Flow/MVP and WP:Flow/Design FAQ. Looking through those (if you haven't already) before the office hours, would be ideal. Thanks. Quiddity (WMF) (talk) 23:14, 15 October 2013 (UTC)[reply]

Due to multiple-human-error (the best kind of error!) the Office Hours meeting was announced with the wrong month. The logs for today's (quiet) meeting, can be seen at m:IRC office hours#Office hour logs.
The updated time and date of our next office hours meeting is: 18:00 UTC on 24 October. Thanks, and sorry about the mixup. Quiddity (WMF) (talk) 21:31, 17 October 2013 (UTC)[reply]

Question about pending changes

On Amanda Bynes (page history), a new editor added a non-free image that I think fails our non-free image criteria so I reverted it (I've got the reviewer user right). I just noticed they've "reverted" my reversion 3 times in a row - without me reverting or anyone else editing between their "reversions". The last was their 10th edit so it autoconfirmed them and the image went live.

Can a 4-day-old account just keep hitting "undo" on a PC-protected page until they exceed 10 "edits" and autoconfirm? Anthonyhcole (talk • contribs) 00:44, 16 October 2013 (UTC)[reply]

Yes. If they're causing disruption, treat them like you would any other disruptive user editing a non-PC protected page. Jackmcbarn (talk) 03:04, 16 October 2013 (UTC)[reply]
Cheers. --Anthonyhcole (talk · contribs · email) 03:25, 16 October 2013 (UTC)[reply]

Video files not working

I'm using internet explorer 9 and when ever a Wikipedia article has animation in the article itself, it works totally fine but when ever an article has a video file and I click on it, it won't load the video and I'm unable to watch the video file. That happens every single time for any video file at all on Wikipedia on the laptop I got from future shop 2 years ago. Is it possible for Wikipedia to adapt to fix that problem? Maybe, Wikipedia can do so by changing the method of showing video files. On the other hand, image files display properly with no problem and I can view them at full resolution. Blackbombchu (talk) 03:09, 16 October 2013 (UTC)[reply]

Internet Explorer needs Java to play the files hosted on Wikipedia. Most other browsers have native support built in. — Edokter (talk) — 15:57, 16 October 2013 (UTC)[reply]

Commons-only file acting strangely

Tracked in Phabricator
Task T57776

Can someone have a look at File:Svengalideck.png? It's also on Commons, and so far as I can tell it doesn't exist locally, having been deleted in May, but it's not showing up properly for me. I just get a file history and a dead link to the full resolution (no thumbnails, no "this file is on Commons," no description, just the dead link and the file history). Is anyone else seeing the same thing? --NYKevin 06:31, 16 October 2013 (UTC)[reply]

I am seeing exactly the same thing. It makes absolutely no sense whatsoever that a file transferred in May would not show up correctly. I've posted something over at commons:Commons:Help desk#en:File:Svengalideck.png in case this is a commons problem and not an en.wiki issue. VanIsaacWS Vexcontribs 06:42, 16 October 2013 (UTC)[reply]
It's also not displaying properly at the Trick deck article. This sounds like a database glitch to me - bugzilla would probably be a better place to report this than Commons. — Mr. Stradivarius on tour ♪ talk ♪ 07:29, 16 October 2013 (UTC)[reply]
Filed in bugzilla —TheDJ (talk • contribs) 08:12, 16 October 2013 (UTC)[reply]
It seems that the file still exists locally, although our copy is broken. If an administrator can delete the local copy, then the Commons copy would probably display properly on Wikipedia. If the file can't be deleted for some reason, then a temporary solution could be to upload a new copy of the file locally under a different name until it becomes possible to delete our local copy of the file. --Stefan2 (talk) 08:21, 16 October 2013 (UTC)[reply]
It's probably best to let someone who knows the database system take a look before we mess around with it. This might be a symptom of a larger problem, and if that's the case then it will be easier to find out what went wrong if we leave things as they are. — Mr. Stradivarius ♪ talk ♪ 08:57, 16 October 2013 (UTC)[reply]
I see nothing strange; the local page shows the description page from Commons, and all images display properly for me. — Edokter (talk) — 16:04, 16 October 2013 (UTC)[reply]

Confirmed, I am no longer seeing the problem. I believe [1] may have fixed it. --NYKevin 16:18, 16 October 2013 (UTC) [reply]

After OK edit saving, the return page is an error (504 or ec)

  • Last days or weeks, sometimes when I save an edit it is saved all right, but then it returns an error message page: 504 gateway timeout nginx/1.1.19 (no more). For example this edit just now.
  • At other times, I get an error conflict page while, again, the edit is saved as expected with no conflict in sight. Example: this edit was saved OK (is seen on the fresh page), returns an edit conflict message pag. Remarkable to me: the ec message diff shows a split parsing (right word? split edit saving?): my entered -~~~~ was saved separately into the textual sign, after parsing my content text. That is, the two diffs (from page history and ec page) do not show the same. screenprint). -DePiep (talk) 09:08, 16 October 2013 (UTC)[reply]
    @DePiep: The 504 "gateway timeout" you're seeing is because the page is simply way too large. It's currently 200k bytes which is way above the recommended limit. I've lowered the archive time to 30 days instead of 45, hopefully that will reduce the size of the page to make it more manageable. Legoktm (talk) 05:52, 17 October 2013 (UTC)[reply]
OK, thx. Consider resolved here then. And unrelated to the second issue, false edit conflict (below). -DePiep (talk) 14:38, 18 October 2013 (UTC)[reply]
  • Self edit-conflict of "~~~~" against saved signature: I took time to read those diff sections which rejected your use of tildes "~~~~" as conflicting with your saved timestamp-signature, where you had immediately saved those changes as the latest revision (shown in screenprint). That result is so twisted, that it took me a few minutes to understand, and it does seem like an attempt to re-save the same edit twice. I need to re-read and re-think, but I thank you for posting the screenprint, and I will watch for similar future false edit-conflicts of tildes "~~" clashing with the saved signatures. I get so many edit-conflicts, per week, that I had stopped reading them and just re-ran the edit with copy/paste text. It also doesn't help that the developers think edit-conflicts are ok, and not a massive, high-priority problem to be solved (this month) by changing diff3.c and setting read-lock semaphores in the system. Even in a distributed system, with mirrored databases, there will be write-lock and read-lock operations which limit access to the latest stored revision of a page. The whole concept of test-and-set access to resources has been known to software developers for decades. Most conflicts are easy to auto-merge, and computer scientists have known for years how to recover from edit-conflicts. It is merely a case of comparing 3 revisions of a page, while a read-lock prevents another user from trying to perform an edit-save against the same prior revision, during the intervening few seconds. -Wikid77 (talk) 04:02, 17 October; shortened text) 19:03, 17 October 2013 (UTC)[reply]
So, perhaps the large size of the page sometimes triggered a timeout-exit which caused the self edit-conflict against the just-saved revision. In remembering past false edit-conflicts, I think many occurred in large pages, or smaller pages with many large templates as might reach a timeout limit. -Wikid77 (talk) 16:32, 17 October 2013 (UTC)[reply]

Inconsistency between WIkipedia API, action=raw and html view

I discovered inconsistency between the Wikipedia API, the action=raw view and the regular HTML view. e.g. the village Oberalm in:

  1. regular HTML-view http://en.wikipedia.org/wiki/Oberalm has a population_total of 4247 people
  2. Wikipedia API: http://en.wikipedia.org/w/api.php?action=query&prop=revisions&rvprop=content&format=xmlfm&titles=Oberalm&rvsection=0 -> Einwohner (which stands for population)=3989
  3. action=raw view: http://en.wikipedia.org/wiki/Oberalm?action=raw Einwohner=3989

I learned that the regular HTML-view uses a template and overwrites the raw-source of the entry. I think this prevent people to update the real number in the source. But in case we do not need a up to date source, shouldn't the template be used whichever way the content is requested?

Interesting sidefact: The German Wikipedia page about Oberalm http://de.wikipedia.org/wiki/Oberalm does nor list Einwohner neither population. Einwohner seams only to be shown on the English Wikipedia.

Pedak (talk) 14:52, 16 October 2013 (UTC)[reply]

So diving down the rabbit hole I discovered that {{Metadata population AT-5}} is the source of the number that is being displayed in the article. Werieth (talk) 15:10, 16 October 2013 (UTC)[reply]
PS dewiki does show the population Einwohner: 4.247 (1. Jän. 2013) Werieth (talk) 15:12, 16 October 2013 (UTC)[reply]
(edit conflict) Yeah the 4247 number is actually displayed in the German article, just not in its infobox. It's in the article content, via the template de:Vorlage:EWD. I'm not sure where the number is actually coming from though, because it's not listed at wikidata:Q387186. equazcion� 15:15, 16 Oct 2013 (UTC)
I see it in the infobox, its about the 8th item down. Werieth (talk) 15:15, 16 October 2013 (UTC)[reply]
de:Vorlage:Metadaten Einwohnerzahl AT-5 is the German version of our template which has the value. Both the article content and template invoke it. Werieth (talk) 15:19, 16 October 2013 (UTC)[reply]
(edit conflict)The article uses Template:Infobox Ort in Österreich, which redirects to Template:Infobox Town AT. But that template discards the "Einwohner" setting provided in the article if Template:Get Austria population has a value for the code from Template:Get Austria location code:
{{Get Austria location code|Oberalm}} gives Template:Get Austria location code and {{Get Austria population|key=50208}} gets the population from {{Metadata population AT-5|50208}}, which finally gives Template:Metadata population AT-5, the number in the article.
That template was most recently updated in July 2013, while the number in the infobox is from 2007 and should probably just be removed since it's filled in automatically as of 2011. Mr.Z-man 15:25, 16 October 2013 (UTC)[reply]
I see now, the population numbers are listed statically in the code of {{Metadata population AT-5}}. Yes the template population parameters should probably be removed from all documentation, and maybe articles via a bot to avoid confusion? equazcion� 15:30, 16 Oct 2013 (UTC)
PS. I feel bad for anyone who comes across an article where this is implemented and wants to correct a population number. I get that this is probably somewhat of a good way to make population numbers consistent among all wikis and their articles, but as far as ease of editing goes, it's rather idiotic. equazcion� 15:34, 16 Oct 2013 (UTC)

So it occurs to me that all population numbers should probably be migrated over to wikidata, rather than being listed in some buried template. Any thoughts on the feasibility of that? equazcion� 15:43, 16 Oct 2013 (UTC)

Well, it is not "some buried template" but rather a whole set of population-data templates which have been kept in sync with German Wikipedia for 3 years, covering 7 nations. However, when I wrote those templates in 2011, I was in hurry to update 2,500 population counts (in 25 edits) and expected others to manage them, but it did not happen, so let's discuss briefly, below, at: "#Managing population-data templates". -Wikid77 05:46, 17 October 2013 (UTC)[reply]

Thank you a lot for the explanation where the numbers come from. Is there any way to get the actual numbers or to get the template also to work in raw-mode, via the API or any other structured way? My problem is: If i read the data in a structured way, people try to validate and see different information in the standard html view, they will complain about it. Any suggestions or workarounds for that? Pedak (talk) 19:08, 16 October 2013 (UTC)[reply]

http://en.wikipedia.org/w/api.php?action=query&prop=revisions&rvprop=content&format=jsonfm&titles=Oberalm&rvsection=0&rvparse Werieth (talk) 19:14, 16 October 2013 (UTC)[reply]

Managing population-data templates

Okay, I admit I did it. I am wondering if the infobox template should be coded with "population=auto-set" (or similar) to remove the outdated population counts "Einwohner=4247" and hint how the current count is "auto-set" by the template. Back in September 2011, when I wrote those 9 Austria population-data templates, {Metadata population AT-1} to {Metadata population AT-9}, I just wanted the 2,400 town populations to be rapidly updated, which I did with just 27 "simple" edits in a few hours of spare time, with full documentation. I was thinking no one else wanted to edit those 2,400 pages and reset the population counts (ya think?). However, to my surprise, another editor actually soon edited all 2,400 Austria town articles to set the town-code numbers but did not blank the population count nor set it to "Einwohner=auto-set" (or such). It would be an easy Bot task to edit all WhatLinksHere to insert "population=auto-set". On the other hand, the prior hard-coded population counts are fairly accurate (within 10%), and so a tool which extracts the population data would at least show a "ballpark estimate" of thousands of inhabitants with the actual year (2007) of the cited data. In retrospect, it was an awesome feat to quickly show current populations (now year 2013, as updating 7,200 counts), while the internal template parameters reported the prior population as of the stated year (and all with just 27 edits in my spare time!). Now there are town infoboxes of 6 nations (more on dewiki) which auto-set town counts from the related population-data templates. Perhaps if a Bot could reset the old counts to the current counts and set year "2013" then tools which extract population counts from the infobox would be more current for the next few years (as long as a Bot would be run to edit the 2,500 towns anyway). As for conversion to Wikidata, show me where Wikidata has accurately updated 30,000 town populations for 3 years, and then we might consider switching from population-data templates. Perhaps we should move this to a Bot-discussion page, to see what would be agreeable as a Bot task. -Wikid77 (talk) 05:46, 17 October 2013 (UTC)[reply]

(note - tldr. only read the last sentence): this does not make any sense as "bot task". this is a classic wikidata item. surely, population numbers will be added to wikidata - hopefully soon. once they are, all these templates should simply be removed, and normal wikidata references should replace them. peace - קיפודנחש (aka kipod) (talk) 23:31, 17 October 2013 (UTC)[reply]

We need a watch list for categorization

It would be very helpful to have some way to "watch" categories... something that would alert us whenever an article gets either added or removed from a specific category. Is this possible? Blueboar (talk) 15:45, 16 October 2013 (UTC)[reply]

I agree. Until we have that, there's the Catwatch userscript, see Wikipedia:User scripts#Watchlist. equazcion� 15:47, 16 Oct 2013 (UTC)
Special:RecentChangesLinked does that. Matma Rex talk 16:26, 16 October 2013 (UTC)[reply]
Yes, this is a very good idea - one the community has repeatedly asked for. bd2412 T 16:52, 16 October 2013 (UTC)[reply]
I appreciate that there are external programs that serve a similar function... but they are not quite what I was asking for. what I am seeking is a "watch this category" function on our menu bar. Blueboar (talk) 18:45, 16 October 2013 (UTC)[reply]
Although not listed at WP:PEREN, I'm pretty sure that this has come up several times before; many of us would like it (myself included), so that fact that it's not been done suggests that it's difficult. --Redrose64 (talk) 22:20, 16 October 2013 (UTC)[reply]
It is difficult, because category listings aren't actual pages that change when things are added or removed from them. I don't think there's any bona fide "thing" that really changes when a category tag is placed or removed from a page, other than a cached database query, and even that only occurs when someone checks a category listing (I think). It might be worth having a log attached to each category that actually changes when category tags are added to or removed from pages, and then the watchlist could easily be tweaked to show those log changes for watched categories, the same way we can now see deletion/protection/move log changes for pages we watch. equazcion� 23:47, 16 Oct 2013 (UTC)
the fact that category lists are not "real" does not mean this request has to be very difficult. (the following represent my understanding, and though i believe it's pretty much correct, i may be wrong): basically, wathlist entries and categories are calculated on page saves. the requested functionality would require to calculate "category diff" on page save (not that difficult - the old cats are listed in the DB, and the new cats are freshly calculated), and for any category in the "diff", add watchlist item to all the watchers of this category. now, without changing the DB, what the user will see is "foreign" pages on their watchlist (i.e., pages they do not actually watch), and in order to realize that this is a category addition or removal, they will need to check the actual diff. (it's possible to include the cat instead of the page in the watchlist, but this will not be very useful, because there will be no indication of what actually changed in the cat). there is another interesing hitch, which is change in category that came from a template edit: if the category in the template is inside an "includeonly" tag, editing the template can add or remove tons of pages from the cat, and the mechanism i described above will totally miss it (we have similar issue with watchlist and templates: when a template changes, people who watch pages that transclude this template do not get indication in their watchlist, even though the page they are watching was, in reality, changed). so what i described is not a perfect solution, but it solves 78.12% of the problem. i think that perfect solution will require some changes to the DB - it won't be enough to change the code.
some disclosure: some time ago the foundation solicited for "google summer of code" ideas, and i tried to pitch this one. nobody took it...
peace - קיפודנחש (aka kipod) (talk) 23:25, 17 October 2013 (UTC)[reply]
You could have a bot running that updates a listpage that lists the contents of the category, and then watchlist the listpage. -- 76.65.131.217 (talk) 06:40, 19 October 2013 (UTC)[reply]

Back button disabled when I sign in (on Firefox)

Referring back to this question, I'm now experiencing the same behavior as when I created a redirect (this was not happening until very recently). When I try to go back to what I was doing before I went to Wikipedia, I am not allowed to and I keep going forward to the article I was looking at when I signed in. I was told by someone who doesn't know how to find it that it is Firefox 24.0.— Vchimpanzee · talk · contributions · 15:53, 16 October 2013 (UTC)[reply]

I see that you also posted this problem at Wikipedia:Reference desk/Computing#Back button disabled when I sign in (on Firefox). Per WP:MULTI, I've replied there. --Redrose64 (talk) 22:17, 16 October 2013 (UTC)[reply]

Template parameters don't work in ref tags

Resolved

I'd like to be able to output ref tags from a template, which use given parameters for their content. After much finagling, it seems like that might not be possible. Does anyone know of a way? equazcion� 17:52, 16 Oct 2013 (UTC)

Yes, use {{#tag:ref| {{{1}}} }}. Matma Rex talk 18:04, 16 October 2013 (UTC)[reply]
Awwwesome, thanks! equazcion� 18:07, 16 Oct 2013 (UTC)
  • The #tag also allows parameters inside math-tags: Beyond allowing wp:Subst'ing of templates inside of reftag footnotes, the #tag also allows templates to use parameters in a mathematical formula formatted inside math-tags. Compare the following:
  • <math>f(x) + {{{nam|g}}}(x)</math> →
  • {{#tag:math |f(x) + {{{nam|g}}}(x)}} →
Note how the use of #tag:math allowed the parameter "{{{nam}}}" to be evaluated with "g" as the 2nd function inside the math-tag formula. We can write clever templates which generate complete equations or footnote reftags where the contents of the footnote are specified by parameters. -Wikid77 (talk) 19:27, 16 October 2013 (UTC)[reply]
  • This allowed me to successfully add refs to the {{userscript}} template that contain specific installation instructions for each script at WP:US, in case anyone cared what this was for. Thanks again Matma Rex. equazcion� 19:48, 16 Oct 2013 (UTC)

The Wikipedia Adventure, alpha-testers needed

Hi folks, I've been working for the past 7 months on an interactive guided tour for new editors called The Wikipedia Adventure, as part of a WMF Individual Engagement Grant. The game is an experiment in teaching our aspiring future editors in an educational but playful way.

  • This week I need some alpha-testers to kick the tires and basically try to break it. I'm interested in general impressions and suggestions of course, but I'm really looking for gnarly, unexpected browser issues, layout problems, workflow bugs, and other sundry errors that would prevent people from playing through and having a positive experience.
  • If you're able to spend 1-3 hours doing some quality assurance work this week, you would have: a) my sincere gratitude b), a sparkly TWA barnstar, c) special thanks in the game credits, and d) left your mark on Wikipedia's outreach puzzle and new editor engagement efforts
  • Please note that the game automatically sends edits to your own userspace and it lets you know when that will happen. If you want, you can register a new testing account just for the game, but it won't work properly unless you're logged-in by step 8 of mission 1 when it lets you register on the fly.

If you're interested, please add your name below and have at it. You can post feedback to WP:TWA/Feedback. Thanks and cheers! Ocaasi t | c 20:51, 16 October 2013 (UTC)[reply]

Try out The Wikipedia Adventure

I'm interested and on the bug-hunt. Will report back this week

  1. Add your name here
  2. Or here
  3. Or here...
Questions
sendMessage = function( targetPage, msgPage, linkTo ) {
	var api = new mw.Api();
 
	api.get( {
		'action' : 'query',
		'titles' : msgPage,
		'prop'   : 'revisions|info',
		'intoken' : 'edit',
		'rvprop' : 'content',
		'indexpageids' : 1
	}, {
		'ok' : function(result) {
			result = result.query;
			var page = result.pages[result.pageids[0]];
			var text = page.revisions[0]['*'];
			api.post(
				{
					'action' : 'edit',
					'title' : targetPage,
					'appendtext' : "\n" + text,
					'summary' : 'New Message (simulated automatically as part of [[WP:The Wikipedia Adventure|The Wikipedia Adventure]])',
					'token' : page.edittoken
				},
				{
					'ok'  : function() { window.location.href = linkTo;}
				}
			);
		}
	}	);
}

Ocaasi t | c 21:23, 16 October 2013 (UTC)[reply]

@Ocaasi: Did you ever get my messages in -helpers? πr2 (t • c) 22:17, 18 October 2013 (UTC)[reply]
@PiRSquared17:, ah sorry I haven't been on irc for a few days. I'll check now. Cheers, Ocaasi t | c 22:22, 18 October 2013 (UTC)[reply]

Wikipdia font change?

Hi, fonts throughout Wikipedia have got noticeably smaller for me (Win7/IE10). No noticeable changes to any other websites, no changes to browser zoom level or other settings, no changes to any other system settings. Has something changed at Wikipedia? 86.160.86.177 (talk) 21:07, 16 October 2013 (UTC)[reply]

This has also happened to me I use Firefox - not sure about any changes made but... Pressing Ctrl++ will increase the size of the text..the size of all. There is also a setting to be had in your windows setting as seen at ... Make the text on your screen larger or smaller - from http://windows.microsoft.com. - Moxy (talk) 23:54, 16 October 2013 (UTC)[reply]
Ctrl+ just changes the browser zoom level, which the poster says they haven't changed. I'm not seeing this font change though. I checked Firefox and IE, and I'm not seeing it in either one -- though I'm using IE11, not 10. Does the smaller font still remain after restarting the browser and restarting your computer? equazcion� 00:00, 17 Oct 2013 (UTC)
And just to clarify, are you both currently seeing this change (Moxy and 86.160.86.177)? Anyone else seeing this? equazcion� 00:01, 17 Oct 2013 (UTC)
  • I'm not seeing any change in FireFox. I have IE7 available, but not 10 (although they might at university, and I'd be happy to check). Technical 13 (talk) 00:18, 17 October 2013 (UTC)[reply]
  • I have deleted all local cached files and rebooted, and the problem has gone away. Sorry, I should have tried that earlier. Unfortunately, since I did both operations together, I do not know which one fixed it. Strange. 86.160.86.177 (talk) 00:26, 17 October 2013 (UTC)[reply]
  • Of note, I had the same problem periodically over the weekend; at one point, I had three different font sizes in three different tabs. (FF24/Win7) I just closed the wrong-sized tabs and decided to blame it on atmospheric conditions, since it made no sense. Risker (talk) 00:47, 17 October 2013 (UTC)[reply]
  • You can usually hit Ctrl+0 and it will go back to "default" size. Most of the time zoom levels happen by accident. Some browsers will change zoom level if you are holding down a control key and move the mouse, for instance.--Jorm (WMF) (talk) 00:49, 17 October 2013 (UTC)[reply]
As far as my problems were concerned, they had nothing to do with broswer zoom level. I understand how browser zooming works, and the zoom level, at least as far as it was displayed by IE, assuming IE wasn't lying, had not changed. Also, any combination of restarting the broswer, closing and opening tabs for Wikipedia and other sites, etc., always resulted in Wikipedia being shown in a smaller-than-usual font but no change to any other website. 86.160.86.177 (talk) 01:11, 17 October 2013 (UTC)[reply]

Special Pages and Guided Tours

Hey folks, any idea why...

The goal is to load one's contributions and then have it load the Guided tour step as listed in the url. Cheers, Ocaasi t | c 03:03, 17 October 2013 (UTC)[reply]

I think it's due to Template:Bug Steven Walling (WMF) • talk 05:42, 17 October 2013 (UTC)[reply]
Thanks Steven (WMF). Do you know if I can use a should skip/isPage instead, and if so what isPage is the correct landing? Ocaasi t | c 16:22, 17 October 2013 (UTC)[reply]

Watchlist cleanup check box

Currently here's what I have: "You have 8,545 pages on your watchlist (excluding talk pages)."

I know it's possible to edit the raw watchlist, but I'd like the option to purge it as I look through the currently active list. Can a box be added beside each line so that I can check all those I want to get rid of and thus clear the list as I go along each day? -- Brangifer (talk) 05:59, 17 October 2013 (UTC)[reply]

Isn't Special:EditWatchlist what you need (that's the link for "View and edit watchlist")? If doing bulk changes (like deleting all items that match a particular search), you would use Special:EditWatchlist/raw, copy the text into an editor, get it how you want, paste it back into the edit box, and click Update. However, I don't think there's anything realistic you can do once you have a few thousand items, apart from severely pruning almost everything. Johnuniq (talk) 06:12, 17 October 2013 (UTC)[reply]
I think Brangifer is looking for something that can be used directly from the "view relevant changes" screen, i.e. the regular watchlist. I use popups for this - you can mouse over the article link and then click the "un" link from "un|watch" in the "actions" menu. However, maybe someone's written a user script that makes an actual button/checkbox/link on the watchlist page itself. Is anyone aware of such a thing? — Mr. Stradivarius ♪ talk ♪ 06:30, 17 October 2013 (UTC)[reply]
Yes, that's what I'm looking for. I already know about the Special:EditWatchlist, but I'm only interested in pruning the no longer relevant ones which actually keep popping up on my watchlist when I'm looking at recent changes. A box would be nice so everyone has access to it on their own watchlist, without having to install a script. -- Brangifer (talk) 14:25, 17 October 2013 (UTC)[reply]
This may be more work than you want to do, but here's what I do (this on on a Mac, in Firefox, so the keyboard commands may be different for you). Hold down the Command key and click on each of the pages that I want to unwatch. This opens a bunch of tabs in the background. Close the watchlist and position your mouse arrow over the blue star that indicates that the article is watched. Click it, then use Command-W to close the tab (without moving the mouse). Click the star on the next article, Command-W, click, Command-W, etc. You can unwatch 50 articles in a couple of minutes if you get into a rhythm. – Jonesey95 (talk) 15:04, 17 October 2013 (UTC)[reply]
I'm familiar with that method and have used it, but I want something easier. I already use Firefox most of the time and often have over 20 tabs open. -- Brangifer (talk) 06:32, 18 October 2013 (UTC)[reply]

Proposal to unprotect Wikipedia:Cascade-protected items

Now that Wikipedia:Requests for comment/Template editor user right has passed, we will very soon have a new user right to allow editors to edit protected templates, plus a new protection level that will only be available for use in the template and module namespaces. However, editors with the new user right will not be able to edit cascade-protected pages. Because of this, I have made a proposal that Wikipedia:Cascade-protected items and other similar lockboxes be unprotected and the new protection level be used instead. If you're interested, please leave a comment on this proposal on the discussion page. — Mr. Stradivarius ♪ talk ♪ 06:46, 17 October 2013 (UTC)[reply]

Backlog status Category:Pages with broken reference names

Category:Pages with broken reference names the status of backlog should be more than 50 or better 75 items, because there are always 20 templates in this category. See [2].

Please change: 

This page has a backlog that requires the attention of willing editors.
This notice will hide when this category has fewer than 75 items. (recount)

Thanks --Frze > talk 08:10, 17 October 2013 (UTC)[reply]

At Category:Pages with broken reference names you will find {{backlog|25}}. All you need do is alter that figure. --Redrose64 (talk) 08:24, 17 October 2013 (UTC)[reply]
Thanks  Done --Frze > talk 08:34, 17 October 2013 (UTC)[reply]

New REFBot

Thanks - Vielen Dank!

DPL bot and BracketBot are the best inventions of Wikipedia. It's time for a new Bot. We need the

If a user contributes a broken reference name, an incorrect ref formating (or a missing reflist), please inform the user who caused this error. It is so outrageously hard work to correct all these errors afterwards, from someone who is not holding the factual knowledge. For example: it took me a week to work up the backlog of Category:Pages with broken reference names - more than 1500 items, some disregarded more than two years. Search with WikiBlame for first entry of ref, making the changes, inform the users... annoying. Thank you very much. --Frze > talk 12:25, 17 October 2013 (UTC)[reply]

I ask for little consideration please. It takes several minutes of work - only because of a lack of character. Ten times and more per day. For example see Cite error: The named reference Media2 was invoked but never defined What' wrong? > Compare selected versions > Fix broken reference name. Why doesn't a Bot send a message to the polluter of the error? Why must other users rid of the mess? With BrackBot and DPLBot it is so easy --Frze > talk 06:44, 18 October 2013 (UTC)[reply]
I suggest you bring this up at WP:Bot requests. Ryan Vesey 07:05, 18 October 2013 (UTC)[reply]
Done. Thanks. --Frze > talk 07:17, 18 October 2013 (UTC)[reply]

Wikipedia:WikiProject Articles for creation/Submissions

Dear tech people: On the above page is says that there have been, since 2008, 2714 declined submissions. The truth is that there have been well over 120,000, and while a goodly portion of these have been deleted (several months ago it was about 40,000 deleted, but since the new db-g13, many more have bitten the dust), there are still about 45,000 in existence. Is there any way to make this statistic more accurate? There have been more than 2714 declined articles this month.—Anne Delong (talk) 12:57, 17 October 2013 (UTC)[reply]

  • Updated AfC/Submissions to reflect count of declined: I have made an initial update to that page, as more realistic, to also mention at least the online 57,000 AfC pages (which currently say "declined"), to begin a proper reporting of the counts. That page has received about 30 pageviews per day, so "over 57,000 declined" gives those readers a better ballpark estimate than 2700 declined. A long-term count of 120,000 declined, or 135,000, will require more analysis of the records. -Wikid77 (talk) 23:26, 17 October 2013 (UTC)[reply]

Can userspace additions affect Wikipedia?

Can addition of what I presume is Java code or something in user space be "extended" to affect Wikipedia outside of user space? Notice top of page. Student7 (talk) 19:05, 17 October 2013 (UTC)[reply]

And what's on top of page for you? Screenshot might help, a bit too vague so far to help... --AKlapper (WMF) (talk) 19:08, 17 October 2013 (UTC)[reply]
Well, if it's only Java, probably not; if JavaScript, maybe, but I would need to know what I'm looking for since I don't see anything. Chris857 (talk) 19:10, 17 October 2013 (UTC)[reply]
If you're talking about the topicons on DS's talk page (the question mark icons, etc.), that's not actually either Java or Javascript; that's just standard wikimarkup, and it is benign. As to the general question, no, it's not possible for Java source code to be run from a Wikipedia page; Java code needs to be compiled before it's run, and can only run inside what's called a virtual machine, so that's not really possible through Wikipedia. Javascript is a different story (and an entirely different language), but it should still be impossible in most cases for Javascript in userspace to affect things outside userspace; you have to explicitly give permission for Javascript contained in the wikitext of a page to be executed, and it's not something that you can easily do accidentally. And, as with AKlapper and Chris, I don't see either kind of coding on DS's page, so I wouldn't worry about it. Writ Keeper ⚇♔ 19:31, 17 October 2013 (UTC)[reply]

Collapsible sidebox template

Hi,

I hope this is the correct place to ask this. I am looking to create a collapse sidebox akin to the campaign box template belowalongside. I have conducted several searches, but have been unable to find any blank templates that describe how to do something similar to this. I am looking to create a similar box for a series of related diplomatic articles. Regards.EnigmaMcmxc (talk) 03:15, 18 October 2013 (UTC)[reply]

The {{Campaignbox Normandy}} template is an invocation of {{Campaignbox}} which is itself a specialised form of {{military navigation}}. If your intended use is not military, these won't be suitable for you, and the next level inward is {{navbox}} which is truly general-purpose. Apart from {{Campaignbox Normandy}} itself, all of these have documentation where the various parameters are described. There is more information on navboxes - both the narrow type used in sidebars and the full-width type used at the bottom of articles - at WP:NAVBOX. --Redrose64 (talk) 06:31, 18 October 2013 (UTC)[reply]

Notification glitch

Tracked in Phabricator
Task T52082

I just got a notification about this edit 'Jeremyb mentioned you on Kudpung talk page in "Spacing in your edits"', even though I'm not mentioned at all. Has anyone got any idea why a notification was triggered, or is this a bug? — Mr. Stradivarius ♪ talk ♪ 04:02, 18 October 2013 (UTC)[reply]

I was also pinged and upon responding I did see an error message regarding a template loop, suspecting that triggered the ping. Cheers.—John Cline (talk) 04:48, 18 October 2013 (UTC)[reply]
Why is this tracked as a bug? I see it as more of a feature. Ginsuloft (talk) 18:20, 18 October 2013 (UTC)[reply]

Automatic taxoboxes

I'm almost certain I'm not asking this at the right place, but I still hope you can help me. I noticed that certain taxoboxes are severely flawed (example: Gaviiformes - Gaviomorphae is invalid and should be replaced with Aequornithes). However, most articles on biological species or clades use "automatic taxoboxes" - so I can't edit them. Are these automatic taxoboxes new? I'm certain I didn't encountered them last year. So, how the hell can I edit automatic taxoboxes?

DaMatriX (talk) 14:04, 18 October 2013 (UTC)[reply]

They've been around for about three years, I think. You're probably best off asking at Template talk:Automatic taxobox. --Redrose64 (talk) 17:09, 18 October 2013 (UTC)[reply]
I second the above, and here is a clue: Template:Automatic_taxobox/doc/new#Parameter_.7Ctaxon.3D explains that the taxonomy box in question can be edited at Template:Taxonomy/Gaviiformes (or, in general, Template:Taxonomy/taxon, where "taxon" is the species or clade name.
Once you get there, you may end up down the rabbit hole trying to create a new Template:Taxonomy/Aequornithes and trying to get that linked up with the Gaviiformes template. Before you do all of that, you might want to check on the Talk page for Gaviiformes or for Automatic taxobox to see if there is consensus (scientific or otherwise) for making this change. – Jonesey95 (talk) 17:52, 18 October 2013 (UTC)[reply]

Pages created by....on Toolserver

Please see Pages created by on Toolserver. Has this gone over to Labs? Currently, this only gives the top 100 pages. At the bottom "To see all results, please go to this link. Unfortunately, that just seems to hang and go nowhere. Is there a new tool for this at Labs? — Maile (talk) 16:43, 18 October 2013 (UTC)[reply]

I saw a similar tool on labs Here Werieth (talk) 17:25, 18 October 2013 (UTC)[reply]
Close. Except that it says I created 1295 articles. I figure an accurate count to be somewhere between 550 and 600, counting new articles, redirects, dabs and categories. I hope Labs perfects it a little. This tool lists every talk page I ever created, whether that was to accompany my newly created article, a vandalism warning on an IP page, manually created archive pages to existing talk pages, or any talk page I created to assess an article that I otherwise had nothing to do with. And it also seems to be counting my skins where I set up .js or .css. — Maile (talk) 18:04, 18 October 2013 (UTC)[reply]
It is exactly what you asked for, every page that you ever created. Werieth (talk) 18:07, 18 October 2013 (UTC)[reply]
Technically, true. But what was over at Toolserver that TParis had set up, when it worked a couple of months ago came up with 521 articles. It didn't count the talk pages and what I had created for skins. Is this one still being developed? Maybe there could be a filter of some sort on what we want to see and what we don't. — Maile (talk) 18:24, 18 October 2013 (UTC)[reply]
@Maile66: try toollabs:/xtools/pages/. It should be what you're looking for. Clicking the link at the bottom of that page says you've created 525 articles, but takes a minute or two. 64.40.54.174 (talk) 05:24, 19 October 2013 (UTC)[reply]

VisualEditor weekly update - 2013-10-17 (MW 1.22wmf22)

Hey all,

Here is a copy of the weekly update for the VisualEditor project. This is to make sure you you all have as much opportunity to know what is happening, tell us when we're wrong, and help guide the priorities for development and improvement:

VisualEditor was updated as part of the wider MediaWiki 1.22wmf22 branch deployment on Thursday 17 October. In the week since 1.22wmf21, the team worked on some feature changes, infrastructure improvements, and fixing bugs.

We improved media item resizing in a few ways. Firstly, inline media items can now be resized in the same way that has been possible with block ones (like thumbnails) before. When resizing a media item, you can see a live preview of how it will look as you drag it (bug 54298). Whilst you are dragging an image to resize it, we now show a label with the current dimensions (bug 54297). Once you have resized it, we fetch a new higher-resolution image for the media item if necessary (bug 55697). Manual setting of media item sizes in their dialog is nearly complete and should be available next week.

Beyond this work, there was a lot of preparatory work done to make it possible to re-use elements of VisualEditor outside of the full VE surface (such as in other extensions or gadgets), which should be completed soon, and adding support for switching extra parts of VisualEditor on in the forthcoming mw:Beta Features extension.

A complete list of individual code commits is available in the 1.22/wmf22 changelog, and all Bugzilla bugs closed in this period are on Bugzilla's list.

Following the regular MediaWiki deployment roadmap, this should be deployed here (for opted-in users) on Thursday 24 October.

Hope this is helpful! As always, feedback on what we're doing is gratefully received, either here or on the enwiki-specific feedback page. Please ping me using {{ping|Jdforrester (WMF)}} to make sure I see it promptly if you have any thoughts or corrections.

Jdforrester (WMF) (talk) 19:32, 18 October 2013 (UTC) [reply]

geoiplookup.wikimedia.org

I don't know what https://geoiplookup.wikimedia.org/ is but what is wrong with it? My browsers wait ages for it till they time out. I get the impression it is happening on every WP page. Nurg (talk) 21:53, 18 October 2013 (UTC)[reply]

I'm getting the exact same thing. Basically every page, or when I click preview. Jevansen (talk) 22:11, 18 October 2013 (UTC)[reply]
I'm getting this too, and I'm seeing a pattern in our locations. --Closedmouth (talk) 05:28, 19 October 2013 (UTC)[reply]

Hi. Could someone help me with {{Gallery sequence}}? It's largely a direct copy from fr:Template:Animation... Having a bit of difficulty in making it work here... Sandbox is here. Rehman 04:25, 19 October 2013 (UTC)[reply]

Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(technical)&oldid=577810881"