Wikipedia:Village pump (technical)/Archive 46

Something weird in our neighbourhood
on Wikipedia Signpost/2008-08-25/News and notes some of the wikilinks are not visible in the project page although you can see they are there in the edit screen.

Here is the raw wikitext:

* The Lithuanian Wiktionary has reached 90,000 entries.


 * The Lithuanian Wiktionary has reached 60,000 articles.


 * The Georgian Wiktionary has reached 100 entries.

This then comes out as:
 * The has reached 90,000 entries.
 * The has reached 60,000 articles.
 * The has reached 100 entries.

I've tried this with firefox 3.01 and Safari Version 3.1.2 (4525.22) on Mac OS X 10.4.11 (8S2167). —Preceding unsigned comment added by Filceolaire (talk • contribs) 09:16, 31 August 2008 (UTC)


 * Add a colon before the language prefix, or it will show up on the left. --NE2 09:33, 31 August 2008 (UTC)

Database error
I keep getting:
 * A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was: (SQL query hidden) from within function "ExternalStoreDB::store". MySQL returned error "1030: Got error 136 from storage engine (10.0.2.107)".

-- User:Docu
 * I got that one too earlier today. After a complicated edit too - luckily Refresh pulled it out of the dumpster. See my post above - I'm sure the servers are hungover today, or at least one of them needs a drug test. Franamax (talk) 10:35, 31 August 2008 (UTC)
 * I concur. If they are shown to be guilty then they should have their medals stripped from them! It seems to have suddenly started working again in the last few minutes. I reported it at the help desk The Bald One      White cat 11:39, 31 August 2008 (UTC)
 * This just started for me too. older ≠ wiser 10:50, 31 August 2008 (UTC)
 * Me too (if this saves). Carcharoth (talk) 10:58, 31 August 2008 (UTC)
 * Again - 10.0.2.107 needs euthanizing it seems. Franamax (talk) 11:06, 31 August 2008 (UTC)


 * See https://wikitech.leuksman.com/view/Server_admin_log --- User:Docu

Me too
A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:

(SQL query hidden)

from within function "ExternalStoreDB::store". MySQL returned error "1030: Got error 136 from storage engine (10.0.2.102)".

I received this error when trying to save User:^.^/Status. It has not occurred again. Just thought I might report it anyway. — ^.^ &#91;citation needed&#93; 11:33, 31 August 2008 (UTC)
 * on the same note I keep on receiving similar errors. They occur on random pages that seem to have no link. Normally happens to me when trying to save a page.  ·Add§hore·  T alk /C ont 11:50, 31 August 2008 (UTC)
 * That happened to me when I was trying to add a CSD tag when Twinkle wasn't working. --frogger3140 (talk) 15:40, 31 August 2008 (UTC)
 * I got it an hour or two ago while simply trying to post to a talk page. Luckily I always copy to the paste buffer before I hit "save". Baseball Bugs What's up, Doc? 16:17, 31 August 2008 (UTC)
 * If that happens on Firefox and you've just pressed "Save" to save content, you can just press F5, and nothing is lost. D.M.N. (talk) 16:33, 31 August 2008 (UTC)
 * I've encountered this error twice today, but never before. SharkD (talk) 17:57, 31 August 2008 (UTC)
 * I've been here for about 3 1/2 years and this is the first time I recall seeing it. I'll get "page not found" from time to time, but not this. Baseball Bugs What's up, Doc? 19:41, 31 August 2008 (UTC)

Database Error - Page Not Found
I get a database error where the database "cannot find a page that it should find" at this URL I think the problem is the old ID is missing from the database.

I'm not very sophisticated at the working of Wikipedia but the missing page relates to a very hot topic and I am concerned that someone may have manipulated the database to delete the page. Can this be checked? Tom94022 (talk) 19:39, 31 August 2008 (UTC)


 * It looks like your URL is missing a '4' at the end. That would give http://en.wikipedia.org/w/index.php?title=Wikipedia:Suspected_sock_puppets/Thunderbird2&diff=prev&oldid=234474144 which is the page creation. PrimeHunter (talk) 22:48, 31 August 2008 (UTC)

Lupin's Anti-vandal tool
Hey guys. I installed this tool on my monobook.js a few months ago. To see the changes, I always had to bypass my browser cache ...however, no matter how many times I do that now, I am unable to use the non-admin rollback button, even though it worked before. It always says "Internal Server Error" and when I check the error report, it says "Line 275; Char: 2; Undefined is null or not an object; Code: 0". I don't know what sort of junk that's supposed to mean, but I DO know that the error might be a bug or something that affected User:Lupin/recent2.js. I know that this tool is "out-dated", but I use it anyway and it works on any computer in, say, the library. Does anyone know how this works? Thanks, ~ Troy (talk) 20:20, 31 August 2008 (UTC)

...PS: normal admin rollback works with me only sometimes as I was just recently granted the rollback function (it was always like that). Whenever it doesn't work, it asks me to use non-admin rollback, so that's where the issue is noticeable. ~ Troy (talk) 20:36, 31 August 2008 (UTC)


 * I gave Troy some suggestions here, but I don't know if that's done it or not.--chaser - t 02:57, 1 September 2008 (UTC)

Toilet history problem
I was wondering if anyone can give an explanation of what happened on the article Toilet. Due to Gwarp moves I move protected the article. A further attempt to move the article failed but the article move command has appeared in the article history as per diifs. Why has an entry been made in the history? Keith D (talk) 21:41, 31 August 2008 (UTC)
 * It wasn't a move. It was an edit which added a space to line 1 and made an edit summary which made it sound like a move. PrimeHunter (talk) 22:38, 31 August 2008 (UTC)
 * Thanks - I should have looked closer - too much Grawp at the moment, time for bed. Keith D (talk) 23:01, 31 August 2008 (UTC)

Enable &amp;
The standard &#91;X&#93;HTML elements  and   have been disabled by MediaWiki for no particular reason that I can conceive of. These are perfectly valid elements with with clear semantic meanings. Right now, people are either abusing   (which has a different semantic significance), or misusing   (which has no semantic significance and exists only for futzing with the presentation, and while not officially deprecated by W3C yet has, like   and , been effectively deprecated by the Web development community at large for around a decade now) in order to work around these Wiki-missing elements. For CSS purposes,  should probably be styled exactly as MediaWiki styles , and   identically to   but boldfaced. NB: They are all inline, not block, elements. Just to be clear, they should not be simply "allowed" and then translated into  or anything else, but should appear in the post-processed source as received by the browser just like   does, since they have independent semantic significance beyond their display style. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 23:22, 31 August 2008 (UTC)


 * After looking up what these elements are used for, I agree that both  and   should be enabled. Both of these look quite useful, and overuse of   certainly can make it hard to distinguish source code, user input, and user output.  Pyrospirit  ( talk  ·  contribs ) 23:53, 31 August 2008 (UTC)

—Simetrical (talk • contribs) 00:41, 1 September 2008 (UTC)

Title blacklist affecting my userpage
I can't seem to create any pages in my userspace as the username ^.^ is in the page title blacklist. Is it possible to rectify this problem so I can create pages in my userspace (the ones that are already there were moved over by a crat when I got renamed). At the very least, could a kind admin move User talk:HappyMan/Archive 6 to User talk:^.^/Archive 6. (HappyMan is my ASCII compatible account). Thanks — ^.^ &#91;citation needed&#93; 00:12, 1 September 2008 (UTC)


 * There seems to be a way to set individual entries on the title blacklist to apply only to editors who aren't autoconfirmed. Alternatively, there's also the handy-dandy Title whitelist if some helpful admin speaks whatever programming language these two are written in. Talk page archive moved.--chaser - t 00:19, 1 September 2008 (UTC)
 * Whitelisted. Mr.Z-man 00:29, 1 September 2008 (UTC)

Is 'Random article' truly random?
I've been meaning to ask this for the last couple years; the 'Random article' algorithm isn't really random, is it? If I click it, say, 50 times, I get several communes in France, but no villages elsewhere. Or I might get several articles on genes, but no athletes. The next day, the situation might be reversed. The likelihood of this happening is very low, so I'm thinking it uses an article's ID number (given to it when it was created?), and since there are clusters of articles are created by bots in a short period of time, one draws "too many" from one these clusters. Please, don't comment on probability unless you happen to know the algorithm 'Random article' uses. Thanks, 129.252.106.201 (talk) 04:01, 25 August 2008 (UTC)


 * Every article is assigned a random number between 0 and 1 when it is created (these are indexed in SQL, which is what makes selection fast). When you click random article it generates a target random number and then returns the article whose recorded random number is closest to this target.  Any clustering by content that you think you see is purely in the eye of the beholder.  Dragons flight (talk) 04:10, 25 August 2008 (UTC)


 * Technical_FAQ - Icewedge (talk) 04:14, 25 August 2008 (UTC)


 * In simple English, it's not totally random, but it's not biased towards anything in particular. --Carnildo (talk) 05:09, 25 August 2008 (UTC)


 * Or rather, it is biased, but randomly so, if that makes sense. For example. if you had three total articles, and they were randomly assigned the numbers .5 .51 and .49, you would see the latter two far far more often. Prodego  talk  21:25, 25 August 2008 (UTC)
 * It actually picks a random number and then chooses the page whose preselected random number is the closest one greater than the random number chosen. In the event that there's no suitable number found (if the random number is larger than all page_random values), it wraps around to the lowest values since .  This is no less random, but more efficient from a database point of view. —Simetrical (talk • contribs) 13:40, 26 August 2008 (UTC)

I've tested this and it seems to choose the one with the lower page_id if the page_random values are equal, which isn't implausible on large projects (see Birthday paradox). — CharlotteWebb 17:43, 26 August 2008 (UTC)
 * I got the same article again when I re-randomed a few days ago (I set up the kids' browser with a random article as home page). At one point it was set to reject RamBot articles, as they made up too large a % of the article count. Rich Farmbrough, 01:57 27 August 2008 (GMT).
 * But Rich, that's our most reliable content... — CharlotteWebb 02:10, 27 August 2008 (UTC)
 * page_random is a double, which in MySQL apparently means 53 precision bits. The birthday paradox implies that you'd need on the order of sqrt(n) different articles to cause a collision, instead of the order of n.  In our case, this should mean on the order of 225.5, or about 47 million articles, before you'd expect a collision.  This is very rough reasoning, and is possibly off by an order of magnitude, but there are likely going to be few if any collisions even for quite large numbers of articles, even with the birthday paradox.  This is good, because if two pages share page_random, one will never turn up using the random page feature ― although it's not like that cycles through all pages anyway, so it hardly matters. Anyway, the actual sorting order is undefined when page_randoms are equal.  Since pages will normally be added to the index in the order of page_id, it's quite likely that the database will store the page with the lowest page_id first in the index leaf, and will then incidentally return it first when it needs to return the first result with that page_random.  Depending on implementation, it could also easily be in descending page_id order, or could randomly alternate between orders.  There are probably cases in InnoDB where the higher page_id will be returned. —Simetrical (talk • contribs) 22:44, 27 August 2008 (UTC)
 * 253 seems to be about 9 x 1015 but I only saw 12 decimal places (using phpMyAdmin), so maybe my test was flawed by copying and pasting a truncated value (oops?). — CharlotteWebb 15:24, 28 August 2008 (UTC)
 * Any conversion to decimal will probably lose precision in at least some cases. phpMyAdmin is probably truncating it.  In any event, I did find a few thousand page_random collisions on enwiki, so my estimate was somewhat high (but not drastically).  I didn't expect it to be good to within more than about an order of magnitude, and it was less than an order of magnitude higher than the number of pages enwiki has, so that's no surprise. —Simetrical (talk • contribs) 00:26, 1 September 2008 (UTC)
 * Hmm, it may be helpful to check for collisions like this every month or so and assign them new values for page_random so that they have a non-zero chance of being seen. — CharlotteWebb 13:43, 1 September 2008 (UTC)
 * No point in doing it every month, it should be at page creation. But there's not really much point in it.  No page has a significant chance of being seen anyway. —Simetrical (talk • contribs) 21:07, 1 September 2008 (UTC)

Article sub-pages
Can pages in article space have subpages (such as Water melon/subpage), or are the forward slashes automatically converted into textual data? SharkD (talk) 04:49, 28 August 2008 (UTC)
 * No real subpages in the mainspace, sorry. Subpage-enabled namespaces are marked as such here. WODU P  04:55, 28 August 2008 (UTC)
 * You can use slashes fine, it's just that the result is not considered a subpage. More info: Help:Subpage, mw:Help:Subpages. —AlexSm 05:05, 28 August 2008 (UTC)
 * I see that "hierarchical organisation of articles is discouraged." I was thinking this might be useful in cases where articles are too long, but the "split" articles don't meet notability requirements, and are instead supposed to inherit the notability of the parent article. Oh well. SharkD (talk) 05:21, 28 August 2008 (UTC)
 * Years ago, this feature was enabled in the main namespace, and in Wikipedia's early years, it was a common way of organising information, but it led to awkward titles and the creation of article forks written from a particular POV, so it was decided to abandon the subpages system. Warofdreams talk 08:59, 28 August 2008 (UTC)
 * Do you have an example of an article where you believe "notability" is a concern? — CharlotteWebb 14:37, 28 August 2008 (UTC)
 * Not off the top of my head. But I remember the issue being brought up a couple of times in WikiProject Video games and in AfD discussions. SharkD (talk) 03:26, 1 September 2008 (UTC)
 * It seems there's a new article on the topic. SharkD (talk) 04:31, 1 September 2008 (UTC)
 * Let me rephrase that, do you have any examples of sub-pages, or distinct topics begging to be split onto separate pages, which may be in need of a more canonical title? — CharlotteWebb 13:49, 1 September 2008 (UTC)
 * You may want to take the issue to the discussion page I linked to (also, see here). I haven't been following along, but the whole point of the discussion seems to revolve around whether some articles may inherit the notability of the parent topic (or, rather, whether rules can be bent in some cases). I don't particularly like the idea, but I am suggesting that sub-articles may potentially constitute a separate class (i.e., not a "real", self-contained article) where this is the case. SharkD (talk) 16:27, 1 September 2008 (UTC)
 * Even if there was a consensus to use file-directory style titles, I don't see how the "notability" guideline would be interpreted differently. On the other hand "notability guidelines and common sense" go together about as well as "water and oil", "vinegar and baking soda", or "bleach and ammonia" (as opposed to "fish and chips", "peaches and creme", or "vodka and orange juice"), so really g-d knows what would happen in that case. Seriously though would this mean Christianity in the United States would be under "United States/Religion/Christianity" or under "Christianity/Places/United States", or under something else I haven't thought of? — CharlotteWebb 16:52, 1 September 2008 (UTC)
 * Denoting subpages by using backslashes is ponderous and unecessary. I am not a database expert, but my understanding is that the hierarchical taxonomy being described by having subpages (e.g. Hamlet/Prince Hamlet) is already dealt with by categories such as Category:Characters in Hamlet. In my view, the proposal to have subpages is just a thinly veiled attempt to assert that a subject without reliable secondary sources inhertits notabiility from a more notable subject, which is just not the case.--Gavin Collins (talk) 09:28, 1 September 2008 (UTC)
 * In this particular case Hamlet (character) would probably be preferable. — CharlotteWebb 13:49, 1 September 2008 (UTC)
 * I don't quite see the connection in what you're saying. The existence of the article, Hamlet (the play, not the character), does not lend itself to the categorization scheme you suggest. I.e., placing the Hamlet article in Category:Characters in Hamlet isn't particularly sensible. SharkD (talk) 16:27, 1 September 2008 (UTC)
 * Alright so maybe you mean to use Hamlet/Act 1 instead of Act 1 of Hamlet, Hamlet: Act 1, Hamlet, Act 1, Act 1 (Hamlet), or Hamlet (Act 1). I don't see how this would make things any easier or why anyone would expect it to have any effect on "notability" guidelines. — CharlotteWebb 16:52, 1 September 2008 (UTC)
 * One area where I could see such a thing used is in List-type articles. They're not really articles (they have no prose, and don't explain anything), yet they have some encyclopedic value. For instance, a list of things in Topic X could appear in X/List of Y in X. However, I wouldn't want to see this done, either. I'd rather see a separate namespace be used for lists (I think this would be a good idea) than resorting to this option. SharkD (talk) 19:30, 1 September 2008 (UTC)

Lots of changes to "wikibits.js"
I've made lots of changes to "wikibits.js". The edited version can be found here. The changes to the file are described here. You can experiment with the code by copying it to your "monobook.js". The changes include several improvements to sortable tables. I'd appreciate some feedback. SharkD (talk) 22:33, 29 August 2008 (UTC)
 * I posted some feedback on User talk:SharkD/Sandbox/wikibits 4 as I could not work out how to test it. -84user (talk) 10:37, 30 August 2008 (UTC)
 * Thanks for the feedback! I replied to your comments on the talk page. SharkD (talk) 11:05, 30 August 2008 (UTC)

The bugzilla entries (including patches) for this(-ese) proposals can be found here:. SharkD (talk) 03:13, 1 September 2008 (UTC)
 * One more: . SharkD (talk) 16:17, 1 September 2008 (UTC)
 * Another: . SharkD (talk) 06:00, 2 September 2008 (UTC)

Image category help
How do I place images like Image:FreeCol colony screen.jpg into a category? Will "creating the page" cause problems? Thanks! SharkD (talk) 11:42, 30 August 2008 (UTC)
 * You can do it, but I think WikiCommons would be a better place to created categories of images.--Appraiser (talk) 15:09, 30 August 2008 (UTC)
 * Well, the nice thing about having the categories be on Wikipedia is that they appear along side the articles when viewing the category. I.e., readers viewing the category may find images they hadn't looked for and be pleasantly surprised. SharkD (talk) 15:22, 30 August 2008 (UTC)
 * You can link to the commons categories, see commonscat (not "common scat" mind you). — CharlotteWebb 15:49, 30 August 2008 (UTC)
 * If I use commonscat for category B (located on commons), will this cause all the images in that particular category to appear in category A (located on en:WP)? This is what I'm trying to achieve. SharkD (talk) 20:33, 31 August 2008 (UTC)
 * No, this is just the standard way of linking to a commons category from a project other than commons. Attempting to categorize commons images locally on each project is redundant and confusing (as there is no corresponding local content). — CharlotteWebb 13:52, 1 September 2008 (UTC)

Template/CSS help
I'm trying to achieve this effect in this article using this template. There are two problems: 1) it only works in Internet Explorer, AFAIK; 2) it screws up the alphabetic sorting of the table column. I was hoping someone here might have a solution. Thanks! SharkD (talk) 20:28, 31 August 2008 (UTC)
 * It works fine in FF3, Opera and Safari. You could use a hidden sortkey to sort out the sorting. Algebraist 20:33, 31 August 2008 (UTC)
 * Ah, sorry I didn't specify the issue in better detail! It's been a few weeks since I actually worked on the article/template.
 * The problem in Firefox is when the cell's text is too long. In Firefox a line-beak occurs between the audio link and the cell text, causing the cell text and audio link to appear on different lines. In IE the cell text instead wraps when it comes into contact with the audio link. To see this problem in effect, take a look at the thirteenth row in the table. (You may have to decrease the size of the browser window first, as the problem will not occur if the browser window is very wide.) I'm not sure which browser is technically working "correctly", but the IE behavior is the one that is desired. SharkD (talk) 20:45, 31 August 2008 (UTC)
 * According to the template docs, if there are line-break/overlap issues you should specify the widths for each span using the optional parameters,  and  . Algebraist 23:02, 31 August 2008 (UTC)
 * Hehe. I wrote the doc. Anyway, I tried this and another user reverted it (it only works when the table column has a width specified for it), so I'm looking for a better all-around solution. SharkD (talk) 03:11, 1 September 2008 (UTC)
 * Why not just put the text and link in different columns? Algebraist 13:54, 1 September 2008 (UTC)
 * Putting the links in another column is ugly, and would result in the table tabulating metadata instead of only data. That said, maybe the borders between the two columns could be collapsed, but I think this would be more trouble than it's worth. SharkD (talk) 16:56, 1 September 2008 (UTC)

template help
Can someone help me add a parameter to icu-saved? I'd like to be able to add "Nested=yes" so that it can be collapsed like regular WP tags. The WICU project has seveal tags but this one seems to be the one that stays on articles however they do more of a"treat and release" concept so the regular project banner doesn't seem as approprite in the cases of articles that have been "released". Banj<font color="#FF4400">e <u style="font-family: Zapfino, sans-serif;color:#8000FF">b<font color="#CC0000">oi 03:44, 1 September 2008 (UTC)
 * Could you provide an example of a template that features the desired behavior? SharkD (talk) 04:04, 1 September 2008 (UTC)
 * Template:WPBannerMeta. <u style="text-decoration:none;font-family: papyrus;color:#CC00CC">Banj<font color="#FF4400">e <u style="font-family: Zapfino, sans-serif;color:#8000FF">b<font color="#CC0000">oi  06:48, 1 September 2008 (UTC)
 * Done. Anomie⚔ 11:44, 1 September 2008 (UTC)
 * Yea! Thank you!!! <u style="text-decoration:none;font-family: papyrus;color:#CC00CC">Banj<font color="#FF4400">e <u style="font-family: Zapfino, sans-serif;color:#8000FF">b<font color="#CC0000">oi  02:38, 2 September 2008 (UTC)

WP:BLPN
Something very odd has happened here. I did an edit. and the bottom half of the page has vanished, along with its edit history from what I can make out. This is what it looks like now  Ty  10:00, 1 September 2008 (UTC)
 * Now it's suddenly all reappeared again - the rest of the page and the related part of the edit history.  Ty  10:03, 1 September 2008 (UTC)

And now it's half vanished, along with the edit history related to that part. Also talk page is a red link. Here's the edit history with edits missing:


 * (cur) (last) 10:08, 1 September 2008 Tyrenius (Talk | contribs | block) (161,385 bytes) (→Talk: John Michell (writer): heading level from 2 to 3) (rollback | undo)
 * ((cur) (last) 09:50, 1 September 2008 Tyrenius (Talk | contribs | block) (161,383 bytes) (rv self - something not working here) (undo)
 * ((cur) (last) 09:48, 1 September 2008 Tyrenius (Talk | contribs | block) (160,083 bytes) (move post down to related post) (undo)
 * ((cur) (last) 03:46, 31 August 2008 Privatemusings (Talk | contribs | block) (161,383 bytes) (→Ongoing WP:BLP-related concerns: add danny) (undo)
 * ((cur) (last) 00:37, 20 August 2008 SageMab (Talk | contribs | block) (161,165 bytes) (Talk: John Michell (writer)) (undo)
 * ((cur) (last) 00:30, 20 August 2008 SageMab (Talk | contribs | block) (159,865 bytes) (→Talk:living author: defaming continues) (undo)
 * ((cur) (last) 16:19, 15 August 2008 Ciupicat (Talk | contribs | block) (158,721 bytes) (→  ) (undo)
 * ((cur) (last) 16:18, 15 August 2008 Ciupicat (Talk | contribs | block) (158,721 bytes) (→Ongoing WP:BLP-related concerns) (undo)
 * ((cur) (last) 04:25, 15 August 2008 Edison (Talk | contribs | block) (158,356 bytes) (→Proposal for a vote) (undo)

There's too much that had vanished to show it all here, but here's a sample of what it looks like when it's working OK:


 * (cur) (last) 18:27, 15 August 2008 Wikidemon (Talk | contribs | block) (69,552 bytes) (→Mocking a BLP at Media Matters for America: evil steve jobs) (undo)
 * (cur) (last) 18:09, 15 August 2008 Noroton (Talk | contribs | block) (69,012 bytes) (→Mocking a BLP at Media Matters for America: wrong on all counts) (undo)
 * (cur) (last) 17:45, 15 August 2008 Gamaliel (Talk | contribs | block) m (68,332 bytes) (→Mocking a BLP at Media Matters for America) (undo)
 * (cur) (last) 17:44, 15 August 2008 Gamaliel (Talk | contribs | block) (68,331 bytes) (→Mocking a BLP at Media Matters for America) (undo)
 * (cur) (last) 06:36, 15 August 2008 MiszaBot II (Talk | contribs | block) m (67,987 bytes) (Archiving 1 thread(s) (older than 4d) to Wikipedia:Biographies of living persons/Noticeboard/Archive50.) (undo)
 * (cur) (last) 05:48, 15 August 2008 Wikidemon (Talk | contribs | block) (69,708 bytes) (→Mocking a BLP at Media Matters for America: comment) (undo)
 * (cur) (last) 04:25, 15 August 2008 Noroton (Talk | contribs | block) (68,971 bytes) (→Mocking a BLP at Media Matters for America: response to LaughingVulcan) (undo)

It is at the moment OK again.

 Ty  10:23, 1 September 2008 (UTC)

Page text not found
While looking for potential image copyvios in an upload log, I encountered the following on two images (Image:Jericho rosales.jpg & Image:Kristine hermosa.jpg): The database did not find the text of a page that it should have found, named "Image:Kristine hermosa.jpg".

This is usually caused by following an outdated diff or history link to a page that has been deleted.

If this is not the case, you may have found a bug in the software. Please report this, making note of the URL. Well, it isn't a diff or a history link, and there's nothing in the image logs to suggest they've been deleted. So I'm reporting it here as advised. Thanks. --62.164.255.93 (talk) 10:33, 1 September 2008 (UTC)

Might be relevant: —Simetrical (talk • contribs) 21:20, 1 September 2008 (UTC)

Important RFC on Notability!
Note that Wikipedia:Notability/RFC:compromise is now open for business and your comments would be very welcome. --Gavin Collins (talk) 10:43, 1 September 2008 (UTC)

Create protection failure
was created again, but the log shows that it was protected against creation on July 20, not to expire until Oct 20. What's up?Kww (talk) 15:40, 1 September 2008 (UTC)
 * It was created at You me at six and moved by Deb (who is an admin). Algebraist 15:44, 1 September 2008 (UTC)
 * Ah. Left a note with Deb pointing out her mistake.Kww (talk) 16:19, 1 September 2008 (UTC)

Registered in Wikipedia English, how to post article Wikipedia French
I am registered on the English Wikipedia site and have translated an article into French from an article I posted in English.

However, the French Wikipedia does not appear to recognise my username/login. Do I have to register again? IanS (talk) 18:02, 1 September 2008 (UTC)
 * See m:Help:Unified login. <font fmily="Comic sans"> Corvus cornix  talk  18:31, 1 September 2008 (UTC)

Thank you IanS (talk) 18:39, 1 September 2008 (UTC)

Move problem
Why can't I move 2004 Summer Paralympics medal count to 2004 Summer Paralympics medal table? When I try to do so it says "the title '2004 Summer Paralympics medal table' is protected from being created", but the protection log shows nothing. When I first tried this there was no article at the "medal table" title, so I tried creating a redirect there. It worked, so it's obviously not protected against creation. But I still get the same error message when I try to move the "medal count" article over the redirect. — jwillbur 20:20, 27 August 2008 (UTC)


 * This appears to be a persistent problem. I have come across the problem today with Sugar Hill Historic District (Detroit, Michigan), which was "protected against creation", but I was able to create a redirect. JPG-GR (talk) 19:32, 31 August 2008 (UTC)
 * And, again at Template Attribute Language Expression Syntax. JPG-GR (talk) 03:01, 1 September 2008 (UTC)

MediaWiki:Titleblacklist. &mdash; Werdna &bull; talk 08:13, 2 September 2008 (UTC)


 * Any idea which particular line on the blacklist is causing the problems? I couldn't see any, but then I'm not all that familiar with regexps. — jwillbur 15:22, 2 September 2008 (UTC)

Confirmation
I suggest that there be a confirmation message that appears when a user is not signed in and trying to edit, to confirm with the user that he/she wants to edit without signing in. -- IRP (talk) 03:09, 28 August 2008 (UTC)
 * There is:
 * You are not currently logged in. Editing this way will cause your IP address to be recorded publicly in this page's edit history. If you create an account, you can conceal your IP address and be provided with many other benefits. Messages sent to your IP can be viewed on your talk page.
 * Please do not save test edits. If you want to experiment, please use the sandbox.
 * 24.0.66.109 (talk) 03:13, 28 August 2008 (UTC)
 * 24.0.66.109 (talk) 03:13, 28 August 2008 (UTC)


 * I was suggesting a more noticeable popup message. -- IRP (talk) 03:18, 28 August 2008 (UTC)


 * It's a nice idea, but for those who choose to edit anonymously (and there are some people who have racked up enough Wikipedia experience to be admin material while on an IP address) it would get infuriating in a very short time. Confusing Manifestation (Say hi!) 04:16, 28 August 2008 (UTC)
 * How about making that message a bit more noticeable? A border or lightly colored background? <font color="#0000CC">W<font color="#000066">ODU P  04:28, 28 August 2008 (UTC)
 * Well, it's a balance between trying to get people who have inadvertently logged out to notice that they're logged out and scaring would-be anonymous contributors with a large notice about how they're not logged in. --MZMcBride (talk) 05:00, 28 August 2008 (UTC)

One possible solution is to make your personal logged-in experience a bit different, so you can immediately notice when you're logged out. For example,  in your monobook.css should make your "Save" button green. —AlexSm 05:14, 28 August 2008 (UTC)

Didn't we just have this discussion? Anomie⚔ 10:51, 28 August 2008 (UTC)


 * Ok, how about this: The same message that already exists will not pop up in a popup message, but it should be underneath the 3 buttons: Save Page, Show preview, and Show changes. The message should be highlighted in a noticable color. -- IRP (talk) 20:07, 28 August 2008 (UTC)


 * Yes, we already had this discussion: Village pump (proposals)/Archive 32. And there were some good suggestions there. Some might even be good as gadgets. -- John Broughton  (♫♫) 21:18, 28 August 2008 (UTC)


 * Thanks for the information -- IRP (talk) 21:21, 28 August 2008 (UTC)

How about having a different default skin color for logged in editors vs. editors who are not logged in? <font face="Comic sans"> Corvus cornix talk  22:13, 29 August 2008 (UTC)

Gadgets don't work for anons. &mdash; Werdna &bull; talk 08:11, 2 September 2008 (UTC)

Proposal: create MediaWiki:Wikibits.js
I propose that MediaWiki:Wikibits.js be created and serve as a central location for discussing this file. A page exists for MediaWiki:Common.js and probably other JS files as well. SharkD (talk) 04:48, 28 August 2008 (UTC)
 * I don't see why you would want to create a useless and misleading "local system message" to discuss internal MediaWiki file. On the other hand, a separate discussion place somewhere on Meta or mediawiki.org might make sense, provided that you advertise it on MediaWiki_talk:Common.js, WP:US, etc. —AlexSm 05:00, 28 August 2008 (UTC)
 * Well, there's nothing about the title of the proposed page that screams to me, "local system message!" However, if there is an existing benchmark for these sorts of things, then an alternative proposal would be appreciated. SharkD (talk) 05:13, 28 August 2008 (UTC)
 * All «MediaWiki:...» pages are local system messages by definition, so creating MediaWiki:Wikibits.js would give some users the wrong idea that this file can be changed just for English Wikipedia. —AlexSm 05:18, 28 August 2008 (UTC)
 * Ah! I see! For some reason I thought MediaWiki:Wikibits.js linked to a page on mediawiki.org (i.e., an interwiki link). SharkD (talk) 05:31, 28 August 2008 (UTC)

Issues with date formatting/sorting (described here) might be a good reason to make this file editable on a per-language wiki basis. SharkD (talk) 06:09, 28 August 2008 (UTC)

Is there enough discussion outside of bugzilla to warrant the creation of talk pages for all files that fall under this category? Some of them? Lots of them? I'd rather see a systematic change than see a single talk page orphaned somewhere in space... SharkD (talk) 04:00, 1 September 2008 (UTC)

We have a general discussion forum for software changes. It's called <tt>wikitech-l</tt>. It does not make sense to have discussion pages for <tt>wikibits.js</tt>, unless we also have them for <tt>User.php</tt>, <tt>Title.php</tt>, and so on. &mdash; Werdna &bull; talk 08:10, 2 September 2008 (UTC)

Global block overwrites local block?
Does a global softblock of an IP address overwrite a local hardblock of the same? --Oxymoron83 06:11, 1 September 2008 (UTC)
 * No. — CharlotteWebb 14:14, 1 September 2008 (UTC)

No global block will ever "overwrite" a local block. To say this implies that the local block is somehow removed. The effect of both a global block and a local block is additive. &mdash; Werdna &bull; talk 07:14, 2 September 2008 (UTC)

Please advise on tl-something templates
From what I know, we've got,  ,  ,   - when are we meant to use one, and when the other? Please help! It Is Me Here (talk) 16:38, 1 September 2008 (UTC)
 * Usage notes are provided at Template:Tl. – xeno  ( talk ) 16:42, 1 September 2008 (UTC)
 * I understand, but I don't find the instructions there very clear, to be honest. For instance, what is the difference between  and  ? It Is Me Here (talk) 17:53, 1 September 2008 (UTC)
 * Tl2 is for linking to other langs/projects, Tlx is so that you can show some parameters. – xeno  ( talk ) 18:57, 1 September 2008 (UTC)
 * OK, and what about ? I'm sure I saw it somewhere. It Is Me Here (talk) 07:18, 2 September 2008 (UTC)
 * From Template:tnull - Tnull is used to nullify templates in the monospaced font. It's used for stuff like helpme after you've helped the user, you nullify the template. – xeno  ( talk ) 10:23, 2 September 2008 (UTC)
 * I see - thanks! It Is Me Here (talk) 13:21, 2 September 2008 (UTC)

Page creation and deletion by namespace
Is there information available regarding one or more of the following? Since it's apparently possible to give an exact count of the number of non-orphaned articles (see Special:Statistics), it should also be possible to do the same with other namespaces. However, despite all my searching, I've not been able to find this information anywhere. –Black Falcon (Talk) 02:14, 2 September 2008 (UTC)
 * 1) The total number of pages in each namespace;
 * 2) The number of pages created in a given time period (e.g. day, week, month), broken down by namespace; and
 * 3) The number of pages deleted in a given time period, broken down by namespace.
 * See m:Help:Magic_words for the first (seems to be disabled). I'm not sure about the rest. - jc37 02:18, 2 September 2008 (UTC)
 * You can get a pretty good idea of #2 via Special:NewPages for everything but articles. You'll have to set the feed high and then re-set it to get an exact count, but if you're just interested in the rate, dividing a given number of pages by the time period will give you that. It's no good for articles because there are so damn many of them and they get speedy deleted so quickly. I haven't done enough speedy deletion outside the article namespace to know how quickly that's done or what effect it would have.--chaser - t 06:22, 2 September 2008 (UTC)

If there is a point to this question, you can ask somebody with a toolserver account to run a few database queries. &mdash; Werdna &bull; talk 08:15, 2 September 2008 (UTC)

Weird broken link.
Can someone explain what was wrong with this edit?  I reverted it because it breaks the link on the "music" icon, but I can't figure out why. Thanks. APL (talk) 14:25, 2 September 2008 (UTC)
 * The image has another original size so the coordinates only select the upper left corner which can be clicked. PrimeHunter (talk) 14:36, 2 September 2008 (UTC)
 * Aha! That did it. Thanks. I didn't realize it worked like that. APL (talk) 16:41, 2 September 2008 (UTC)

Looking for some listing tools
I'm not sure if this is the right place, but I figure the people here can point me in the right direction. Over at Featured topics we've been having a hard time keeping in sync the articles listed on that page and the articled tagged as being in an FT using ArticleHistory. To keep these two in line with each other, I would like to find: With these two lists I would be able to quickly find the disagreement between the two, which currently can take a very long time to track down. I have a feeling that someone has arleady devised tools to do each of these functions, so can someone please tell me where to find them? Thank you. --Arctic Gnome (talk • contribs) 18:00, 2 September 2008 (UTC)
 * 1) A way to see a list of all the articles in Category:Wikipedia featured topics all articles, its sub-categories, and sub-sub-categories in alphabetical order. This would let me see which articles are tagged with the ArticleHistory template.
 * 2) A way to see a list of all the articles linked on Featured topics in alphabetical order.
 * I did this the hard way using the API, Excel and Access. I was able to verify that all 438 articles currently match up. Having gone through the procedure once, it would likely take me 20-25 minutes to do again, which I'll gladly do on request. No it's not a listing tool, it's a series of manual steps - but I kept notes. If you do find someone to write you a tool, give them a hint that some of the FT page listings are redirects (currently 18 of the 438 don't match exactly) - so they will need to resolve the true (normalized) article name when they get that info. Franamax (talk) 05:07, 3 September 2008 (UTC)

English Wikipedia Internal Account Creation Interface
Hi guys! I'm trying to create an account on the English Wikipedia Internal Account Creation Interface. I entered in all of the information, clicked the second checkbox, skipped the signature box, and hit submit. I got something that looked like:

"Query failed: INSERT INTO acc_user (user_name, user_email, user_pass, user_level, user_onwikiname, user_secure, user_welcome, user_welcome_sig, user_welcome_template) VALUES ('NuclearWarfare', 'EMAIL WAS HERE BUT I REMOVED IT', 'HASH WAS HERE BUT I REMOVED IT', 'New', 'NuclearWarfare', '0', '0',, 'welcome-personal'); ERROR: Unknown column 'user_secure' in 'field list' "

Can anyone tell me what is wrong? I've read Wikipedia:Request an account/Guide over and I can't see what's wrong. Thanks for your help!

(I was told to cross post this over to here, forgive me if I put it in the wrong place) NuclearWarfare  <sup style="color:green;">contact me <sub style="color:purple;">My work  18:11, 2 September 2008 (UTC)
 * Hmm...that's interesting. I'll have to ask SQLDb or OverlordQ about that, as they run the SQL tables. <span style="font-family:Verdana,Arial,Helvetica;color:steelblue;">X clamation point  00:50, 3 September 2008 (UTC)
 * Give me a couple moments, and, I'll have it fixed up. SQL <sup style="font-size: 5pt;color:#999">Query me!  03:21, 3 September 2008 (UTC)
 * Should be fixed now. SQL <sup style="font-size: 5pt;color:#999">Query me!  03:43, 3 September 2008 (UTC)

Move copyright warning on the edit page
Please move the 'Content that violates any copyright will be deleted...' message above the edit pane. As it is now, the edit pane and summary field don't always fit in the browser window. I have to scroll every time I edit a page. SharkD (talk) 01:25, 3 September 2008 (UTC)
 * Try putting

body #editpage-copywarn, body #editpage-copywarn2, body #editpage-copywarn3, {    display: none !important; }
 * in your User:SharkD/monobook.css file. Also check out mine.  Hardly any boilerplate shows up for me anymore.  Saintrain (talk) 01:39, 3 September 2008 (UTC)

Excluding editors unlikely to vandalize from filtered recent changes
I often use Lupin's anti-vandalism tool to scroll recent changes for vandalism. The tool works great, but the feed is cluttered with perfectly legitimate edits and vandalism reversions that get caught b/c it operates with a bad-words list. One way I've often considered to exclude these is to create a whitelist of admins and others whose edits don't get picked up by the tool. The AWB checkpage and rollbackers would be a good start to such a list (this isn't by any means an effort to create a new "trusted" user class). Would it be possible to implement this somehow, as all Lupin's code appears to be in his userspace? (I would contact Lupin, but he or she hasn't edited in almost eight months.)--chaser - t 22:58, 31 August 2008 (UTC)


 * The User:Lupin/recent2.js tool grabs data from the Special:Recentchanges RSS feed, which currently allows you to ignore bots, anons, etc. with <tt>&hidebots=1</tt>, <tt>&hideanons=1</tt>, etc. in the url. Adding more options for other access levels e.g. <tt>&hideadmins=1</tt>, <tt>&hiderollbacker=1</tt>, <tt>&hideboardvote=1</tt>, etc. would be the most efficient way to ignore certain user-groups, because loading only the diffs you want to see is less wasteful than loading all of them and telling the script to disregard many of them, and would require only minimal changes to the script (namely the url it grabs from). Talk to the devs about this. — CharlotteWebb 14:12, 1 September 2008 (UTC)
 * Whether a change was made by a user who's anonymous or a bot is stored in the recentchanges table. Whether they're an admin, rollbacker, boardvote, etc. is not.  It would require a join to the user_groups table and would be a lot less efficient.  In some cases you'd probably have to scan the whole table, too: try searching for edits by developers . . . Bots and anons are "efficient enough" because you can expect a substantial but not overwhelming percentage of edits to be by each group.  So at most you should have to scan two or three or ten times as many rows as you really need, not thousands of times as many. —Simetrical (talk • contribs) 21:18, 1 September 2008 (UTC)
 * Okay, failing the first option, how about including the user-rights of each user whose edits appear in the RSS feed? Then the script would know which edits to assume good (ignore) without having to hit the API or whatever for each unrecognized name. — CharlotteWebb 22:34, 1 September 2008 (UTC)
 * It could be done, sure, if there's enough demand for it. —Simetrical (talk • contribs) 14:10, 2 September 2008 (UTC)

Did you read Simetrical's reply above? He clearly says that, when generating the recentchanges table, checking for those things is not conducive to reasonable performance. It could, of course, be included in the recentchanges table, but it seems like a bit of a big ask for minimal benefit. OTOH, I'm working on some new fields for recentchanges, so it could be snuck in there, presuming there was sufficient demand/need for it. &mdash; Werdna &bull; talk 06:20, 2 September 2008 (UTC)
 * It would be an issue to filter by it. It wouldn't be an issue to just provide the data.  One extra query on all the user_id's would do it.  (Joining would be a bad idea because it would repeat rows.) You couldn't include it in the recentchanges table unless you did something ugly like concatenating the group names, because users can have multiple groups, so it should really be in its own table.  Plus you'd have to maintain it, if it were denormalized like that, which would be a pain. —Simetrical (talk • contribs) 14:10, 2 September 2008 (UTC)
 * If I'm the only person asking about this, there's probably not sufficient demand (unless there's a super-easy way!). Anyway, thanks everybody for your time.--chaser - t 07:04, 3 September 2008 (UTC)

When editing a page, whitespace above editing box
When editing a page (that is, on the page you get to after clicking the "Edit this page" link), there is some whitespace above the editing box that wasn't there before today, which I think is caused by something like:

getting changed to:

Where was this change made?

—Lowellian (reply) 04:15, 1 September 2008 (UTC)


 * In EditPage.php: look at the comments to 40184 and maybe some other revisions as well. —AlexSm 04:24, 1 September 2008 (UTC)


 * Thank you for the info, AlexSm. But this is not cool at all; that unnecessary whitespace is very annoying and shoves the edit box down too far. As far as I can tell from looking at the revisions, under the previous code, the




 * part was only generated when there was actually a Preview DIV with content in it (which makes sense, because if there is a Preview block, we want there to be some space between the Preview block and the edit box for visual clarity). But now, that whitespace is generated even when there is no Preview (that is, you haven't clicked the Preview button, so there is only a Preview DIV with no content). This is unnecessary whitespace that looks rather glaring. Any idea how to get the attention of the developers to fix this, or alternatively, how I can modify my monobook.css to get rid of the whitespace?


 * —Lowellian (reply) 19:36, 1 September 2008 (UTC)

Saying "I don't like it, for the following reasons" would be sufficient to express your displeasure with the change. &mdash; Werdna &bull; talk 07:13, 2 September 2008 (UTC)


 * Yes, but who should I contact to get this changed? Or like I said, alternatively, how I can modify my monobook.css to get rid of the whitespace? —Lowellian (reply) 17:40, 2 September 2008 (UTC)


 * It seems the extra whitespace has been removed by now, exactly by revision that I linked above. —AlexSm 15:17, 3 September 2008 (UTC)


 * Oh, okay, I see, the fixed revision just took a few days to take effect. That's great news. Thank you, developers; I really appreciate this. :) —Lowellian (reply) 23:32, 3 September 2008 (UTC)

NavPopups problem
Hi there. I recently installed the NavPopups script in my .js and everything was fine after the first basic install, but after then adding customisation options they have stopped appearing altogether. Where have I gone wrong? It Is Me Here (talk) 15:54, 3 September 2008 (UTC)
 * Are you getting any javascript errors (Tools/Error console in Firefox)? Algebraist 17:59, 3 September 2008 (UTC)
 * You missed out the equals sign in the 'popupDragHandle' line. I don't know if that could be causing the problem. Algebraist 18:00, 3 September 2008 (UTC)


 * Yep, that was it - silly me! Thanks for the help! It Is Me Here (talk) 18:02, 3 September 2008 (UTC)

AOL XFF problem
See Administrators%27_noticeboard/Incidents. It appears that XFF information from 64.12.0.0/16 is no longer trusted, so editors using AOL are affected by a range block placed to prevent anonymous IP hopping vandalism. My understanding is that if the xff headers were trusted, the edits would be recorded as coming from the editor's home ISP rather than AOL and they would not be affected by the block. Thatcher 16:13, 3 September 2008 (UTC)
 * Strictly from a layman's POV, AOL does own the whole 64.12/16 block but only 64.12.96/24 seems to be trusted-XFF. I also note here that AOL was marked (long ago) as not sending XFF headers consistently. Seems like an issue with how AOL partitions its address space and forwards XFF headers. Don't know if that helps at all. The revision history of the XFF-trusted file would be instructive, but I see no way to figure that out from where I sit. Franamax (talk) 05:16, 4 September 2008 (UTC)
 * And from the XFF page on Meta, it looks as though "Contact xff AT wikimedia DOT org for listing and delisting inquiries" is where you may wish to ask. Franamax (talk) 05:19, 4 September 2008 (UTC)

media player doesn't work on Internet Explorer
The media player applet doesn't seem to work on Internet Explorer. This appears to be the case for both IE6 and IE7. When I try to play something on IE, the message "class com.fluendo.player.Cortado not found" appears and the applet does not load.

Does anyone know how to resolve this? --Ixfd64 (talk) 05:35, 4 September 2008 (UTC)

[Mark this page as patrolled] is not appearing...
I'm trying to mark pages as patrolled through Special:NewPages but the option to is not appearing on the articles. There is an rcid in the url and the articles arn't listed in Special:Log/patrol. Can anyone help me please? – Jerry  teps  05:49, 4 September 2008 (UTC)

Edit summary not working properly
When you change an edit summary to document changes to more than one section (as described at Edit summary), only the last section name enclosed in "/* */" will be properly parsed. See here and here. This used to work properly. I am using Firefox 3.0.1 with several javascript apps in my monobook.js (I don't think this would be the problem, but I don't know for sure). — OranL (talk) 20:01, 2 September 2008 (UTC)


 * I think this is an output issue, not really a problem with using summaries in that manner: I archive the steward permissions page regularly, and I'm seeing edit summaries with that issue going back to July. The problem is, I recall them working correctly at the time. I suspect there's been a recent change to mediawiki that's causing the output to not display the way we're used to. Kylu (talk) 01:54, 3 September 2008 (UTC)


 * I noticed that too, and I liked the old way: sometimes I feel like adding replies to several section with just one edit. Look at the history of Linker.php, this was probably 39373 by Brion that broke multiple /*section headers*/. By the way, another recent change is 39600 do-not-escape-entities, so if you put   in some old edit summary now it is displayed as space. —AlexSm 03:58, 3 September 2008 (UTC)


 * I'm going to try adding a bug to MediaWiki Bugzilla. Has anyone not experienced the bug? I'm trying to find out if all configurations are affected, or if it's specific to Firefox or Windows (both of which I highly doubt, but I want to make sure). — OranL (talk) 23:25, 3 September 2008 (UTC)


 * That's a bug in how mediawiki's rendering the output of the edit summary, not something that varying browser is going to change. Kylu (talk) 05:37, 4 September 2008 (UTC)


 * I searched through the bugzilla database and was unable to find any bug logged on this yet. I have logged it as bug 15481. — OranL (talk) 20:19, 4 September 2008 (UTC)

Software project
I was wondering if anyone would be interested in starting a software project. Basically what it would do is query CatScan and output the results in table format. The results would have multiple columns depending on the categories an article belongs to. It would be used for members of Wikipedia Projects to organize work. SharkD (talk) 01:28, 3 September 2008 (UTC)
 * I just created an account yesterday with a free web hosting service. The service supports ASP and has an ASP.NET beta program. Can I use the DotNetWikiBot framework on this system? I'm a complete newbie when it comes to server-side programming. Thanks. SharkD (talk) 15:52, 4 September 2008 (UTC)

Special:Watchlist not working
At Special:Watchlist I keep getting a blank white page with html source: <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=Content-Type content="text/html; charset=windows-1252"></HEAD> <BODY></BODY></HTML> I have MonoBook and IE7, and have not made any browser or account changes recently. It worked less than an hour ago. PrimeHunter (talk) 21:44, 3 September 2008 (UTC)


 * Hmm, works for me. Indeed, I noticed your post on my watchlist.  —Ilmari Karonen (talk) 21:50, 3 September 2008 (UTC)


 * Also, there's no way Wikipedia should be sending anything in windows-1252; we do everything in utf-8 here. This feels very strange.  —Ilmari Karonen (talk) 21:53, 3 September 2008 (UTC)


 * Now it varies between the blank page and no page at all. I don't know how it works but I wonder whether my IE7 browser on Windows Vista could have made the quoted html when it couldn't get a correct page. I have a few thousand pages on my watchlist, I don't remember the count. Are others with large watchlists having problems? PrimeHunter (talk) 22:11, 3 September 2008 (UTC)


 * I just discovered it works when I change my watchlist preferences to only show 1 day. There are 2054 pages on my watchlist. It still fails when I choose 2 or my normal 3 days in preferences, and also when I choose 1, see a working watchlist, and then click the '3' at days to change the display without changing preferences. PrimeHunter (talk) 22:20, 3 September 2008 (UTC)


 * If I choose 1 day in preferences, view the watchlist, click "Hide logged-in user edits", then I can display both 3 and 7 days, but if I have 3 displayed and click "Show logged-in user edits" then it fails again. It definitely looks like some kind of size limit. PrimeHunter (talk) 22:24, 3 September 2008 (UTC)


 * I have about 2600 pages on my watchlist, and I can view the whole list for the last 30 days (which is how far back the recentchanges table goes). Takes bloody ages to load, but it works.  I'm using Firefox 3 on Kubuntu Linux.  —Ilmari Karonen (talk) 22:35, 3 September 2008 (UTC)
 * Don't forget that watchlist can be in quite different "modes": simple or expanded (all changes), simple or enhanced (changes grouped together). I mean, everybode has to mention what kind of watchlist the're using, and only then you can compare. —AlexSm 22:46, 3 September 2008 (UTC)


 * I can view 5000 edits at Special:Recentchanges without problems. I have no checkmark at "Expand watchlist to show all applicable changes", or at any of the Hide's, or at "Enhanced recent changes" under Recent changes. But it now appears it's not a size limit but a specific (unknown) entry. http://en.wikipedia.org/w/index.php?title=Special:Watchlist&days=1.274 currently works but http://en.wikipedia.org/w/index.php?title=Special:Watchlist&days=1.275 doesn't. The exact value of the working day count is slowly increasing as time apparently moves away from the mysterious entry. I can go back 7 days by clicking "Hide logged-in user edits", but it fails at 1.275 days if logged in edits are shown and all other are hidden. PrimeHunter (talk) 23:12, 3 September 2008 (UTC)


 * I have created an alternative account (had long considered that for other reasons). I copied PrimeHunter's watchlist to the new account and got the same problem which disappeared when most of the watchlist was deleted. I will now attempt a sort of binary search removing around half the watchlist at a time to look for a bad entry. PrimeHunter2 (talk) 23:48, 3 September 2008 (UTC)
 * Got it. I now have a watchlist only containing Homat el Diyar. The watchlist currently fails at http://en.wikipedia.org/w/index.php?title=Special:Watchlist&days=1.313 which is around this edit. Don't ask me why. Can anybody confirm the problem? Add a little to days when testing. PrimeHunter2 (talk) 00:06, 4 September 2008 (UTC)
 * Yes, that breaks my watchlist on FF3, Opera and Safari. I get a blank page with blank source. On IE I get a blank page with the same source as you got. Algebraist 00:09, 4 September 2008 (UTC)
 * Thanks for the confirmation. Finding the cause had me scratching my head for a while. PrimeHunter2 (talk) 00:19, 4 September 2008 (UTC)

(undent) Okay, testing shows I get a blank page and a 500 Internal Server Error. Specifically, the response headers are: 500 Internal Server Error Cache-Control: private, s-maxage=0, max-age=0, must-revalidate Connection: close Date: Thu, 04 Sep 2008 00:32:17 GMT Via: 1.0 sq34.wikimedia.org:3128 (squid/2.6.STABLE21), 1.0 knsq28.knams.wikimedia.org:3128 (squid/2.6.STABLE21), 1.0 knsq5.knams.wikimedia.org:80 (squid/2.6.STABLE21) Server: Apache Vary: Accept-Encoding Content-Length: 0 Content-Type: text/html Client-Date: Thu, 04 Sep 2008 00:32:19 GMT Client-Peer: 91.198.174.2:80 Client-Response-Num: 1 Set-Cookie: enwiki_session=[CENSORED]; path=/; HttpOnly X-Cache: MISS from sq34.wikimedia.org X-Cache: MISS from knsq28.knams.wikimedia.org X-Cache: MISS from knsq5.knams.wikimedia.org X-Cache-Lookup: MISS from sq34.wikimedia.org:3128 X-Cache-Lookup: MISS from knsq28.knams.wikimedia.org:3128 X-Cache-Lookup: MISS from knsq5.knams.wikimedia.org:80 X-Powered-By: PHP/5.2.5 X-Vary-Options: Accept-Encoding;list-contains=gzip I think it's time to call in the devs and see what's causing it. —Ilmari Karonen (talk) 00:36, 4 September 2008 (UTC)


 * Fixed now. The invalid title which caused this will show up as unlinked instead. -- Tim Starling (talk) 04:10, 4 September 2008 (UTC)


 * Thanks. My earlier diff was miscalculated because I forgot to set local time on the alternative account. It's this move which was reverted (to great anger of the mover by the way) before I detected the error. PrimeHunter (talk) 12:02, 4 September 2008 (UTC)

Print and handheld style overrides
I've added support for MediaWiki:Handheld.css and MediaWiki:Print.css site-specific style additions for the handheld and printable styles. I threw in some quick sample styles to flatten the layout tables on the Main Page in handheld style; see my blog post on the matter for sample images in Opera's small screen mode. --brion (talk) 03:57, 4 September 2008 (UTC)


 * Sweet! --Splarka (rant) 07:37, 4 September 2008 (UTC)

javascript tools
What software do javascript authors use to develop code that interacts with wikipedia--without writing it online and needlessly creating hundreds of code-page revisions? Thanks, Whiskeydog (talk) 05:12, 4 September 2008 (UTC)


 * Greasemonkey is pretty popular. You can also do a lot with just [preview] on an edit to a .js page. I personally like to save a page locally, add  to the head, insert my own , and then load it via   URI (you can't use AJAX without some serious security relaxing though). You can also (depending on security settings) use in some browsers a trick like   in the location bar. --Splarka (rant) 07:37, 4 September 2008 (UTC)


 * My favorite way is a local web server and a monobook.js that calls localhost (although Opera 9.5+ disallowed this). —AlexSm 14:37, 4 September 2008 (UTC)


 * Hundreds or thousands of javascript edits will upset no one if hosted off-site. You can store on and import from any wiki farm or even pastebin.ca or your blog or anywhere else where you have write-access. — CharlotteWebb 15:41, 4 September 2008 (UTC)


 * Thanks for all the responses. I hadn't realized that a script could be imported via monobook.js from another domain! Whiskeydog (talk) 02:46, 5 September 2008 (UTC)

Unsolved questions
In the article, the author(s) could formulate a link with unsolved questions. There he could put things he doesn’t know about the topic. Someone, who is reading wikipedia, could find the link and answer the questions. This would help to improve the articles and would be an excellent new tool for expanding wikipedia.

Hope you enjoy the idea.

Alexandre Lipski Gonçalves —Preceding unsigned comment added by 201.83.217.113 (talk) 15:40, 4 September 2008 (UTC)


 * Having a place to ask questions is a good idea but there are already places, the talk page of the article and Reference desk. I don't think more is needed. PrimeHunter (talk) 16:35, 4 September 2008 (UTC)


 * Actually, I believe there are templates that ask for particular help. They can be placed in individual article sections. I don't know offhand where those templates are. Anybody else know? --Timeshifter (talk) 23:28, 4 September 2008 (UTC)


 * The closest I know is Expert which links to the talk page where specific questions might be asked. And then there is Unsolved but that doesn't ask for help and should only be used to indicate that nobody knows the answer. PrimeHunter (talk) 00:08, 5 September 2008 (UTC)


 * Expert is one I remember. Maybe some other templates need to be created. --Timeshifter (talk) 02:29, 5 September 2008 (UTC)

Trouble with parser functions
I am attempting to make a template (User:Someguy1221/AFC) that takes a single parameter to return a string of text, but no matter how much I fiddle with it, it only returns the default value. Anyone know what I'm doing wrong? Someguy1221 (talk) 23:10, 4 September 2008 (UTC)
 * Should work now. --- RockMFR 23:21, 4 September 2008 (UTC)
 * Thanks! Someguy1221 (talk) 23:27, 4 September 2008 (UTC)

MediaWiki:Sp-contributions-footer
Would there be any objections if I were to create this particular MediaWiki: page? Examples of its use can be seen on Meta and Wikisource (see bottom of here and here) &mdash; Anonymous Dissident  Talk 08:56, 10 August 2008 (UTC)
 * I want it I want it I want it!! <b style="color:forestgreen;">Happy</b>‑<b style="color:darkorange;">melon</b> 20:43, 10 August 2008 (UTC)
 * YES! Interestingly enough I was just thinking about this too after my recent editing on WB. - Icewedge (talk) 07:25, 11 August 2008 (UTC)


 * Nicely done. :) Is there a particular reason MediaWiki:Sp-contributions-footer and MediaWiki:Sp-contributions-footer-anon use different images and css attributes? In particular, Image:Wikipedia-logo.svg (38px) and Image:Information.svg (40px)... and the tables themselves seem to render with slightly different widths, on my browser. I favor consistency, but have no particular preference regarding which style we keep. – Luna Santin  (talk) 10:07, 11 August 2008 (UTC)


 * Luna Santin: You are right. I'll update both to use our standard MediaWiki message box styles, thus they will look like the MediaWiki:Sharedupload. With 100% width, standard light grey background and so on.
 * I have coded up examples in my user space at User:Davidgothberg/Test30. I don't know which image to choose. I am leaning towards the Wikipedia logo (it looks good), or the icon tools image (since it says more about what content the boxes have).
 * Note! In my examples I also suggest a slight change in the MediaWiki:Sharedupload box. (The "This is a file from the Wikimedia Commons" box.)
 * --David Göthberg (talk) 22:49, 11 August 2008 (UTC)


 * Interesting. I definitely appreciate the consistency. As far as images, I believe I would personally favor Image:Information.svg or Image:Icon tools.png (though I believe you mentioned you liked Image:Imbox notice.png previously, which is also fine with me) -- they hint nicely as to the box's contents, as you said. – Luna Santin  (talk) 09:09, 12 August 2008 (UTC)


 * Based on the two of us it seems the Image:Icon tools.png is the choice. But I think we need some more users to look at this. So please, can more people take a look at User:Davidgothberg/Test30 and comment here?
 * --David Göthberg (talk) 06:25, 13 August 2008 (UTC)


 * I also prefer this icon; it matches the purpose of the links much better. I'd certainly prefer not to see the Wikipedia logo. Waltham, The Duke of 12:20, 15 August 2008 (UTC)

Is there any way I can add the template to the bottom of my user page? Thanks! SharkD (talk) 22:12, 18 August 2008 (UTC)


 * SharkD: I see that you answered yourself by making the template sp-contributions-footer.
 * Everyone: I think I have found a much better image to use for these boxes: Image:User-info.svg [[Image:User-info.svg|40px]]
 * Perhaps we should use a more colourful version of that image, but you get the idea. But these boxes are interface messages so perhaps it is good that the image looks a bit neutral. See my example boxes at User:Davidgothberg/Test30 to see how it looks in the boxes.
 * --David Göthberg (talk) 08:21, 19 August 2008 (UTC)


 * I really think that "IP user" should be substituted for the three places that "anonymous user" appears, as everyone should know that most user names are far more anonymous than IP addresses. 199.125.109.90 (talk) 17:14, 21 August 2008 (UTC)


 * Yes, I agree with the IP user 199.125.109.90, it is time we get rid of the misnomer "anonymous user" for the not logged in users.
 * --David Göthberg (talk) 17:48, 21 August 2008 (UTC)


 * The term "IP user" is terrible. It's incomprehensible jargon to practically anyone on the planet.  Call them "unregistered users". —Simetrical (talk • contribs) 15:19, 22 August 2008 (UTC)


 * Not specific enough. IP user is fine, and it is well explained. It's not something that would ever be used out of context. I know I would much rather be referred to as an IP user than as an un-something. 199.125.109.64 (talk) 15:40, 22 August 2008 (UTC)
 * IP user is not "well explained". It's incomprehensible unless you happen to be one of the tiny technical elite who knows what an IP address is, and even then it's incomprehensible unless you realize that MediaWiki names unregistered users by IP address (which is unheard of for web software). —Simetrical (talk • contribs) 02:15, 24 August 2008 (UTC)
 * I like the new image; I don't mind the discreet blue colour (actually, I think I like it), and I agree that the image should look neutral. In the sandbox I notice two additional new suggestions... As I've said, I prefer this icon to one with more colour (in this case, the one wearing a red tie). As far as the black one is concerned, I get the meaning ("shady" members), but I don't think it will be appreciated much, and it stands out too much anyway.
 * Although I might consider it if it were stroking a white Persian cat. Waltham, The Duke of 22:16, 22 August 2008 (UTC)


 * As most of you probably have already noticed: I have changed to the "user info" image I show above for all the related MediaWiki messages.
 * Waltham: Yeah, the black one looks scary. I bet some would like it for IP users! :)) The two coloured images were the only related ready made ones I could find on Commons. I have tried with softer green versions on my own computer, but didn't upload them since they didn't look as good as the grey-blue one we are using now.
 * Simetrical and 199.125.109.64: Both "IP user" and "unregistered user" sounds right to me. Both are way better than "anonymous user". I think we need input from more users to decide which one is best. So people, which do you prefer: "IP user" or "unregistered user" or something else?
 * --David Göthberg (talk) 11:15, 23 August 2008 (UTC)
 * I prefer IP user as it is both shorted and as 199.125.109.64 (I think) pointed out, unregistered could be seen as negative. —Atyndall &#91;citation needed&#93; 11:17, 24 August 2008 (UTC)
 * I am being pedantic about semantics, but maybe use "editor" rather than "user". All editors are users, but not all users are editors (some are just readers), and there is no purpose in having "IP user contributions" for IP readers-only. — Twas Now ( talk • contribs • e-mail ) 11:39, 24 August 2008 (UTC)


 * Twas Now has a point there. Here are the first sentence of the message with the different namings. Version 0 being the old message:
 * 0: This is the contributions page for an anonymous user, identified by the user's numerical IP address.
 * 1: This is the contributions page for an IP user, identified by the user's numerical IP address.
 * 2: This is the contributions page for an unregistered user, identified by the user's numerical IP address.
 * 3: This is the contributions page for an IP editor, identified by the editor's numerical IP address.
 * 4: This is the contributions page for an unregistered editor, identified by the editor's numerical IP address.
 * I think I prefer versions 1 and 2, but I am okay with 3 and 4 too. To me "user" sounds more friendly and welcoming than "editor". By the way, since these footer boxes are very visible I think that the naming we use in them is likely to become the standard naming we use here at Wikipedia.
 * --David Göthberg (talk) 15:48, 24 August 2008 (UTC)


 * I have no preference, but I'd like to make two observations:
 * IP user/editor is more educational, as it makes the connection with IP address and shows the Wikipedian in question what an IP user/editor is without having to be a nerd or read any help pages.
 * As far as Twas Now's point is concerned, he is right in that only those IP addresses with edits are relevant here, so it makes no sense to refer to plain readers in the toolbox. Furthermore, I think we can say that editor is a title automatically bestowed on anyone who's made an edit, not least because the two words are obviously related.
 * On the other hand, since only those who have made edits will have a contributions page (where the toolbox at the bottom appears), the IP user in question cannot be a plain reader anyway. The "this is the contributions page of an IP/anonymous user" message would only appear for editors, who do belong to the category of IP users.
 * Full circle. :-) Waltham, The Duke of 13:04, 25 August 2008 (UTC)


 * So glad Version 0 is getting looked at. I much prefer "IP editor" to any other option."Hey look, you're already an editor. One of us. One of us". Seriously, it is much more personal, friendly and instructive than "IP user" (my second choice). Hopefully affording us the title will rub off on both IPs and those registered. Dream the dream. 86.44.24.10 (talk) 02:32, 29 August 2008 (UTC) is pimping his sidebar proposal


 * User is standard terminology on Wikipedia. Note: User:Dream, User:Duke, User:David. If Wikipedia had been using Editor:The Duke of Waltham then IP editor would be appropriate. Since WP uses User:The Duke of Waltham, then it should be IP user for consistency. 199.125.109.96 (talk) 06:31, 2 September 2008 (UTC)


 * Good point. Considering this, and Mr Göthberg's view, I think I'll also choose user over editor; I still agree with Twas Now's view, but I have already mentioned that I don't find it very relevant.
 * Additional benefit of IP user: it is the shortest available option (not a factor I considered in my choice, though).
 * Having said that, I believe we have consensus on the IP part. Waltham, The Duke of 15:27, 4 September 2008 (UTC)


 * I agree, 199.125.109.96 has a very good point when he points to that user pages are named "User:". Previously I only had a slight preference for "user" over "editor", but now my preference is much stronger. And yeah, it seems most of us prefer "IP" over "unregistered" (darn that is hard to spell!).
 * Waltham: Right, "IP user" is the shortest term we have that still is clear. And I believe that is a bigger benefit than you might think. Since when Wikipedians discuss things on talk pages they tend to use as short forms as possible. (They are way too lazy for their own good!) So it is good if we standardise around a short but still clear form, thus the same form can be used all over the place (including in documentation) and thus causing less confusion. So on talk pages using "IP user" is much better than the current all too common short form "anons".
 * --David Göthberg (talk) 12:14, 5 September 2008 (UTC)

Proposal: add "my sandbox" link to top navbar
I propose that a "my sandbox" link be added to the top navbar, in between "my talk" and "my preferences". This might encourage users to work on articles in their sandbox before moving them to article space. SharkD (talk) 04:52, 28 August 2008 (UTC)


 * Perhaps a Gadget? --MZMcBride (talk) 06:11, 28 August 2008 (UTC)


 * Yes, that would work too. Still, I'd like it if it were considered for all registered users. SharkD (talk) 06:14, 28 August 2008 (UTC)


 * This would be a very good feature for all users. I like the placement of the link that SharkD suggests, between "my talk" and "my preferences". And I like the suggested name "my sandbox".
 * It should not be a gadget since this is exactly for the kind of users who still probably have no idea what a gadget is. Instead we can make a gadget to turn off the link for those who don't want/need it anymore.
 * We should probably make it so the sandbox link isn't red. (For instance by using my sandbox instead of my sandbox.)
 * What page name should the sandbox have?
 * /Sandbox
 * /sandbox
 * /Test1
 * /test1
 * That is, upper or lower case first character? And should we name it "/Test1" to make it clear that one can make more test pages? (When I tell people to "Please try it in your own user space first" then I link them to Special:Mypage/Test1.) On the other hand, that "/Test1" link kind of gets wrong once one has more test pages, so simply "/sandbox" is perhaps more timeless? For "/Test1" I have a slight preference for upper case. While for "/sandbox" I have a slight preference for lower case since that is our standard naming for template sandboxes.
 * This feature could also be good for IP users. But we perhaps need to think a bit about where the IP user's sandbox should be placed.
 * --David Göthberg (talk) 11:59, 28 August 2008 (UTC)


 * Perhaps IP users should just be pointed to THE sandbox with logged-in users pointed to their own?
 * —Preceding unsigned comment added by Clubjuggle (talk • contribs) 12:23, 28 August 2008


 * I agree with Sharks' original proposal and second Clubjuggle's IP suggestion - X201 (talk) 12:43, 28 August 2008 (UTC)


 * I agree with directing IPs to THE sandbox. I disagree with David Göthberg in that I don't think that the links need to be forced into blue links. Both the user and user talk pages start out as red links for newly-joined users until they actually get around to creating these pages. The same can be true for user sandboxes. SharkD (talk) 13:56, 28 August 2008 (UTC)
 * I think /Sandbox (upper case) would be fine. A custom message could appear there when starting the page describing how to create multiple (and sub-hierarchical) sandboxes, kind of how a custom message appears before creating any other aritcle. SharkD (talk) 14:02, 28 August 2008 (UTC)
 * There are some of us who have multiple sandboxes or other subpages. I have a couple of pages of notes for articles I want to create or develop, and several I use to test templates or tables. --——  Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  12:55, 28 August 2008 (UTC)


 * Gadget850: Well, there's nothing preventing you from having more sandboxes, and even from putting a link list on your "my sandbox" page to your other sandboxes and test pages. :)
 * Everyone: Linking IP users to Sandbox might be a slight problem. Since already now that sandbox gets so frequently edited that I think those poor newbies must get edit conflicts all the time. But since we probably will be adding the "sandbox" link using javascript, then we can make it so it points to many different sandboxes. For instance we could have say 32 different sandboxes. (And we can increase that later if needed.) But we shouldn't choose the sandbox randomly, since the same user will expect to see the same sandbox the next time he/she visits it. So we could use the first byte in the user's IP address to choose which sandbox to use. That means the same user will usually see the same sandbox even when he visits some time later, since usually only the lower bytes in the IP address changes between sessions. I think that should be a no-brainer for our javascript coders to fix.
 * --David Göthberg (talk) 13:11, 28 August 2008 (UTC)
 * Take a look at my sandbox page: User:SharkD/Sandbox. It serves more as a directory for all my other sandbox pages, which are stored as sub-pages under the main sandbox page. SharkD (talk) 13:59, 28 August 2008 (UTC)

Where are we going to put the sandbox link for IPs? Algebraist 13:13, 28 August 2008 (UTC)
 * The link could go in the same place, but direct to the wiki sandbox. SharkD (talk) 13:59, 28 August 2008 (UTC)

Simple JS implementation: <pre style="overflow: scroll;"> addOnloadHook(function {    addPortletLink('p-personal', wgArticlePath.replace("$1", "Special:Mypage/Sandbox"), 'My sandbox', 'pt-sandbox', 'Your personal sandbox', null, document.getElementById('pt-preferences')); }); Of course, if we added it to the sidebar instead, we could do it without JavaScript. —Ilmari Karonen (talk) 14:15, 28 August 2008 (UTC)


 * I think it would be great to have a "my sandbox" or (for (IP editors) a "sandbox" link on the navbar. I don't think it's so critical to explain to editors how to create multiple subpages; I'd much prefer the message box that shows up (when someone clicks on the red link) to simply explain that this is a personal subpage for testing of editing, with a link to WP:SUB for those who might be interested.  But that's just details to be worked out later; the core concept seems really good.


 * As for IP editors, having their "sandbox" link go to one of 100+ sandboxes, depending on the first block in the IP address, has pluses and minuses. On the plus side, it minimizes edit conflicts that occur when lots of people are editing the standard sandbox simultaneously; on the minus side, it means that the bot or bots doing the cleanup need to cover 100+ pages.  But again, I think this is a secondary issue; it would be great to be able to tell IP editors to just click the "sandbox" link at the top of their screen rather than provide them with a link to WP:SAND, and I think there is a reasonable chance that we'd have fewer "test" edits on real pages if there was a "sandbox" link visible at the top of screens.  And maybe we'd even get a few more people to realize that they can edit Wikipedia.  (On a lot of other wikis, if you click the "edit" tab, you just get a message saying that you have to log in first to edit.)  -- John Broughton  (♫♫) 21:06, 28 August 2008 (UTC)


 * John Broughton: Right, the IP user sandboxes needs to be cleaned regularly by a bot, just like the current Sandbox. And I can imagine there are cases when an admin has to delete the sandbox to get rid of especially nasty stuff from the edit history of the sandbox. And that was why I suggested using a fairly low number of such sandboxes. And by the way, I suggest naming them Sandbox1 and so on.
 * Ilmari Karonen: Right, we could put the link in the sidebar. Although I think this is an important enough feature that it should probably be put on the navbar at the top of the page. (SharkD's original suggestion above.)
 * Ilmari and our other javascript experts: Does javascript have access to the IP user's name or IP number so we can use it to pick which sandbox to send the IP user?
 * --David Göthberg (talk) 18:38, 29 August 2008 (UTC)


 * Special:Mypage/sandbox seems to work just fine for anon users (lower case is better IMHO, in the same form as skinname.css/js) --Splarka (rant) 08:04, 30 August 2008 (UTC)


 * @David Göthberg: JavaScript has no way of finding a user's IP address alone - one has to use server-side code (i.e. PHP) or (ugh) a Java applet. x42bn6 Talk Mess  04:09, 31 August 2008 (UTC)


 * Splarka: Yes, Special:Mypage/sandbox works for IP users too. But that would mean after a year or so we would have perhaps a million IP user sandboxes, some of which might hold pretty nasty content. And some search engines do index all pages on Wikipedia...
 * x42bn6: Ouch, seems you are right. I checked around, the "wgUserName" javascript variable is only filled in for logged in users. And since those variables are set in the rendered page header it would probably destroy the squid caching if the user name (IP number) of IP users were set in the page header. So for now it seems our best option is to direct the "sandbox" link for IP users to Sandbox.
 * For the logged in users I prefer we use lower case "/sandbox" for the "my sandbox" link. Since as Splarka mentioned above, we already use lower case for the user's own monobook.js and monobook.css. And we use lower case for the now standardised subpages for templates such as "/doc", "/sandbox" and "/testcases".
 * Then after some weeks if people like these things then I suggest we request the making of a special page named Special:Sandbox that links to Special:Mypage/sandbox for logged in users and to a numbered public sandbox for IP users. But let's leave that aside for now.
 * Since all comments above have been so positive I think we can declare a consensus to do this. The only thing that remains is to decide on upper case or lower case "/Sandbox" or "/sandbox". (Sorry to be picky about this, but we can not change that spelling later since then a lot of users will have a problem to find their old /sandbox.)
 * --David Göthberg (talk) 15:28, 31 August 2008 (UTC)


 * I am not even a javascript programmer, but I have figured out how we can do the numbered sandboxes for the IP users: When a new user visits we can randomly assign one of the numbered sandboxes, and then store the result in a javascript cookie. Then if the user comes back an hour later we can assign the same sandbox to that user. Of course, if the user has set his browser to clear cookies between sessions then he gets assigned a new sandbox. But at least the user will see the same sandbox in all open windows during the same browser session, which is more important than seeing the same sandbox some hours later. It seems simple enough that I could cut and paste some cookie examples I found on the web to do it. But to not break anything I should probably leave that to our javascript experts.
 * But I'll deploy a simpler code at MediaWiki:Common.js first. See the code I am showing at MediaWiki talk:Common.js.
 * --David Göthberg (talk) 20:25, 1 September 2008 (UTC)


 * Sounds good. SharkD (talk) 06:02, 2 September 2008 (UTC)


 * This suggestion has met resistance over at MediaWiki talk:Common.js.
 * --David Göthberg (talk) 13:07, 5 September 2008 (UTC)

Heading structure broken
Just a heads-up, as I've also been advised to take the following to MediaWiki. Per Help desk:

The heading structure on every page appears to be broken. There is an H1 (the page title) immediately followed by an H3 ("From Wikipedia, the free encyclopedia" - which isn't even a heading, and should probably be marked up as a paragraph); and the page ends with a bunch of H5s ("Views", "Personal tools", "Navigation", "Search", "Interaction", "Toolbox") with no parent other than H1 (and the last headings in the article, which are not meant to apply to it).

Headings should follow a strict hierarchy, H1 > H2 > H3, and so on. WCAG says:

"3.5 Use header elements to convey document structure and use them according to specification. [Priority 2] For example, in HTML, use H2 to indicate a subsection of H1. Do not use headers for font effects."

and:

"Since some users skim through a document by navigating its headings, it is important to use them appropriately to convey document structure. Users should order heading elements properly. For example, in HTML, H2 elements should follow H1 elements, H3 elements should follow H2 elements, etc. Content developers should not 'skip' levels (e.g., H1 directly to H3). Do not use headings to create font effects; use style sheets to change font styles for example."

See also Wikipedia talk:Accessibility. Andy Mabbett | Talk to Andy Mabbett 17:57, 1 September 2008 (UTC)


 * —Simetrical (talk • contribs) 21:45, 1 September 2008 (UTC)


 * Thank you - you beat me to it, because I am still waiting for my bugzilla registration confirmation e-mail to arrive. It's very disappointing to see that the bug was raised on 12 September 2004, just short of 4 years ago, and yet a fix has not yet been implemented. Andy Mabbett | Talk to Andy Mabbett 21:59, 1 September 2008 (UTC)


 * You realize that there are 3087 unresolved bugs for MediaWiki and Wikimedia? What exactly makes you think that every problem is likely to get fixed within four years of being reported?  This particular one doesn't cause any real problem that I've heard of, but fixing it would break all sorts of user scripts and custom styles. —Simetrical (talk • contribs) 14:14, 2 September 2008 (UTC)


 * It causes the real problem that you will have heard of, when you read the cited WCAG guidance, above. Where did I say that I thought that every problem would be resolved within four years? If your latter assertion is true, then the lesson is to be stricter in adhering to such guidelines in the first place. Andy Mabbett (aka Pigsonthewing); Talk to Andy Mabbett; Andy Mabbett's contributions 00:10, 3 September 2008 (UTC)
 * Violating WCAG guidance is not a "real problem" unless people actually complain that in some way their browsing experience has been harmed by it. There have been no complaints I've heard except by standards purists, no complaints by anyone who actually felt that the dubious behavior was harmful to them in some way.  I completely agree that it's wrong and should be changed at some point, but it shouldn't be surprising that it receives very low priority, since it would be a pain to fix (breaking some unknown amount of CSS/JS) and there are no concrete benefits.  Nothing is "broken" here, we just don't completely adhere to relevant standards. —Simetrical (talk • contribs) 14:51, 3 September 2008 (UTC)
 * Everyone whose browsing experience is mucked up because they use screen readers (which treat heading structures as they should be treated) will disagree with you that nothing is broken and that fixing the problem will have no concrete benefits. It is difficult to imagine any CSS that would be broken by fixing this, and I'm skeptical that any system-wide JS would be effected either.  That user-created tools will need to be updated should not be a major factor (if one at all) in WikiMedia following basic web standards.  If anything, by following them more closely it will make it easier for third-party developers and everyday JS-using editors to write tools for WikiMedia sites, since things will work as expected instead of weirdly and requiring exceptions.  Argument to expediency is rarely a strong argument, and almost never a good one in a standards and development discussion, since every corner cut is very likely to lead to problems down the road that simply grow over time and become harder to resolve the longer they sit unresolved (i.e. the user scripts fixes you are certain will be required). —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 01:41, 4 September 2008 (UTC)
 * PS: All that needs to happen to resolve this is changing the abused  to some other block element like a   or , depending on the exact code context (I'd have to look at the source to be sure). It could even keep the same id, class and/or other CSS information associated with it, with only minimal if any changes to the .css files.  The fix strikes me as almost ridiculously trivial to implement, and also trivial for script writers to update for. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 01:44, 4 September 2008 (UTC)
 * You would probably receive more sympathy if you were less condescending and self-righteous. We've received no reports from users of screen readers, or anyone else, saying that they're experiencing an actual problem from this.  If you're going to claim it is one, you need to give a concrete case of someone whose browsing experience was affected, not make up further chains of hypotheticals that really serve to mask the fact that your objection is based on the fact that we're violating standards per se.  I have no objection to fixing things merely for being standards violations, but they have deservedly lower precedence than fixing things that actually cause problems. You seem to have missed the fact that Monobook also misuses h5 in the portlets, by the way.  And maybe other places I don't know of offhand.  I have broken user scripts before by committing "ridiculously trivial to implement" document structure changes, so I don't feel quite as blasé about changing things around as you do.  (This, the addition of a pBody wrapper, broke a script that many people were using.  Best of all was that they copy-pasted it from each other instead of including from a central source, so each one individually had to be instructed on how to fix it.) —Simetrical (talk • contribs) 15:59, 5 September 2008 (UTC)

Some Usercheck fields have stopped working
I made a template fairly recently, Usercheck-full, which combines the fields for users from all other Usercheck templates. However, some of its links have stopped working of late. Namely, the total link has stopped working, instead linking to a page which informs you that "query.php is dead" (this also means that User11, where I got the code for the link from, has also stopped functioning correctly), and the summary link now just links to a page that says "Object not found!" (this also affects User15). Please advise. It Is Me Here (talk) 09:32, 2 September 2008 (UTC)
 * The templates need to be updated to use api.php. Query.php was shut down a few days ago. Franamax (talk) 00:24, 3 September 2008 (UTC)
 * I made a copy of User11 and changed it. Does this work? Franamax (talk) 00:41, 3 September 2008 (UTC)
 * Yeah, that seems fine, thanks. Could you update usercheck-full too, please? It Is Me Here (talk) 06:47, 3 September 2008 (UTC)
 * Done! Tested for normal case and error case (user unexist). Not tested for valid user with zero contribs. Implemented for User11 (please test!). Implemented for usercheck-full (please test!). That is one heck of a user check. You missed the links asking me to take my shoes off and to discard any liquids I'm carrying. And how exactly did boarding an airplane come to resemble entering a rock concert anyway? When did my 77-year-old Danish aunt get suspected of terrorism? Anyway, I don't see a problem at User15, can you elaborate? Cheers! Franamax (talk) 09:20, 3 September 2008 (UTC)
 * Well, thanks, I think (sorry if my template didn't go down too well!). It's the summary link that doesn't work with User15 and Usercheck-full for me: http://www.math.ucla.edu/~aoleg/wp/rfa/edit_summary.cgi?user=It+Is+Me+Here tells me:
 * " Object not found!
 * The requested URL was not found on this server. The link on the referring page seems to be wrong or outdated. Please inform the author of that page about the error.
 * If you think this is a server error, please contact the webmaster.
 * Error 404
 * www.math.ucla.edu
 * Wed Sep 3 08:57:23 2008
 * Apache/2.2.4 (Linux/SUSE) "
 * The total link is now fine once again, both in User11 and Usercheck-full - thanks for doing that.
 * Now that we're on the subject, I've just realised that the UHx, UtHx and UtE links don't work in Usercheck-full (for me, at least; it interprets the spaces in my user name as + signs). If you could fix that too, then that would also be very much appreciated. This issue also affects User1 plus and User2 plus (the latter seems to have been completely broken by something).
 * It Is Me Here (talk) 16:12, 3 September 2008 (UTC)
 * My comments on the comprehensiveness of your template were typical of my lame attempts at humour :) I actually like it and have added it to the list at User information templates to make it more visible.
 * Anyway, the "summary" field in User15 and usercheck-full now uses mathbot's tool. I also fixed (well, "changed") the Uxx fields in usercheck-full. Also I changed UtE to create a new talk page section rather than edit the whole page - is that a better way for it to operate?
 * I'll wait for comments before making any changes to the Uxx-plus templates. I've got curly-braces on the brain right now after those last edits. :) Franamax (talk) 04:01, 4 September 2008 (UTC)
 * Nice one! The only thing is, UtE has now become redundant as it does exactly the same thing as message; you might want either to revert its function back, or just delete it altogether (see the history for more duplicate categories that I had deleted). Also, the Online / Offline thing does not seem to work for me (see my user page) - do you know where the problem might lie with that? It Is Me Here (talk) 07:01, 4 September 2008 (UTC)
 * Yeah, I changed UtE back. As far as "Online" (I was holding back on this issue), I believe that StatusBot made too many edits and was shot down like a mangy stray dog that keeps chasing your sheep and ripping open your garbage bags. See SB's page - there are some alternatives there you may wish to consider. Otherwise that bit should be removed.
 * Another consideration would be to have hovering tooltips, see "page moves" for an example. This is done with the tag and lets you describe the link function rather than just show the link URL. This may violate our accessibility policy, I'm not sure on that. I'm minded now to go through all the Userxx templates and add tooltips and the User information templates extended documentation throughout. If anyone else is reading this thread, please chime in with your comments! Franamax (talk) 07:37, 4 September 2008 (UTC)
 * OK, I got rid of the Online / Offline thing from Usercheck-full, but couldn't work out which other templates it was on - please remove it from any other usercheck templates it features in if you can figure out where that is. It Is Me Here (talk) 21:24, 5 September 2008 (UTC)
 * Another consideration would be to have hovering tooltips, see "page moves" for an example. This is done with the tag and lets you describe the link function rather than just show the link URL. This may violate our accessibility policy, I'm not sure on that. I'm minded now to go through all the Userxx templates and add tooltips and the User information templates extended documentation throughout. If anyone else is reading this thread, please chime in with your comments! Franamax (talk) 07:37, 4 September 2008 (UTC)
 * OK, I got rid of the Online / Offline thing from Usercheck-full, but couldn't work out which other templates it was on - please remove it from any other usercheck templates it features in if you can figure out where that is. It Is Me Here (talk) 21:24, 5 September 2008 (UTC)

Watchlisting rollbacks
On seeing this interesting question, I tried my (out-of-date) local wiki and it appears that when you use rollback, the article is not added to your watchlist. Can anyone shed light on whether there is a preference to enable this? Franamax (talk) 04:45, 3 September 2008 (UTC)
 * This is true even in the most current version. — CharlotteWebb 15:25, 3 September 2008 (UTC)
 * One of the perks of rollback is that it doesn't add the article to your watchlist. It makes dealing with widespread vandalism (which is its entire purpose) that much easier. If you want to add the pages to your watchlist, use tabbed browsing and open each rollback link in a separate tab; then you can go through the rolled back pages and watchlist the ones you want. EVula // talk // &#9775;  // 04:10, 4 September 2008 (UTC)
 * Bleaghh! Now that I know, I'll do that. I guess. I would rather have the preference - if I use rollback, almost by definition I want to watchlist that article to be sure the vandalism ceases. As far as "its entire purpose" - that was not exactly the way non-admin rollback was presented to the community, although there was much talk of its efficiency for automated tools. (Which I just cautioned a good-faith vandal-fighter for overusing) Franamax (talk) 04:57, 4 September 2008 (UTC)
 * Huggle has an option to automatically add reverted pages to your watchlist -- Gurch (talk) 15:56, 5 September 2008 (UTC)

Images
There's some kind of a problem with images at Commons. There's a discussion at WP:ANI. I first noticed it at Napoleon Marache. If you go down to the section titled "Game against Morphy" there are several pictures of chess boards. One of the images that makes up this image, Image:Chess zver 26.png, an image from commons, is coming up as a blue text link instead of an image and screwing up the whole layout of the page. I am using Firefox. The problem seems to be browser-dependent, based on the discussion at ANI. I have refreshed my cache and the problem persists. --Steven J. Anderson (talk) 06:15, 5 September 2008 (UTC)


 * Might be something to do with this. 3000 images were accidentally deleted from the database, 2500 were recoverd but about 500 have been lost. See commons:Commons:Village_pump for more information. Woody (talk) 13:30, 5 September 2008 (UTC)


 * OK, that explains the issue with ambox, where two of the icons (for style and content) are missing. Any objections to replacing them with Image:Imbox content.png for the time being? UltraExactZZ Claims~ Evidence 14:06, 5 September 2008 (UTC)


 * Yeah, when reading the thread about the images over at WP:ANI I immediately checked all the images I am "responsible" for. And as you Ultraexactzz point out the two images Image:Ambox content.png and Image:Ambox style.png were gone. So I immediately re-uploaded them (that is, at the same time you were typing your above message).
 * That those two images were gone shows that the list of damaged images that Woody links to above is not complete. The two ambox images are not listed in that list. So people, better check all images that you are "responsible" for, that is images for which you know where to get a new copy.
 * By the way, the ambox and imbox images are not identical. They look identical in most browsers and the MediaWiki "Search for duplicate files" function thinks they are identical. But they are not. They have special tweaked transparent backgrounds to match the slightly different background colours of the ambox and the imbox. This means they look "transparent" in older web browsers too. (Such hand tweaking is only worth the trouble for such widely used icons.)
 * --David Göthberg (talk) 14:28, 5 September 2008 (UTC)

Wiki slow to load pages
Is it me, or is wikipedia operating very slowly? I especially have problems with any pages with images. But many pages are very slow, some articles taking minutes. Is this common to everyone? &mdash; Mattisse (Talk) 17:37, 5 September 2008 (UTC)
 * Which articles and with which browsers are you testing? I measured the load time for Firefox 2.0 on these largish articles on 5 August, and upon retesting just now (no bottlenecks with network or CPU at my end), the load time is roughly slower. By "roughly" I mean within an order of magnitude (from a half to triple the previously measured time). DNA loaded in 16 seconds then, now it's 24 to 40 seconds. Some monster articles are a lot slower: Russia took 16 seconds now it takes 38 seconds. Alpha Kappa Alpha loads consistently slower, around 14 seconds. €2 commemorative coins takes around 12 to 15 seconds, so not much slower. I just tested a few with Google Chrome and load times were significantly faster (Russia loads fully in 8 seconds and Alpha Kappa Alpha in less than 3 seconds for example). -84user (talk) 22:00, 5 September 2008 (UTC)
 * I actually upgraded to Firefox 3 to see if that made a difference. It didn't. I agree that some long pages are very slow, while others are fine. Articles with images seem the worst. Some articles, the images never did arrive. Good article nominations was taking a long time for a while but seems better now. I am concluding, from what you are saying, that it is not just me. Thanks! &mdash; Mattisse (Talk) 22:07, 5 September 2008 (UTC)

math rendering
Hi! What does this mean? $${\rm hp}$$ " {\rm " appears to be a shortcut for mathrm? If so, are there any more such shortcuts? Thanks Saintrain (talk) 18:42, 3 September 2008 (UTC)


 * \rm stops the text from being rendered in italic like variable names (compare $$xyz$$ and $$\rm xyz$$). You can try Help:Displaying a formula for more info. Admiral Norton (talk) 22:26, 3 September 2008 (UTC)
 * Thanks. That's what it looked like but it's not documented in Help:Math or Help:Displaying a formula, hence the 1st question.
 * Is "\" an abbreviation for "math" in this context? Does it work for \it (== \mathit), etc?  Is it documented someplace?  Is it a deprecated function or should the above docs be updated to reflect it's use?  Saintrain (talk) 18:57, 4 September 2008 (UTC)
 * It is deprecated but still widely used within LaTeX. No idea what that implies for its status in the wiki renderer. And no, the backslash is not an abbreviation, it is the character used in LaTeX to signal the start of a formatting macro. \it and \mathit are two different macros that happen to have similar meanings. —David Eppstein (talk) 05:11, 5 September 2008 (UTC)
 * Thanks David. It's gone (at least where I found it).  Saintrain (talk) 03:51, 7 September 2008 (UTC)

google books link breaks when used within [ ]'s or in templates
The following google books URL breaks when used within [ ]'s or within templates, see this diff within a template. http://books.google.com/books?id=y1_R-rjdcb0C&pg=PA726&vq=The+third+son,+James&dq="Alexander+Hamilton"&source=gbs_search_s&sig=ACfU3U36bnOJW4HE1bpJ004GuxaiXnX7Vg It breaks at the "Alexander+Hamilton". Here is what it does:  Was not sure what to do/where to report? --Captain-tucker (talk) 15:29, 5 September 2008 (UTC)
 * It looks like the link works if you use single quotes instead of double quotes areound Alexander+Hamilton. Here's the test.Laenir (talk) 16:09, 5 September 2008 (UTC)
 * That would be because the double quotes are breaking the HTML; quotes are used in the actual link being generated by the MediaWiki software, hence the borkage (technical term). EVula // talk // &#9775;  // 18:41, 5 September 2008 (UTC)
 * More likely the issue is that double quotes seem to be specifically excluded from literal inclusion in a URL by the software; if I'm reading this correctly, the same will occur if the URL contains <>, [], space, and control characters 0-31 and 127. The easy fix is to urlencode the double-quote characters (as ) in the link, like this: . Anomie⚔ 19:52, 5 September 2008 (UTC)

When I go to that URL, it automatically encodes it as http://books.google.com/books?id=y1_R-rjdcb0C&pg=PA726&vq=The+third+son,+James&dq=%22Alexander+Hamilton%22&source=gbs_search_s&sig=ACfU3U36bnOJW4HE1bpJ004GuxaiXnX7Vg. Looks like your browser may be too smart? --NE2 20:18, 5 September 2008 (UTC)


 * You don't get that on the front end of the Google Book page for that book:
 * http://books.google.com/books?id=y1_R-rjdcb0C&pg
 * Which you can also get to by using Google books
 * --—— Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  20:52, 5 September 2008 (UTC)
 * --—— Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  20:52, 5 September 2008 (UTC)

The purpose of my using that specific link was that I found an article consisting of a paragraph copied word for word from a book that I located on google books. So the URL was created by the google books site by searching for a specific set of words in that paragraph so I could include that link in a db-copyvio tag. I noticed the broken link within the db-copyvio tag. Since I would imagine that many editors use google books to cite references that it might be a problem. --Captain-tucker (talk) 21:10, 5 September 2008 (UTC)

According to RFC 1738, quotation marks are unsafe in URLs, and MUST be encoded:

Unsafe: Characters can be unsafe for a number of reasons. The space character is unsafe because significant spaces may disappear and insignificant spaces may be introduced when URLs are transcribed or  typeset or subjected to the treatment of word-processing programs. The characters "<" and ">" are unsafe because they are used as the delimiters around URLs in free text; the quote mark (""") is used to  delimit URLs in some systems.  The character "#" is unsafe and should   always be encoded because it is used in World Wide Web and in other   systems to delimit a URL from a fragment/anchor identifier that might   follow it.  The character "%" is unsafe because it is used for   encodings of other characters.  Other characters are unsafe because   gateways and other transport agents are known to sometimes modify   such characters. These characters are "{", "}", "|", "\", "^", "~",   "[", "]", and "`".   All unsafe characters must always be encoded within a URL. For   example, the character "#" must be encoded within URLs even in   systems that do not normally deal with fragment or anchor   identifiers, so that if the URL is copied into another system that   does use them, it will not be necessary to change the URL encoding.

Any site that does not do this, and treats unencoded quotation marks as valid URL characters, is broken and should be fixed. MediaWiki correctly prohibits unescaped double quote marks from being considered part of URLs. —Simetrical (talk • contribs) 01:31, 7 September 2008 (UTC)

What links to a specific diff
Hey, is there a fancy tool somewhere that will list every WP page that links to a specific diff?--The Fat Man Who Never Came Back (talk) 05:30, 6 September 2008 (UTC)


 * If you know the exact URL, you can use Special:Linksearch, eg: oldid=102210806. However, there are endless variations on a diff URL (rearrangement of parameters, article vs script path, whether the title was included, diff=next/prev/#, etc) you'd probably need to download the database or ask for a toolserver query (a regex grep on external links table). --Splarka (rant) 07:28, 6 September 2008 (UTC)


 * You can use this revision history search. — Twas Now ( talk • contribs • e-mail ) 10:02, 6 September 2008 (UTC)

Navbox defaults
If a template uses a Navbox and has a default (eg it is set to be collapsed). Is it possible to, how does one, override that default and make one particular instance of the use of the template do some other thing. I append two navboxes, one uses the default and one attempts (but fails) to override the display state.

[redacted] [redacted] -- SGBailey (talk) 07:57, 6 September 2008 (UTC)
 * The <tt>state</tt> is already hardcoded to <tt>collapsed</tt> in the template. You could just include a parseable parameter (ie. something like <tt> state= </tt>) and then you can use <tt> |state= </tt> that you've used above.  Thanks AreJay (talk) 16:38, 6 September 2008 (UTC)


 * Any chance you could have a look at User:SGBailey/Sandplate and User:SGBailey/Sandbox. I've tried what I think you said and it doesn't work. -- SGBailey (talk) 18:52, 6 September 2008 (UTC)


 * Minor glitch in my suggestion above. The collapsible argument should be parsed through the <tt>state</tt> param. I've gone ahead and fixed it...seems to work fine on User:SGBailey/Sandbox.  Thanks AreJay (talk) 19:59, 6 September 2008 (UTC)


 * Thanks. I think I understand the idea here now. -- SGBailey (talk) 20:45, 6 September 2008 (UTC)

Someone check, fix, and test "List of all-female bands"
There is a profusion of links directing to the incorrect site, such as "Delta Darts" and "Delirium." Linking to Delerium from the catagory List of all-female bands should take me to the female band and not require me to disambiguate. I don't care enough about this mess to do it myself. Just thought this "community" would like to know. Now you have another task to accomplish. Good luck with all that ! P.S. Don't leave me any messages, I don't care anymore (I have an account that I abandoned due to rogue administrator abuse). But since nobody "OWNS" the information in an encyclopedia, and this project serves the public good, then I think it is responsible of me to at least summon this effort to notify the offending or responsible parties. —Preceding unsigned comment added by 76.118.179.225 (talk) 14:14, 6 September 2008 (UTC)

2008 main page redesign
Discussion is still ongoing in the proposed redesign of the main page. A brief survey was conducted to get a sense of community opinion, and most designs have been modified to reflect the bulk of community desires. The number of proposed designs has also been reduced, and each design is now accompanied with a screen shot. Please help establish consensus, and move this proposal to the next stage of voting. -- Nick Penguin ( contribs ) 16:19, 6 September 2008 (UTC)

Incorrect counts from PAGESINCATEGORY function
This parser function seems to be returning incorrect numbers. For instance, Category:Wikipedia featured topics all articles has more than entries. The page counts come from the category table and are presumably updated as cats get added to/removed from pages. This means that a server burp will leave the table in an incorrect state.

It appears in category.php that if the page count drops below zero, a recount of the contents is forced. As an extra piece of icing, this recount comes from the categorylinks table, so if referntial integrity has been lost between it and the page table, the count will still be wrong.

Just off the bat though, is there any way to force a recount of contents of a category other than depopulating it? Franamax (talk) 00:03, 5 September 2008 (UTC)
 * Well, something's fixed it. Algebraist 08:33, 5 September 2008 (UTC)
 * Yes, that something was me in the course of a long journey. The most comprehensive retelling is probably the links on this page. The problem is not solved though, in fact I think I may have also run into the "icing" problem I mentioned above. I'll update the pagesincategory discrepancies I've found a little later. See here and most of my latest contribs for the whole sad story.
 * It's quite late here, so I'm booking off for the night. Meanwhile, I'll ask again: is there a way to force a recount of contents of a category? This is crucial. Franamax (talk) 09:05, 5 September 2008 (UTC)
 * Not to my knowledge - I only know how to workaround it. I run into this with the Speedy Deletion categories; I keep a list of them with pagecounts on my userpage. The trick is that I have to figure out what the actual count is versus the result from PAGESINCATEGORY, and then increment it with a SUM function. So, for example:   (The Non-Notable category) returns , which will not match what the actual category shows at Category:Importance or significance not asserted pages for speedy deletion. So my userpage actually says  NaN , which corrects the count by knocking 8 off of the total. The trick is that I have to keep correcting it when I see errors, but it's close. UltraExactZZ Claims~ Evidence 14:55, 5 September 2008 (UTC)

What we should probably do is recalculate it properly with an 0.1% probability or something like that, whenever the code requests the count. It shouldn't be noticeable in the overwhelming majority of cases, and it should be extremely intermittent in the remainder (e.g. Category:Living people takes a few seconds). Trying to avoid the problem by allowing explicit purges basically means that the non-tech-savvy get screwed over.

Another special case that could be handled is that when someone views a category with fewer than 200 pages in it, the software should be able to automatically detect that that's wrong, and should be able to safely recalculate it (since it already knows the correct number from displaying the category page).

Alternatively, we might have a job that periodically refreshes all the counts in the table in an incremental fashion. This has the advantage of being asynchronous, so it won't cause noticeable lag for anyone.

Or maybe someone could come up with a way of maintaining summary tables that actually works accurately. :(

Anyway, if the categorylinks table is wrong you're SOL. We have to rely on that for category info unless we want to try reparsing every single page in the database to refresh category counts. But categorylinks is usually accurate, and when it's not, null-editing the specific page should work. —Simetrical (talk • contribs) 16:14, 5 September 2008 (UTC)

The situation should be improved when goes live. —Simetrical (talk • contribs) 16:33, 5 September 2008 (UTC)
 * Thanks for that Simetrical. It might be helpful to note the possibility of discrepancies in large categories on the pagesincategory manpage, and maybe also that a purge will be necessary to update the target page after viewing the category.
 * I was actually thinking that a tick-box could be put in the Special:CategoryTree page, "Update category counts", probably admin-only since it could be somewhat expensive. Franamax (talk) 18:19, 5 September 2008 (UTC)
 * It wouldn't be very expensive at all. Comparable to parsing a large page, say.  Users formerly did that query all the time on Special:Categories, before the category table existed.  But adding manual purge interfaces all over the place is really not the best sort of design.  There should be an automated way of ensuring that this is accurate without manual intervention. —Simetrical (talk • contribs) 01:25, 7 September 2008 (UTC)
 * I'm thinking here about categories with population > 200 which have lost count integrity. Your enhancement (40499) is completely pain-free. But there is currently no automated way to spot cat>200's in need of recounts and update category, is there? Leaving forced recounts open to all via special page wouldn't work, at any given time there would be five people trying to get an exact count for the really big ones, which the job queue would make a constantly moving target. The missing spot is probably categories with 200-1999 members or so, where users would still have an interest in reporting exact counts, and the membership wouldn't be changing especially fast.
 * What's the solution then? A toolserver bot that would regularly and reandomly recount categories and heuristically modify its rate based on the errors it encountered (selection biased toward the lower-count cats>200)? Change the &purge action so that it would force cat recounts? I dunno... Franamax (talk) 21:48, 7 September 2008 (UTC)
 * You could just have jobs that periodically recalculate all category counts one by one regardless of whether they're known to be wrong. It only takes a few seconds max, for the largest categories, and for almost all it takes a lot less time.  Doing each category once a day should be harmless. —Simetrical (talk • contribs) 22:42, 7 September 2008 (UTC)

Autoformatting standard violation
I believe that the wikimedia software has volunarily adopted ISO 8601 with respect to autoformatting of dates, but is violating the standard whenever it transforms a date with a year less than 1583 into or out of the ISO 8601 format. I have commented on this in bug 4582. If the developers who are following that bug decline to fix the problem, what is the proper forum to seek binding arbitration? --Gerry Ashton (talk) 17:53, 5 September 2008 (UTC)
 * There is no "binding arbitration". This is a wiki. — CharlotteWebb 22:56, 5 September 2008 (UTC)
 * Indeed, and our developers are not "bound" to implement anything, either.  krimpet ✽  0:72, Primidi, 21 Fructidor CCXVI
 * What does "binding arbitration" mean? SharkD (talk) 02:48, 6 September 2008 (UTC)
 * binding arbitration —Simetrical (talk • contribs) 01:37, 7 September 2008 (UTC)

There is no binding arbitration. The developers decide what features to implement. The final decision in the event of a dispute is up to Brion Vibber, who is lead developer of MediaWiki. Of course, MediaWiki is free and open-source software, and you're free to maintain your own patches for your own installations, or for that matter fork the project if you like. If you want a change on Wikimedia wikis, Brion Vibber is also CTO of the Wikimedia Foundation, so if he says no then you're going to have to either change his mind or give up.

I don't understand how compliance with ISO 8601 has anything to do with autoformatting unlinked dates, though. That bug is ridiculously long and I'm definitely not going to read through all of the comments. —Simetrical (talk • contribs) 01:37, 7 September 2008 (UTC)


 * Simetrical, thanks for the succinct explanation. That bug does cover a number of issues; let me summarize the ISO 8601 issue.


 * Between statements in mediawiki, the English Wikipedia, and reference sources like the Chicago Manual of Style, one might (or might not) conclude that dates in the format YYYY-MM-DD are governed by ISO 8601. If that format isn't governed by ISO-8601, then who knows what it's governed by in Wikipedia. When one writes that the Battle of Hastings was fought on 14 October, 1066, people expect it to be in the Julian calendar. If it is autoformatted to 1066-10-14, and a well-informed reader thinks ISO 8601 applies, the reader will reject the date, in part because ISO 8601 isn't supposed to be used before 1583 except by mutual agreement, and also because the battle was fought on 20 October in the proleptic Gregorian calendar, which is always used with ISO 8601. Other readers won't think the standard applies, and think the reformatting is OK. --Gerry Ashton (talk) 03:45, 7 September 2008 (UTC)


 * Gerry Ashton: I read what you wrote over at bug 4582. You have misunderstood some basic things: The ISO 8601 standard is from 1988 and formulates a way to write dates that already was used long before that in several parts of the world. I am from Sweden and I started school in 1976, and then we were already using dates in that form. That is, for me as a Swede the date and time of your comment above would be written 2008-09-07 03:45 (UTC). And to me as a Swede it is perfectly clear what it means if a text says: "The Battle of Hastings was fought on 1066-10-14". And I think the Chinese have been writing dates like that for a couple of thousand years, long before the Gregorian calendar was even invented.
 * So if a Swede or a Chinese set their "my preferences - Date and time" to show dates like 2008-09-07 then you should not force some other way to display dates if a date happens to be before 1583. The whole purpose of the "my preferences - Date and time" setting is to allow us to see dates in the way we prefer.
 * --David Göthberg (talk) 11:35, 7 September 2008 (UTC)


 * What do you want us to do, try to figure out automatically whether a date was intended to be Julian or Gregorian? This is a user-level issue, not a software issue.  Dates in articles don't need to be machine-parseable, they're meant for humans to read, and humans can figure out from context whether the date is Julian or Gregorian ― or if not, the article should clarify.  Even if the software could somehow fix this reliably, which it couldn't, I don't see any value in it. —Simetrical (talk • contribs) 22:45, 7 September 2008 (UTC)

Table shorthand
Normally in tables you have to place the  markup for a new row on its own line. I propose a change to allow placing the row and cell markup on the same line. I.e.,

This change would mean a lot less scrolling for large tables. SharkD (talk) 02:43, 6 September 2008 (UTC)


 * Another good option is to use "HTML wikimarkup":
 * We often use HTML tables since among other things that means we don't need to escape the pipes "|" in the tables by using, which makes both more readable code and saves a lot of transclusions of which costs less server load. (The pipes need to be escaped in several situations, for instance if a wikitable is placed inside a #switch statement.)
 * And even when we don't need the escaping some of us thinks HTML wikitables are more readable than lots of braces and pipes, since the same kind of braces and pipes are used for the parserfunctions and the image notation and so on. And as you kind of point out above, in wikitables the whitespace such as newlines has meaning! Ouch.
 * --David Göthberg (talk) 16:08, 6 September 2008 (UTC)
 * --David Göthberg (talk) 16:08, 6 September 2008 (UTC)


 * A problem with using HTML is that there isn't always a way to predict when to close a row. This is not a problem with wikimarkup. SharkD (talk) 03:17, 7 September 2008 (UTC)

Another thing I noticed is that when nesting tables, a blank line is always placed at the beginning and end. I.e., this works (but also causes the blank line):

This doesn't work: {| SharkD (talk) 03:15, 7 September 2008 (UTC)
 * |}


 * Regarding when to close rows: The HTML tables we use in Wikipedia pages aren't really XHTML tables, rather they are "HTML wikitables". That is, they are parsed by MediaWiki. Which means you don't need to close the table cells and table rows, MediaWiki does that for you. Just as it does with wikitables that use braces and pipes. For instance if you use this code:


 * Then MediaWiki parses that and renders this XHTML code in the output web page:


 * Neat isn't it? It also repairs some other mistakes you might do when coding tables, just like it does for wikitables with braces. And if you use HTML wikitables then you don't get any of the blank line problems that you mention in your comments above. Compare these two examples:


 * Note! The above example is really bad table coding, it even lacked the  tags (start of row). So please don't code your tables in that way. But as you can see, MediaWiki does fix that and outputs clean XHTML table code. Just for reference, here is how I really would code it:


 * I should perhaps mention that when we built the ambox we discovered there were some case when we needed to close the table cells. (I don't remember the details but I think it had something to do with whitespace in parameters fed to the ambox.) So if you get weird table behaviour then do close your cells and rows.
 * --David Göthberg (talk) 10:43, 7 September 2008 (UTC)
 * Thanks for the explanation. It's good to know. SharkD (talk) 00:15, 8 September 2008 (UTC)

displaying the number of pages in subcategories
I noticed the German Wikipedia displays both the number of subsubcategories and also the numbers of pages and of files contained in a subcategory next to that subcategory, in brackets (example). Are there any plans to implement the same here? <span style="font-family:lucida sans, console;">user:Everyme 10:46, 6 September 2008 (UTC)
 * I notice that commons also has this extension. <span style="font-family:lucida sans, console;">user:Everyme 04:31, 7 September 2008 (UTC)
 * It's probably done by JS. It should be easy enough to write up for the software. —Simetrical (talk • contribs) 22:49, 7 September 2008 (UTC)

Template/table transposition help
I would like to transpose (i.e. switch the right-bottom directions) the table created by this template. The problem is that there's no way to know beforehand how many and which of the fields will be filled. Thanks. SharkD (talk) 03:10, 7 September 2008 (UTC)
 * I've got a few ideas, but it really depends on how it is being used. Is this template going to be added to articles frequently? Will the list of systems passed into it be changed frequently? There are a few optimizations that can be done by the person using the template that would vastly simplify the solution. --- RockMFR 03:40, 7 September 2008 (UTC)
 * I've wrote up something at User:RockMFR/sandbox that you can take a look at. Syntax is <tt> </tt>, where the first parameter is the number of systems. --- RockMFR 05:28, 7 September 2008 (UTC)
 * Thank you! That worked very well. I've modified your example slightly and have along the way introduced a new CSS problem in Internet Explorer only. The problem is that the border along the far right side disappears for some reason. I believe using widths of 100% causes the table(s) to become clipped somehow. This problem does not occur in Firefox. You can see an example of the problem here. Thanks again for your help! SharkD (talk) 22:11, 7 September 2008 (UTC)
 * The problem seems to be gone now. Thanks again! SharkD (talk) 00:10, 8 September 2008 (UTC)

Decision table
Are the WP:xxx rules stated anywhere in the form of a decision table? —Preceding unsigned comment added by 71.100.2.227 (talk) 09:22, 7 September 2008 (UTC)


 * No. How would that work? — Twas Now ( talk • contribs • e-mail ) 09:37, 7 September 2008 (UTC)
 * Very easily. Just list all of the conditions upon which the actions will be based. Then list all of the possible combinations of conditions. Next, list all of the possible actions and then indicate which actions are justified by each combination. The indicator can be a sequence number.


 * In preparation for bots you need do little more but for humans its best to make a polychotomous key. —Preceding unsigned comment added by 71.100.11.142 (talk) 15:20, 7 September 2008 (UTC)
 * Wikipedia's rules are rarely, if ever, strictly followed like that. <font face="Broadway">Mr.Z-man 16:05, 7 September 2008 (UTC)


 * Right, well I know there is some diplomacy and politics involved, however, in order to be diplomatic and consistently fair the rigor of a decision table is very helpful, i.e., previous cases can be used to help fill it out. —Preceding unsigned comment added by 71.100.11.142 (talk) 16:08, 7 September 2008 (UTC)


 * No, I don't think you understand. We do what makes the most sense in an given situation, not necessarily what is the most consistent. If we formalize a set of actions and consequences, people will expect that to be followed, when we don't do that. <font face="Broadway">Mr.Z-man 16:32, 7 September 2008 (UTC)


 * What you are forgetting is that there is always a difference between the real and the ideal the plan and the practice. How much do you edit? My last major contributive edit was merely to express a macroeconomic chart in the form of a table. My effort was decried as "original research." What followed was an edit war I could not win. Even if I were being paid to edit i might have second thoughts about the possibility of making a meaningful improvement. With rules stated in the form of a decision table on the other hand I might have a chance to present a legal case. As is whoever wins the edit war is the user that wins. —Preceding unsigned comment added by 71.100.11.142 (talk) 17:28, 7 September 2008 (UTC)


 * You don't name the article, and you didn't edit it under this IP address, which doesn't help the rest of us much. Did you try No original research/noticeboard?  Third opinion?  Request for comment?  I am sorry that an edit you feel strongly about did not meet with the approval of other editors, but that happens to all of us.  Have you considered the possibility that they might be right?  Robert A.West (Talk) 18:04, 7 September 2008 (UTC)
 * Legal case? What part of "Wikipedia is not a bureaucracy" don't you understand? Its not about winning and losing. <font face="Broadway">Mr.Z-man 18:42, 7 September 2008 (UTC)


 * All you need to know is Five pillars (most relevant to this discussion: "Wikipedia does not have firm rules") and Simplified ruleset. For more detail, you can see List of policies. — Twas Now ( talk • contribs • e-mail ) 01:38, 8 September 2008 (UTC)

Help fix edit-conflict induced loss of material
Hi, this diff involves the loss or corruption of several comments or sections of AN/I, it appears to be the result of an edit-conflict. Anyone know how to restore the unintentional damage? Thanks. DuncanHill (talk) 13:27, 7 September 2008 (UTC)

Screen Garbage / Formatting
Hello, VP Tech readers!

Two of us editors were resolving a problem over at AN/I. The problem had been resolved, and the other editor had marked the incident "Completed." I then posted this diff. IIRC I was using the Section edit button. I hit "Save Page." What I then saw (I think) was the base page, followed by all the content text of the page overlaid on the screen without any formatting.

I am in process of making/saving a screenshot of what it looked like. (Though I'm not entirely sure if I can upload it under "all my own work" or what license applies. *argh* )  I hope my description is sufficient for now.

Fortunately, I was able to undo my edit before anyone else posted. The page *still looked the same*. I Control-F5ed. Same. Thinking that I had just accidentally nuked AN/I somehow, I went and left a message on the other editors page. Then I reloaded the page... and it looks fine now? My diff looks fine now? (If relevant, I often have multiple tabs open, and sometimes multiple tabs of the same page for compares, diffs, etc.)

I'm thinking this is a bug - but have no clue if it's a bug with MediaWiki, something to do with Firefox 3, both or neither. Hence my post here. Is this a bug anyone else has experienced? <span style="background:	#ADDFAD;color:yellow">Laughing Vulcan 15:38, 7 September 2008 (UTC)

Watchlist individual sections of article and talk pages?
I would like to be able to watchlist individual sections of talk pages. I dislike having to check talk pages so often due to all the changes that occur in sections I am not interested in. I avoid a lot of discussion pages because of all the time wasted for this reason.

It would be nice if there was a filter that blocked anything showing up on my watchlist that is not from that particular talk section. --Timeshifter (talk) 23:23, 4 September 2008 (UTC)
 * In theory some kind of tool could be written but it would be dependant on people properly using the "edit section" button... – xeno  ( talk ) 23:25, 4 September 2008 (UTC)


 * On clicking the "edit" button next to a section one sees the edit form for that section. At the bottom is a checkbox for "Watch this page". Maybe there can be an additional checkbox for "Watch this section".


 * Now that I think about it, this would be a good option for article sections too. Article editing can get busy at times too. And I am often interested only in watching one section of an article. --Timeshifter (talk) 23:44, 4 September 2008 (UTC)


 * The consensus seems to be that this would be too hard to do. See Perennial proposals. Algebraist 00:56, 5 September 2008 (UTC)


 * It would all go to hell if someone merges sections. Watchlisting articles is about all we can do; anything else is too much work for not enough payoff (hence its position on the perennial proposals page). EVula // talk // &#9775;  // 18:44, 5 September 2008 (UTC)

(unindent) Some links from Perennial proposals lead to a possible solution in the works for talk pages. I don't see anything currently being worked on for watchlisting sections of articles. From reading the discussions it seems that the solution may be simple. If the sections were numbered more specifically and uniquely inside the wiki software and database, then the problem is solved. As long as that unique number is not changed due to work on other sections, then watchlisting is feasible. Merged sections may delete that number but that would show up on the watchlist too, and it is something people would want to know about. I think the solution for articles is much simpler than for talk pages, due to the vast number of options being introduced with the talk page solution. --Timeshifter (talk) 00:00, 6 September 2008 (UTC)


 * Did you look at the three Bugzilla reports linked from there? Implementing this (for articles or talk pages) would require redesigning the database infrastructure, which no-one is willing to do. Algebraist 00:05, 6 September 2008 (UTC)


 * Yes. The clearest statement is the one on the perennial proposals page: "Reasons for previous rejection: While most people agree this would be a good idea, the technical implementation would be very difficult. Currently watchlisting is done per-page and each page has a unique ID number in the database. Sections however, are very fluid. They are assigned a number on the page they are on, but adding or removing sections above it will change the number. The name of the section can also be changed easily and new sections can be created with the same name." So the solution is to assign a larger number that does not change when other sections are added. That would solve 99% of the problems. The other 1% can be ignored for now. That being the problems with people changing the section name, or creating sections with the same name. That doesn't happen very often, and people wouldn't mind re-watchlisting those sections. The benefits of single-section watchlisting outweigh the problems. The real problem is overburdened technical staff at Wikimedia. They need help from more paid staff. But in my opinion ideologues among a few vocal Wikipedia editors, and possibly in the Wikimedia Foundation, are opposed to the few realistic ways left to raise more money. That being search tool income from Google, etc., and opt-in ads on Wikipedia pages. See Advertisements. I am sorry to bring this up again, but I get tired of hearing about the same problems causing us not to do things. --Timeshifter (talk) 02:44, 6 September 2008 (UTC)

The fact of the matter is that we don't currently track sections in any meaningful way, and the benefit of being able to watch individual sections is far outweighed by the complexity and extent of the changes that would need to be made to allow that. Of course it could be done, but it's not worth it. I mean, geez, we don't even allow watchlisting a page independently of its talk page. There's no good reason that the watchlist feature needs to be very granular ― if there's some reason to watch individual sections of a page, that's maybe a hint they should be broken off into their own articles (possibly transcluded, the way many maintenance pages work). —Simetrical (talk • contribs) 01:22, 7 September 2008 (UTC)
 * And no, this has nothing to do with lack of development time. It has to do with development prioritization.  If we had more developers (which we're starting to get) we'd be assigning them to lots of great new features, which almost certainly would not include this one.  It's just not worth the effort to write or maintain it, or the added interface complexity for users, given that the current system works really well. —Simetrical (talk • contribs) 01:24, 7 September 2008 (UTC)


 * It is a perennial request and there are many reasons it would be good to do it. I disagree that it is only a matter of development prioritization. It is also a matter of development money to pay for more developers. There are many perennial requests that are not getting done. I asked before elsewhere, and no one answered: What if money could be raised for specific development purposes? --Timeshifter (talk) 04:11, 7 September 2008 (UTC)


 * Reward board and see if you can find someone. As a preliminary estimate, I could probably do it in six months for $50,000, but this would involve significant amounts (days to weeks) of Wikipedia downtime to convert the database to the new format. --Carnildo (talk) 04:55, 7 September 2008 (UTC)
 * It sure wouldn't have to take six months full-time. And we don't have any downtime for schema updates here, you know.  We just take slaves out of rotation and apply them one by one.  Even if we did, I'm pretty sure no ALTER TABLE is going to take days with our data set and hardware. —Simetrical (talk • contribs) 22:40, 7 September 2008 (UTC)
 * Note that most of the developers for MediaWiki aren't paid. <font face="Broadway">Mr.Z-man 05:26, 7 September 2008 (UTC)


 * Well, you'd have to get it approved in principle. That might be difficult.  Even if you could pay someone to write up a proposal, it would have to be reviewed by existing core devs, and if it's a large patch, that will take time if they're willing to do it at all.  Feel free to try paying someone to do it, if you have a few thousand bucks burning a hole in your pocket.  Just don't be sure you'll succeed. —Simetrical (talk • contribs) 22:40, 7 September 2008 (UTC)


 * Thanks everyone for these detailed responses. I believe there are people who want to help with money. But now that you tell me that most developers are unpaid I am even less hopeful that much will get done anytime soon with many of the perennial requests unless the Wikimedia Foundation tries out some alternative methods of raising serious money. See again: Advertisements. Also, the Bugzilla reporting system for the MediaWiki software is seriously flawed too in my opinion. I found out that Firefox developers also use the Bugzilla reporting system, and that explains to me why the Firefox 3 browser has some serious problems that prevent me from upgrading from Firefox 2 to Firefox 3. See Wikipedia talk:Bug reports and feature requests. So we have a perfect storm of problems that prevent many people from adequately reporting bugs/problems/suggestions, watchlisting individual bugs/problems/suggestions, discussing bugs/problems/suggestions, and solving bugs and problems. The tools and the funding are just inadequate. Angel funding has increased lately for Wikipedia. But that kind of money is hard to depend on, and even harder to increase. Now Google has put out a browser too. The competition is great, but we at Wikipedia and Firefox need to step up our game. We need better discussion tools, and more funding. --Timeshifter (talk) 00:41, 8 September 2008 (UTC)


 * Sorry, you can't watch individual sections of pages, nor is it something we plan to add at this time.


 * Note that for talk pages specifically, more advanced thread-friendly systems are under development (LiquidThreads) which may end up deployed at some point in the next year, depending on how ongoing development continues. (Work is currently funded to a contractor, David McCabe.) This will likely solve the problem you would want to solve with section watching (following a specific discussion). --brion (talk) 21:41, 8 September 2008 (UTC)


 * Thanks for the reply. I know you are a very busy guy. Is it possible that LiquidThreads could be adapted in any way for use in watchlisting article sections? I did not get that impression. Village pumps are talk pages, so I guess LiquidThreads could be used for most Wikipedia projects and forums. This would be a great help. Article section watchlisting isn't as important except on the busier articles. Sure would be helpful though in watching popular controversial articles, and watching frequently vandalized articles. --Timeshifter (talk) 04:19, 9 September 2008 (UTC)

MACID discussion, recast
Are some WP administrators penetrating MACIDs of sockpuppets opposed to secular facism? 68.103.31.116 (talk) 09:33, 4 September 2008 (UTC) There is nothing to stop an admin from '''using resources from a company or an intelligence agency to use tools to push his POV and knock out editors opposed to his POV. (?)''' Much like a criminal prosecutor, cases are picked to prosecute, and other cases go free. The latter matches his POV and the former stands against his POV. MsTopeka (talk) 12:02, 4 September 2008 (UTC)


 * I'll take the silence in answering to mean fear of retaliation. Awful. Almost Biblical... Why the Phantom of the Opera? Why take that name? He thinks this an opera? He thinks his role is as the Phantom? Problem is - I forgot the plot. No, now I recall: There is no plot. No meaning whatsoever except very silly fear induction for children.


 * More answerable: Why would an archived debate get a "hat"? Is that to block search engines? MsTopeka (talk) 00:07, 5 September 2008 (UTC)


 * (P!=NP)== http://en.wikipedia.org/wiki/P_%3D_NP_problem#Consequences_of_proof + http://en.wikipedia.org/wiki/Image:Genetic_Code_Bias_2.JPG 68.103.31.116 (talk) 03:11, 5 September 2008 (UTC) MsTopeka (talk) 03:14, 5 September 2008 (UTC)


 * Mr.Z: Please do not hide this discussion. This is fundamental question, which is of a technical nature that is very appropriate for this forum:  Are some WP administrators penetrating MACIDs using resources from a company or an intelligence agency as tools to push his POV and knock out editors opposed to his POV? MsTopeka (talk) 18:22, 5 September 2008 (UTC)


 * I have no idea what MACID is; perhaps the silence was because nobody else knew what the question was, too... EVula // talk // &#9775;  // 18:42, 5 September 2008 (UTC)
 * MAC address, presumably. Not that that helps understand the OP's remarks. Algebraist 18:57, 5 September 2008 (UTC)
 * Correct. And many of us edit here thinking we are anonymous editors following the WP rules.  So, what do we have in place at WP to stop an Admin from using a more powerful tool - that is available to certain corporations and/or governments - to come through your MACID to your personal information, and then decide they don't like your POV, put you up as a socketpuppet or other WP violation, and then turn you off forever, having disposed of said POV.  If a few of you are willing to watch and appeal this process, I will demonstrate what will happen if I type in a few usernames who can do this.  They might be afraid to do it right here on pump, but let's see. MsTopeka (talk) 19:22, 5 September 2008 (UTC)
 * I'm not sure what "tool" you're talking about, unless the user happens to be editing from the server farm in Florida, there's no way for us to get the MAC address, only the address of the most recent router (or something like that). And putting a hat on a discussion won't stop it from being indexed by search engines, though putting NOINDEX on a page will. (And if you want your question to look like a serious one, take off the tinfoil hat before posting.) <font face="Broadway">Mr.Z-man 19:41, 5 September 2008 (UTC)
 * The answer to that question should be perfectly clear. There is only one "tool" evident in this thread.  S HEFFIELD S TEEL TALK 14:36, 8 September 2008 (UTC)
 * Are you referring to the use of "checkuser" to determine if multiple accounts are coming from the same (or very similiar) IP address? Useight (talk) 00:34, 6 September 2008 (UTC)
 * No, I am speaking of a threat of violence on a wedge issue edit of mine, parallel communications with Godwin, being tracked by a local forum www.kcfs.org, reminding people of US law, and then getting kicked out for "making a legal threat", aided by a consortium of anti-Chrsitian hierarchial editors who believe Darwin is fact - a very strong secular fascist view - supported by Godwin's Law, a pun. MACIDs came up, too, along with the strong hand of computer security people playing POV police on WP. MsTopeka (talk) 02:50, 6 September 2008 (UTC)
 * Making accusations without proof is worthless. And your equation of people who believe in evolution with fascists is not acceptable, and I request that you retract your comments.   Corvus cornix  talk  18:10, 6 September 2008 (UTC)


 * Except my apologies. But please see read dysgenetics and understand how frustrating it is to see a child in public school learning about Darwin with no mention of Adolf Hitler while our WP general counsel rules here with the mighty arm of his Satanic Godwin's Law and abortion for parts is legal.MsTopeka (talk) 23:37, 7 September 2008 (UTC)
 * Now you are getting into personal attacks. Please stop while you're behind.   Corvus cornix  talk  06:29, 8 September 2008 (UTC)

DONE MsTopeka (talk) 14:44, 6 September 2008 (UTC)

nstab-main
I've noticed a few comments in the past months about what the first tab should read on pages in the main namespace. Currently, it says "article", but not everything in the mainspace is an article; disambiguation pages, lists, and the main page are all in mainspace too. The MediaWiki default for this is actually "page", which encompasses all of this whereas "article" is really a bit misleading.

So, what do you think? Is "page" a better choice, or should we stand by "article"? —Remember the dot (talk) 02:27, 6 September 2008 (UTC)
 * I oppose this change. It should say "article" when the page is an article.  It would be nice if disambigs, lists, etc. said the correct thing, but that's not essential.   Superm401 - Talk 05:39, 7 September 2008 (UTC)
 * Why is it essential that article pages say "article" at the expense of other mainspace pages? —Remember the dot (talk) 06:04, 7 September 2008 (UTC)
 * Most pages in main namespace are articles, so I think the tab should say "article". As for all other pages, well, there are always some exceptions, for example User:Redirect fixer and User:MediaWiki default are not real users, users subpages are also not real users, and Image: namespace contains a lot of sound files. —AlexSm 06:20, 7 September 2008 (UTC)
 * Users' subpages at least are marked "user page" and have a link back to the user's userpage, which is appropriate. The Image namespace's name is unfortunate, it should really have been named something more generic, like "File". My point is that we can avoid having counterintuitive exceptions by choosing text (in nstab-main's case, the default text) that fits everything. —Remember the dot (talk) 07:57, 7 September 2008 (UTC)

I don't think many people really care. &mdash; Werdna  &bull;  talk  09:18, 7 September 2008 (UTC)
 * I know I don't. The main purpose of the tab is to distinguish between the subject page and the talk page; whether it says "article" or not isn't something I can get excited about. :) --brion (talk) 21:32, 8 September 2008 (UTC)

?Second/Third Party Cookie?
My browser is Firefox--with tight cookie control. While navigating between internal links only @ the articleA Great Day in Harlem, a Firefox cookie message box informed me that a third party cookie was trying to install in my computer. Of course I blocked it. It's name was www.lizwolfe.com.

Has anyone seen anything like this? —Preceding unsigned comment added by Tapered (talk • contribs) 02:00, 8 September 2008 (UTC)


 * Which links did you click? — Twas Now ( talk • contribs • e-mail ) 06:48, 8 September 2008 (UTC)
 * I'm not seeing (in Special:Linksearch) any pages currently linking to this domain. Perhaps it was linked from an older revision of A Great Day in Harlem, but I didn't look very far. — CharlotteWebb 17:26, 8 September 2008 (UTC)

Boxes overlapping category contents
How do I stop the Wiktionary and Commonscat boxes overlapping the members of the category at Category:Cornish language please? DuncanHill (talk) 02:26, 9 September 2008 (UTC)


 * Fixed Gary <b style="color:#02b;">King</b> ( talk ) 03:18, 9 September 2008 (UTC)
 * Splendid, thank you! DuncanHill (talk) 03:21, 9 September 2008 (UTC)

Infobox not working
Can anyone work out why the infobox on Edwin Dunkin is not working please? DuncanHill (talk) 03:24, 9 September 2008 (UTC)
 * I fixed it - I had a ] when I should have had a ]] to close a wikilink in it, and that broke the whole box. DuncanHill (talk) 03:27, 9 September 2008 (UTC)

Custom edit messages
Per the August 11 Signpost, we can now add custom edit notices per page. This looks like a much better solution than using comments in the article. Does anyone understand how to implement this? --—— Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  11:47, 1 September 2008 (UTC)


 * If MediaWiki:Editnotice-2 is created, it will appear when editing any user-page. If MediaWiki:Editnotice-2-Gadget850 is created, it will appear when editing your user-page. Try it [[Image:Smiley.svg|15px|]]. — CharlotteWebb 14:40, 1 September 2008 (UTC)


 * Check out template editnotice. It has documentation that explains how this works. (And no, I didn't make that template.) One thing worth mentioning is that only admins can create editnotices since they are placed in the fully protected "MediaWiki:" namespace.
 * --David Göthberg (talk) 15:10, 1 September 2008 (UTC)


 * Ah well thanks for that link. Now I feel stupid for going to read the source code [[Image:Frowny.svg|15px|]]. On the other hand if the notices are being added through a template it would be possible to make certain messages editable by regular users (or at least auto-confirmed) by using a vivid array of parser-functions and template sub-pages. Strictly by the "it's a wiki, get over it" principle this wouldn't be a bad idea if the key intent is to at least partly replace <tt>&lt;!-- hidden html comments --&gt;</tt> which can be added or removed by anyone. I don't see it happening though. — CharlotteWebb 15:33, 1 September 2008 (UTC)


 * Thanks to all. This will be helpful in maintaining list articles that have criteria for inclusion. --—— Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  15:49, 1 September 2008 (UTC)


 * CharlotteWebb: Well, I too went reading the source code and figured out that the naming was something like "MediaWiki:Editnotice-Namespace-Pagename". But how to write the namespace and the pagename was not clear. So I decided to search with Special:PrefixIndex/MediaWiki:Editnotice to see if someone else already had tried the different spellings. And thus I found the existing notices and the template used on them. Now afterwards I feel slightly silly since it would have been much simpler to use Special:Search and search all namespaces for "editnotice".
 * Gadget850: Yeah, this will be very useful. But I think the devs have been a bit too fast in deploying this one. Even though I have worked a lot with namespace detection I find it confusing to work with namespace numbers. Sure, it makes the server code simpler, but it can't be that much code to translate between local namespace names and the numbers. Using namespace names would be much more user-friendly.
 * --David Göthberg (talk) 15:55, 1 September 2008 (UTC)
 * And the template is even more useful when I remove  from my .css. I can't remember what I was testing with that. --——  Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  16:15, 1 September 2008 (UTC)
 * I've removed the id from the template - I don't know why it was there. --Random832 (contribs) 18:45, 3 September 2008 (UTC)
 * Yeah, it would probably better to omit the namespace number and just use the &#123;&#123;FULLPAGENAME&#125;&#125; e.g. "Editnotice-Wikipedia:Village_pump_(technical)" instead of "Editnotice-4-Village_pump_(technical)". — CharlotteWebb 16:21, 1 September 2008 (UTC)
 * If we want to have the naming convention changed it should be done now, not after we have thousands of these message pages. — CharlotteWebb 16:26, 1 September 2008 (UTC)
 * I would favour this change, assuming there's no technical problem with it. Has anyone posted it to Bugzilla yet? Algebraist 18:27, 1 September 2008 (UTC)
 * I'm sure the impending response would be "but namespace titles might change, especially on foreign language projects with incomplete localization is incomplete (or at least needing better translation by a more fluent speaker)", however I doubt a project would reach the point where it needs Editnotice pages (certainly not in a level of bulk where renaming the namespaces would be a big deal) without already having high volume of editing and hopefully a critical mass of native/near-native speakers who have already straightened out all the L10n issues. But I could be wrong. Even so it would be easy enough to automatically rename directly from the database whenever namespace are changed.
 * I have asked Krimpet to comment here as she is the one who created this feature. — CharlotteWebb 19:14, 1 September 2008 (UTC)

(Could this be applied/adapted to the request at MediaWiki talk:Common.js? I would guess not, from the implementation explanation [it looks page specific, not generalizable to page-groups], but someone who knows this might be able to help me with that. :) -- Quiddity (talk) 18:12, 1 September 2008 (UTC)
 * Guessing not would be correct. — CharlotteWebb 19:15, 1 September 2008 (UTC)

CharlotteWebb asked be to comment - I'm back after being away for a few days, so here I am. :) I chose this somewhat nomenclature for edit notices for several reasons. The foremost is that, indeed, namespace names can be easily altered by system messages. Due to MediaWiki's architecture, it's desirable to be as non-dependent on namespace names as possible. (Imagine, for example, if we decided to add an extension of some sort to deploy identical editnotices from a centralized location on different wikis with different namespace names.) Also, the name of MediaWiki: messages is already subject to some arbitrary limitations (slashes cannot be contained in messages, for example, hence messages for subpages must replace '/' with '-'); thus, the full page name cannot be used in the message name. For this reason, I adopted the current "Editnotice-#-TITLE" nomenclature. (If tracking down names of messages proves to be too confusing, perhaps putting all the notice in the per-namespace notice, along with a ParserFunctions switch to select the desired message per page, would be better?)  krimpet  ✽  18:22, 3 September 2008 (UTC)

I thought that mediawiki had standard namespace names (what we use here in english, but "Project" instead of "Wikipedia") that exist independent of localization. --Random832 (contribs) 18:31, 3 September 2008 (UTC)
 * Not for custom namespaces (e.g., "Portal"). To be honest, though, I think it probably makes more sense to use the localized canonical namespace name here ― what's the problem if it gets changed?  Some custom edit notices don't work for a couple of days? —Simetrical (talk • contribs) 15:45, 5 September 2008 (UTC)

Do we have a policy on letting users have edit notices on their own userspace pages? I've worked out a system for allowing a non-admin to maintain the editnotice without requiring a protected edit request every time and without allowing anyone else to edit it. (should this be allowed?) --Random832 (contribs) 18:39, 3 September 2008 (UTC)


 * This is a new feature, so we don't appear to have any guidelines. I can't see any issues with edit notices on a user page that we don't already address, such as civility. Many users have notices of some sort on their talk page addressing centralized discussion and the like. I used UserTalkReplyhere before edit notice. --—— Gadget850 (Ed)  <sup style="color:darkblue;">talk  -  20:30, 3 September 2008 (UTC)

Rather than condoning every man and his dog creating a new MediaWiki: page, why don't we put something at MediaWiki:Editnotice-3 that checks for a .js or .css subpage for that user and loads it as the editnotice if it exists? I think something like:

would be sufficient. It's a bit of a hack of the .css-page-edit-restrictions, but sensible under the circumstances. It's essentially what Random833 has already implemented. <small style="color:red">(also) <b style="color:forestgreen;">Happy</b>‑<b style="color:darkorange;">melon</b> 07:54, 4 September 2008 (UTC)


 * Hacks are :-/ And having a parser function called on every &action=edit is not great for performance... --MZMcBride (talk) 08:09, 4 September 2008 (UTC)


 * It would only be a parser call on edits to the User: namespace. On the other hand, given that the existing pseudo-editnotices are not protected, perhaps there's no need to make it a .css subpage. On the other hand, the who-can-edit-what-in-userspace system could do with a complete overhaul... <small style="color:red">(also) <b style="color:forestgreen;">Happy</b>‑<b style="color:darkorange;">melon</b> 08:19, 4 September 2008 (UTC)


 * A typical action=edit is followed by an article save, which doesn't just run a single parser function, it runs the whole parser on a potentially very large page. I wouldn't worry about performance here.  However, I'm not sure what the implications are of allowing unprotected pages to be displayed as interface messages . . . there are probably none, but it worries me slightly.  Is this already done elsewhere? —Simetrical (talk • contribs) 15:47, 5 September 2008 (UTC)


 * Simetrical: Yes, there seems to be a serious security issue with the system messages. They are sent to the browsers without going through the proper parsing. For instance HTML tags are not converted to XHTML tags, which to me indicates that it might be possible to add other non-allowed HTML instructions and script code and other things in MediaWiki messages. Things that normally is filtered away by MediaWiki when someone tries to put that in a page. Things that make it possible to attack the browsers. Although I have not tested yet if MediaWiki still do the security filtering on the MediaWiki messages. Before we have checked that we should not open up those notices for anyone to edit.
 * --David Göthberg (talk) 16:11, 5 September 2008 (UTC)


 * They're in the mediawiki space, so I don't think they can be opened up to anyone to edit. However, I have another question - Do we have a policy page or discussion on where these are to be used? I have editprotected requests for notices on several articles, such as Chronology of console role-playing games, to place a notice about Japanese names on the edit-notice. Is there a discussion somewhere to determine the consensus for what goes here? UltraExactZZ Claims~ Evidence 17:50, 8 September 2008 (UTC)


 * No. Should there be? Of course if I disagree with an edit-notice I'm more likely to disregard it than complain about it (unless the concern is excessive length in which case I'll be bitching sure enough). I expect the content of most individual edit-notices will be instruction creep, so surely a notice about how to use edit-notices would be meta-instruction creep [[Image:Smiley.svg|15px|]]? Surely this discussion belongs at MediaWiki talk:Editnotice-8-Editnotice-0-Chronology of console role-playing games... -— CharlotteWebb 18:06, 8 September 2008 (UTC)


 * Ultraexactzz: What Happy-melon was suggesting above was to transclude a page from user space onto an editnotice in MediaWiki space. And that works, and would mean that data from a user space page would be shown in MediaWiki space. And that would mean the user could edit the page in his user space and that change would be shown in the edit notice. We actually do that with some other (non-editnotice) MediaWiki messages. That is, the MediaWiki message uses a template from template space. Of course we take care to protect the template first, otherwise any user could edit the template and thus damage a system message.
 * Everyone: We now have the page Editnotice and its talkpage Wikipedia talk:Editnotice. That is a good place to continue discussing editnotices.
 * --David Göthberg (talk) 07:09, 9 September 2008 (UTC)

Automated categorization
Can anyone advice of some tool or bot to categorize the english counterpart (if any) of the articles in sv:Kategori:Personer verksamma vid Uppsala universitet to en:Category:Uppsala University faculty? Many thanks, //Urbourbo (talk) 11:23, 5 September 2008 (UTC)


 * That sounds like an good expansion task for existing interwiki-bots, with a few caveats regarding the way certain equivalent categories are used on different projects. Collectively, de:Kategorie:Mann and de:Kategorie:Frau will directly contain all biographies, but en:Category:Men and en:Category:Women will not. I'm sure other examples exist. — CharlotteWebb 13:51, 5 September 2008 (UTC)


 * Yes, there surely will be some manual analysis needed for each category pair compared; however such a tool would still greatly enhance the "internationalisation" of many categories with hundreds of articles (which I assume more people than myself would not labor to compare manually)... Which bots (and bot creators) could be candidates for expansion?... :) All the best, /Urbourbo (talk) 10:02, 6 September 2008 (UTC)
 * The same ones which update interwiki links between equivalent articles would be best suited to adding them to equivalent categories as they are editing the pages anyway and have an account on every project (or probably SUL actually). But wouldn't you know I can't think of any names off-hand... — CharlotteWebb 15:42, 9 September 2008 (UTC)

MediaWiki API: I can rollback registered editors, but not anons
This seems a bit odd to me. I'm playing around in spare moments with the API. action=edit works OK, and action=rollback works OK for registered users (diff). But, action=rollback fails for anons. Am I missing something obvious? I see there is an "anon=" flag set in the response to the token request, but I don't see what to do with this information.

I'm sending (example diff to rollback)

and getting back



Mr Stephen (talk) 22:28, 8 September 2008 (UTC)


 * 15527: already fixed in the software. —AlexSm 00:39, 9 September 2008 (UTC)


 * Of course, it's still going to be a while before the change goes live. We are running . The bug is fixed in r40628. Calvin 1998 (t·c) 01:44, 9 September 2008 (UTC)


 * It's not just me then! Thanks for the replies.  Mr Stephen (talk) 06:49, 9 September 2008 (UTC)

How to create lists in tables?
How does one use the list markup within tables? I keep ending up with only the asterix. I know there are alternatives, such as using HTML or the &bull; character, but I'd rather not use those. SharkD (talk) 23:29, 8 September 2008 (UTC)


 * One way would be something like the following, mixing wikitext and html markup:


 * Yes, it's ugly. There may be a less ugly way, but the above is what comes immediately tomind. -- Boracay Bill (talk) 23:46, 8 September 2008 (UTC)


 * Just make sure you add a line break after the |, like so:


 * (See code for how it's done). — Twas Now ( talk • contribs • e-mail ) 00:30, 9 September 2008 (UTC)


 * Yes, the key is to put the asterisk at the beginning of the line. Otherwise the MediaWiki software will not recognize the asterisk as the beginning of a list item. --Timeshifter (talk) 04:06, 9 September 2008 (UTC)
 * Thanks for the help. Unfortunately, I'm working with huge multi-column tables (see here), and breaking onto new lines or using HTML markup will make them less readable. SharkD (talk) 06:08, 9 September 2008 (UTC)
 * When all else fails, create a template. :) SharkD (talk) 06:34, 9 September 2008 (UTC)


 * If the columns where you want list items are the right-most table columns, you could do something like:


 * Wikitext for that is not really awful. -- Boracay Bill (talk) 06:43, 9 September 2008 (UTC)