Wikipedia:Village pump (technical)/Archive 34

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Improved diff gadget problem

I just tried the "improved diff view" gadget, but when I actually view a diff and click on the "view improved diff" thingy, it crashes my browser (Safari 3.1 on WinXP). Any ideas? DuncanHill (talk) 17:33, 9 April 2008 (UTC)

I have experienced this crash as well. I have also tested this in the nightly build of Safari and there it is no longer an issue, so it seems that apple fixed the problem. So you will likely have to wait with using the Gadget until Apple releases a new version of Safari :( --TheDJ (talkcontribs) 18:27, 9 April 2008 (UTC)
Well, I don't know what a nightly build is, but thanks - at least I'm not the only one affected! DuncanHill (talk) 18:56, 9 April 2008 (UTC)
I just tried it with IE7, and it sent IE's CPU usage up to 100% and IE locked up (had to close it from the Task Manager). DuncanHill (talk) 19:06, 9 April 2008 (UTC)
As far as I remember this gadget also doesn't work in Opera. And I'm pretty sure it could be fixed to work with the current Safari version. This is what happens when gadget is added without prior discussion. —AlexSm 19:26, 9 April 2008 (UTC)
Could a warning be added on the gadgets page, something like "This gadget will crash IE, Safari and Opera"? DuncanHill (talk) 19:29, 9 April 2008 (UTC)
This is what happens when gadget is added without prior discussion. WikEd was already, and had been for quite some time, a gadget at the time I added this (which is a subset of WikEd). --Random832 (contribs) 19:57, 9 April 2008 (UTC)
WikEd is 1) popular (i.e. tested by a lot of people), 2) is actively maintaned, and 3) has some real obstacles to make it compatible with other browsers. On the other hand, WikEdDiff is fairly simple (if we don't consider diff engine itself) and could be easily made cross-browser before becoming a gadget. And could have some other features (useful on diff pages) as well. —AlexSm 20:13, 9 April 2008 (UTC)
Please could the users experiencing problems with wikEdDiff please give some more information about the "crashes" on User_talk:Cacycle/wikEdDiff in order to fix the tool. For example, what do you mean by 'crash' - freezing, terminating, or just throwing a JavaScript error message. Thanks in advance, Сасусlе 18:36, 12 April 2008 (UTC)
As the gadget is no longer listed on the gadgets page in preferences I am unable to replicate the problem. But as I recall, in Safari 3.1, it appeared to freeze and then I got a "Safari has encountered a problem and needs to close" message. DuncanHill (talk) 18:44, 12 April 2008 (UTC)


There seems to be a certain tension between different people's views of what should be added as a gadget. Can we please all discuss and agree on whether gadgets that only work on a subset of browsers should be allowed or not? --Random832 (contribs) 19:59, 9 April 2008 (UTC)

I don't really mind that some gadgets only work on some browsers (of course I would prefer if they all work on mine!) but I do think it is very important that on the gadgets screen in preferences that which browsers a particular gadget will work in is very clearly stated. DuncanHill (talk) 20:19, 9 April 2008 (UTC)
Honestly, while I don't oppose the gadgets being integrated, I think I'd like to see the list on some other page (a "secondary" preferences page, of some kind). And along with them, all the "preferences" which are noted to require javascript.
I think that the page is too long for the average user/reader, and the "tools" (even by just being listed) could be more confusing than helpful. - jc37 20:31, 9 April 2008 (UTC)
Anything that doesn't work in IE should automatically be allowed, as it just provides a reason to ditch such a horrible browser. :) EVula // talk // // 21:23, 9 April 2008 (UTC)
People edited from, erm, restricted environments may not have a choice. =( Pegasus «C¦ 23:22, 9 April 2008 (UTC)
Maybe what we could try is putting some code in MediaWiki:Common.js to detect the current browser and hide or flag any gadgets that are known to be incompatible with that browser. Tra (Talk) 21:52, 9 April 2008 (UTC)

Please see Wikipedia_talk:Gadget. Сасусlе 07:09, 12 April 2008 (UTC)


Few quick questions that I was hoping someone would help me answer:

  1. What, if any, tools are Wikipedians using to flag articles/submissions for deletion?
  2. What kind of time expenditure per hour is spent filtering malicious (i.e., racist, hateful, pornographic, etc.) content?
  3. Is there a desire among the editing community to sift through the garbage more efficiently? or is that just part of the "fun" of the editing process?
  4. How frequently does malicious content get submitted (40 per 1000)?

Your help is appreciated. I'll be checking this periodically if you'd like to have more of a dialog.
Cheers. —Preceding unsigned comment added by Lootgrass (talkcontribs) 22:19, 9 April 2008 (UTC) Lootgrass (talk) 22:22, 9 April 2008 (UTC)

Submissions don't get flagged, they get reverted by clicking undo. Articles get flagged with a number of tags {{db-bio}}, {{prod}} etc. 99% of time per hour is spent filtering malicious content. By some editors. None by others. They put it there. By the way Wikipedia is not censored so pornographic only applies if it violates the laws of the state of Florida. There are a host of Bots that do sift through the garbage and auto revert many junk edits. 40 per 1000, that would be pretty good. It isn't much higher than 400 per 1000. Click "Recent changes" a few times and see for yourself what you think the ratio might be. (talk) 22:43, 9 April 2008 (UTC)
  1. One tool I found in the the index was this one: Wikipedia:WikiProject User scripts/Scripts/Easy db
  2. There are a lot of people who spend a lot of time deleting vandalism and spam, but they aren't required to fill out timecards. So there is no data on this. You could do some approximations using the data shown here, though it's limited to articles (vandalism occurs in other namespaces), to determine the number of edits per hour that are reverted, and multiply it by some factor (your decision) for the amount of time per edit. But that would be a guess.
  3. Part of the fun of editing is doing things efficiently. There are hundreds of tools (again, see the index) to help editors be more productive. See Wikipedia:Counter-Vandalism Unit and Wikipedia:WikiProject Spam, for example.
  4. Again, hard data is limited, but the problem seems to be growing (countermeasures seem to be keeping the effective rate of vandalism at a manageable level, though); 4% is probably low. -- John Broughton (♫♫) 15:56, 12 April 2008 (UTC)


Alone amongst Wikipedia pages, this project page, WP:WikiProject_Stub_sorting/Stub_types, won't load into Firefox (, for me at least. Is there any known issues here, please?

Same as above. Kathleen.wright5 23:14, 12 April 2008 (UTC)
Won't load for me in Safari 3.1 DuncanHill (talk) 00:10, 10 April 2008 (UTC)
And its just got flaky loading into IE7 - sometimes will, sometimes won't ... :-( TerriersFan (talk) 00:16, 10 April 2008 (UTC)
It works for me but it takes time to load. The page I receive is 1313851 bytes so I'm not surprised by that. Maybe the large transclusions should be spread over multiple pages. PrimeHunter (talk) 00:24, 10 April 2008 (UTC)
Fore those who cannot load it, the transcluded pages in section 2 are:
Wikipedia:WikiProject Stub sorting/Stub types/General
Wikipedia:WikiProject Stub sorting/Stub types/Culture
Wikipedia:WikiProject Stub sorting/Stub types/Education
Wikipedia:WikiProject Stub sorting/Stub types/Commerce
Wikipedia:WikiProject Stub sorting/Stub types/Government, law, and politics
Wikipedia:WikiProject Stub sorting/Stub types/Leisure
Wikipedia:WikiProject Stub sorting/Stub types/Sports
Wikipedia:WikiProject Stub sorting/Stub types/Religion, mythology, faiths, and beliefs
Wikipedia:WikiProject Stub sorting/Stub types/Geography
Wikipedia:WikiProject Stub sorting/Stub types/History
Wikipedia:WikiProject Stub sorting/Stub types/People
Wikipedia:WikiProject Stub sorting/Stub types/Science
Wikipedia:WikiProject Stub sorting/Stub types/Technology
Wikipedia:WikiProject Stub sorting/Stub types/Transport
Wikipedia:WikiProject Stub sorting/Stub types/Military and weaponry
Wikipedia:WikiProject Stub sorting/Stub types/Organizations
Wikipedia:WikiProject Stub sorting/Stub types/Miscellaneous
Wikipedia:WikiProject Stub sorting/Stub types/Footnotes
The edit link is [1]. PrimeHunter (talk) 00:28, 10 April 2008 (UTC)

It's been a bit flaky for a long time, due to its size - which is why it was split into the 20 or so subpages. They're still transcribed to a main list for ease of access for people and servers that can access them, but in practice the subpages are more useful. Grutness...wha? 00:49, 13 April 2008 (UTC)

It seems to be OK now. Kathleen.wright5 02:17, 16 April 2008(UTC)

Database dump

What is the benefit of using the database dump? I tried downloading it to find out but cancelled it because was taking a very long time (many hours). I presume that you can search and replace much faster but how do the edits get resolved with other new edits when you upload? Lightmouse (talk) 12:29, 11 April 2008 (UTC)

The database dump is for if you want to set up your own site based on the content of wikipedia pages. (or if you want to run various statistical tests on the data) --Random832 (contribs) 14:10, 11 April 2008 (UTC)
Also, by using the database dump to do statistical analysis, it takes a major strain off of the servers, which would collapse if every single venture/test ran off the live servers.Harryboyles 15:49, 11 April 2008 (UTC)
Thanks. Lightmouse (talk) 14:30, 12 April 2008 (UTC)

Account Creation

There is a discussion taking place regarding the possibility of adding a MediaWiki extension to the English Wikipedia to aid in Account Creation. Please see the discussion at Wikipedia talk:Request an account#Account Request extension. Feedback is appreciated. - Rjd0060 (talk) 22:10, 12 April 2008 (UTC)

Article names ending in "/"


When navigating up from [Wikipedia:Centralized_discussion/Image_placeholders] (using the firefox extension "uppity" (handy==lazy)), I was sent to [Wikipedia:Centralized_discussion/] which doesn't exist. No problem, I knew what to do!

Is there an advantage to allowing "/" to end an article name? I don't see one.

It doesn't seem to be a big problem though , only ~71 article name end in "/" (from []) and the few I checked were properly redirected. Just curious. Saintrain (talk) 02:05, 13 April 2008 (UTC)

On smaller wikis, it's a good spambot trap; spambots will often vandalise a page ending in / rather than the page itself. --ais523 15:48, 14 April 2008 (UTC)

Adding SimpleTable Extension

I propose adding SimpleTable extension to Wikipedia.

I think that the syntax of wikitable is not so easy, and feel pain when inputting wikitable. Persons who think similarly might be not a little. If this extension can be used in wikipedia, inputting tables will be easier. So adding this extension will help wikipedians. Please examine this positively. J8takagi (talk) 15:37, 10 April 2008 (UTC)

This could be useful, but to me there is little reason not to go further and set up a separate table namespace, with a specialized WYSIWYG table editor. Other wikis have this, and there is already a precedent (the Image namespace behaves differently than other namespaces). Then editing tables would be very, very easy. (For more on this proposal, see the "Tables" topic in the editor's index.) -- John Broughton (♫♫) 15:33, 12 April 2008 (UTC)
Thank you for the reply. I am interested in Wikipedia:Table: namespace and editor. it would be wonderful if this feature is added to wikipedia and editing tables become very easy. But adding it takes long time and there may be many problems when adding it, I think. Adding SimpleTable extension is easier and takes shorter time. Is this the major reason not to go further and adding SimpleTable Extension before setting up a separate table namespace? J8takagi (talk) 14:13, 13 April 2008 (UTC)
I don't see any harm in the SimpleTable expansion, though I doubt that most editors will become aware of it, if implemented, and of those that do, relatively few will do so. It's really more a matter of energy - if you can get this implemented without much discussion, fine. But if the community is going to spend a lot of time and effort debating it, then that time would be better spent on a solution that would benefit the vast majority of editors who are faced with editing tables, not the small minority who are adding information already in CSV form. -- John Broughton (♫♫)

Logs appearing on watchlist

Hats off to whatever developer (group of developers) pulled this one off.--VectorPotentialTalk 23:53, 10 April 2008 (UTC)

Thanks :D Voice-of-All 01:02, 11 April 2008 (UTC)
Good job, I agree. However, in enhanced expanded watchlist (imho, the only type of wathlist worth using) log entries are inconveniently grouped by log type, like 03:43 (Move log) (Page history) [User1; User2; User3]. Since I'm not really interested in users who made the moves, I'll have to uncollapse this pretty much every time. Also notice useless «page history» part. And last thing: I understand that logs entries are shown the same way as in Special:Log, but having article name first would be more convenient and more consistent with Watchlist/Recent Changes look. I hope this can be somehow improved later. —AlexSm 04:19, 11 April 2008 (UTC)
I had them grouped by page, but someone changed it. I'll bring it up. Voice-of-All 16:10, 11 April 2008 (UTC)
Yes thanks tremendously. Can I ask what triggers the display of user-right entries? I assume watchlisting the user's talk or userpage? Happymelon 17:51, 11 April 2008 (UTC)
Correct. The "target" field of userrights is just the user page, basically. —Simetrical (talk • contribs) 15:45, 13 April 2008 (UTC)

Deletion log missing

Copied from WP:AN

Has this image been deleted or not? It looks like it has, but no entry in the log! Issue complicated because the italics ('') markup appears in the filename. Carcharoth (talk) 23:37, 9 April 2008 (UTC)

Looking at the user's upload log, I'd say that it has... I have no idea why there's no entry in the deletion log. Black Falcon (Talk) 01:20, 10 April 2008 (UTC)
Would the italics markup thing have confused the log. I would have thought not, as it should represent it as %27. I'll drop a note over at VP:T. Carcharoth (talk) 20:51, 11 April 2008 (UTC)

Anyone have any explanation for the above? Carcharoth (talk) 20:52, 11 April 2008 (UTC)

Well, I just tried it myself here - deletion log appears. So the existence of '' in the title alone is not enough to cause an issue. Or it might be an entirely different cause altogether. --- RockMFR 23:26, 11 April 2008 (UTC)

Seems to be the same problem shown some topics above. --Oxymoron83 23:57, 11 April 2008 (UTC)

Undelete acts funny - The following consists of deleted revisions of [[:Image:South Park - Major Boobage - time 03'03.jpg]]. - but the deleted revisions show up and can be accessed. --Random832 (contribs) 23:43, 12 April 2008 (UTC)

Does it show up in the deleting admin's log? Do we have any idea who the deleting admin is, or when it was deleted? --Random832 (contribs) 23:48, 12 April 2008 (UTC)

Occasionally events are not logged as they should be. I have yet to hear of any smoking gun that anyone has found for this phenomenon. It's not reproducible, so it's kind of hard to track down. There's a bug open on it somewhere, but I can't seem to find it. —Simetrical (talk • contribs) 15:56, 13 April 2008 (UTC)

Problem with rollback script

This is most likely some annoyingly simple error on my part, but I'm having an issue with JavaScript. I'm trying to write a modified version of a rollback script; my modified version is located at User:Pyrospirit/rollbacksummary.js. What I'm trying to do is have the string $u be replaced with the username of the editor being reverted, whose name is found in the &from= part of the rollback link. Here's where I run into problems. To get the username of the editor, I use the following code:

var from = this.href.match(/&from=[^&]+/);
var user = from.replace(/&from=/, '');

The problem is, that line generates the following error (from the Firefox error console):

Error: from.replace is not a function
Source File:
Line: 22

Now, I am admittedly quite inexperienced with JavaScript, but replace() used on a string? I'm pretty sure it is indeed a valid function. So what am I doing wrong here, and how can I get this script to work? If you happen to notice any other bugs while looking at the script, please let me know as well. Thanks, Pyrospirit (talk · contribs) 04:55, 13 April 2008 (UTC)

User:Gracenotes/rollback.js does do a $user replacement, btw. Check out the MDC documentation for String.match. That returns an array of strings if a match exists. Arrays don't have a replace method, but the members of that array (strings) have the method. You can use capturing groups to select certain parts of a regular expression, by the way: this guide (also from MDC) explains regular expressions in JavaScript. GracenotesT § 05:07, 13 April 2008 (UTC)
Thanks, that information helps a lot. I was looking for your script to see how you did it, but for some reason I couldn't find it. I guess I'll just do it the same way you do in your script. Now that you mention it, I think I've seen that technique before, but I just forgot about it. Pyrospirit (talk · contribs) 15:32, 13 April 2008 (UTC)

Search result page text

This seems to be the right place. When you search something and no results come up it says:

unsuccessful searches are often caused by searching for common words like "have" and "from", which are not indexed.

This is incorrect, when you search 'have' or 'from', many results appear. Surely this needs to be changed? George D. Watson (Dendodge).TalkHelp 17:17, 13 April 2008 (UTC)

URLs going to wrong places

Just a second ago, I C&P'd a URL in a prot log purported to go to an old WP:AN revision into my browser window. I was instead taken to an old revision on HMS Hermes (R12). Out of curiosity, is this deliberate? -Jéské (v^_^v X of Swords) 19:36, 13 April 2008 (UTC)

When you go to an oldid, the "title" field is actually irrelevant: it takes you to the page where the revision happened, wherever that may be. For instance, the last edit to this page is, but if you change it to you end up at Agrostis capillaris. I suspect, therefore, that there was a simple typo on someone's part in the oldid value. Happymelon 20:05, 13 April 2008 (UTC)

Linking [[:]]

I was wondering if it would be possible to link [[:|this]]. I know that it's not supported because it's an operational character, but would it be possible to redirect it to Colon (punctuation)? because most people are simply going to give up. I tried but I got "Bad Title". I suppose we'd need a developer or at least an admin who knows css. Andrew Kanode (talk) 23:59, 13 April 2008 (UTC)

Do you mean ":"? ([[Colon (punctuation)|:]], see WP:PIPE) – Leo Laursen –   07:06, 14 April 2008 (UTC)
No, you can't link to ":" because it is used to make links to non standard namespaces. For example [[:Image:Nakai in Kanagawa Prefecture.png]] links to Image:Nakai in Kanagawa Prefecture.png. -- lucasbfr talk 12:00, 14 April 2008 (UTC)

It'd be possible to "auto-redirect" (the way wiktionary does with capitalized titles) from attempts to navigate to or search for such titles via javascript. I'll get something coded up later this week. That won't make the syntax [[:]] work though. --Random832 (contribs) 19:07, 14 April 2008 (UTC)

Green text & printing

I use the green text/black background gadget, and very good it is too at reducing eye-strain. However, when printing a page I get green text (when I would prefer black). Is there a way to make the printable version output black text, or do I have to switch the gadget off when printing? Thanks. DuncanHill (talk) 01:15, 14 April 2008 (UTC)

I informed its author, I'm positive who will fix it real soon. --TheDJ (talkcontribs) 11:28, 14 April 2008 (UTC)

Login page

I admit I haven't researched it extensively, but would it be easily possible to have the focus on the Username Text box when you load Special:UserLogin? I'm pretty sure that would need a modification on MediaWiki, but why isn't it done already? -- lucasbfr talk 12:03, 14 April 2008 (UTC)

I guess some site JS could be made. When HTML 5 is out, an autofocus attribute can just be given :) Voice-of-All 13:01, 14 April 2008 (UTC)
Adding the following to MediaWiki:Common.js would achieve the desired effect:
if (wgPageName == "Special:UserLogin") {
    addOnloadHook(function() {
I don't mind either way. Nihiltres{t.l} 15:37, 14 April 2008 (UTC)
In most cases we don't force focus, though perhaps it would be nice here. However we _do_ usually fix up the tab order on this sort of form; just hit 'tab' once and you'll be right in there. (Unless you're using Firefox 3.0 beta, which is broken.) --brion (talk) 00:38, 15 April 2008 (UTC)

User talk:DelvarMe

Can someone look at the script that is hidden on this talk page and advise what is going on, please? TerriersFan (talk) 19:21, 14 April 2008 (UTC)

It's mostly one big html comment. Looks like instructions for an interpreter of some sort (game? maybe Tcl or ruby?) But it doesn't do anything on wikipedia. —EncMstr 20:37, 14 April 2008 (UTC)


Hey is it just my computer but sometimes, in the search box, the line has a small horizontal bar extending to the right of it. —Preceding unsigned comment added by Rainboqer (talkcontribs) 23:13, 14 April 2008 (UTC)

Moved from T:MP by ffm 00:00, 15 April 2008 (UTC)
That's a text direction marker, indicating whether you're typing left-to-right or right-to-left text. If you switch your keyboard to Hebrew or Arabic input mode, you'll see the little dot/bar move to the left side. I believe it'll appear if there's some Arabic or Hebrew text on the page. --brion (talk) 00:31, 15 April 2008 (UTC)


I just used my account for the first time in two years and found a warning on my Monobook page. History seems to be clean, is there something else I should watch out for? --Yooden 

It's a boilerplate warning so unsuspecting editors don't get lured into trouble. You've done your due diligence and it looks fine anyway. Franamax (talk) 18:39, 10 April 2008 (UTC)
Thanks for the quick response! --Yooden 
We like using four tildes to sign nowadays, so the date of your posting shows up with your sig btw. Franamax (talk) 18:48, 10 April 2008 (UTC)
That's not new, I'm a signature rebel. Thanks for the tip anyway. --Yooden 
Oh yeah? Well maybe I'm a wrong-answer-on-VPT rebel, you never know ;) Franamax (talk) 19:01, 10 April 2008 (UTC)
I have full trust in the decent and respectable VPT users to point out such follys. --Yooden 
If signature rebels (like yourself) post comments to Talk pages archived by MiszaBot, like this one, it will probably not be able to archive the threads in which you participate. Misza depends on seeing a date on *every* comment to know whether the whole thread is stale enough to archive. EdJohnston (talk) 04:46, 11 April 2008 (UTC)
Won't it just archive slightly sooner, as if Yooden's comments weren't happening? It doesn't strike me as a huge problem. rspeer / ɹəədsɹ 05:22, 11 April 2008 (UTC)
Good point, I assumed that the bot would use the actual change dates, not the one found in the page. Let me think about this. --Yooden 
I've seen before, not sure where, advice to not-date a post to prevent archiving by MiszaBot, I'll be interested in the answer. (And I'll try 3 tildes now to see what happens :) Franamax (talk) 06:30, 11 April 2008 (UTC)
OK, try again, counting one-two-three tildes! Franamax (talk)
Thinking about it, relying on the dates within the text would be pretty difficult. Every paragraph would prevent archiving if only the date would matter. Or the bot would have to recognize signatures but avoid archiving if they are undated. Using actual change date is much more straightforward.
Anyway, I just activated archiving on my /Talk, and I have threads with only dated messages. Let's see what happens. --Yooden 
The bot does not search the edit history to find the dates for anything. It merely scans the text looking for signature and date strings. It can't remove unsigned comments, and I understand that undated comments, or those that have dates that don't use the '(UTC)' date format, will hang around forever until someone manually archives them. The closest to an online explanation of this behavior is at [2], at the bottom of the page. EdJohnston (talk) 13:21, 11 April 2008 (UTC)
As long as there's at least one timestamp in the thread, though, it will get archived. --Random832 (contribs) 14:04, 11 April 2008 (UTC)
Right on the mark, only threads entirely without date were unmoved. So only monologues are affected, I can live with that. --Yooden 
I'm a signature rebel. No, you're not - your an inconsiderate editor. You don't care that omitting the time and date makes it more difficult for other editors to follow the discussion, or to decide whether to respond or not (because old comments often aren't worth responding to). If omitting a single tilde somehow benefited you significantly, I could see the rationale - but it looks like you're just after attention, or that you get some sort of perverse delight out of irritating other editors. -- John Broughton (♫♫) 15:13, 12 April 2008 (UTC)
At least you make it easy to decide whether or not to repsond to your posts. Thanks! --Yooden 
Obviously, you have decided to respond to Mr Broughton's posts, if I am to judge from this first case.
I have to agree with him, though; this is a collaborative project and one needs to be civil and helpful to one's fellow editors. We shouldn't have been where we are now if this were not the prevailing spirit on Wikipedia. Waltham, The Duke of 05:26, 16 April 2008 (UTC)

Sortable tables

Is there a way to disable sorting on tables over 100 rows? Or not run the calculations needed to prepare a table for sorting until after I click the sort button? Whenever I load a page with a large table, my browser freezes for some time. --Random832 (contribs) 20:02, 14 April 2008 (UTC)

I don't think the table is sorted until one of the sort buttons are pressed. Until then, the table appears in the order written in wikitext. Large tables do seem to take significant time to load, but I've assumed it was due to the browser figuring out column sizes or something. Which browser are you using, and on what platform? —EncMstr 20:26, 14 April 2008 (UTC)
Looking at wikibits.js ... try to set ts_alternate_row_colors = false in your monobook.js and see if it helps. By the way, I fail to see what's the puprose of this ts_alternate() ... only if I wanted to highlight even/odd table rows with personal CSS, and that's it? —AlexSm 20:39, 14 April 2008 (UTC)
This script shouldn't be doing anything on page load. If it is and someone would like to fix it, I'd be interested to see patches. —Simetrical (talk • contribs) 15:58, 15 April 2008 (UTC)

Rollback attributed an edit to the user being rolled back, not to rollbacker

I rolled back an edit that made to January 13. However, now it appears that him/herself reversed the edit.

Here you can see the feedback that my rollback was successful; the rollback token is visible in the URL of the image in case that's helpful. Also, here it is in pdf format.

Perhaps the user did reverse his/her edit at the same time as my rollback. But even then, I would have thought that my rollback would not have shown to be successful. Any ideas? Thanks. --Art Smart (talk) 23:36, 14 April 2008 (UTC)

Well, here's how it would happen: the system would indeed roll the page back to the state before the other guy's previous edits, but because there was no change, it wouldn't actually get saved. As you say, it should probably detect this condition and explicitly say that this was what happened so you don't worry about why your rollback didn't get saved when the message was successful. If we don't get to it shortly, make sure it gets filed in bugzilla... --brion (talk) 00:35, 15 April 2008 (UTC)
Okay, that makes sense. No big deal. I just wanted to make sure there wasn't some new major issue with Rollback. Thanks for the answer. --Art Smart (talk) 12:10, 15 April 2008 (UTC)

Bot Malfunction for Sacramento, California

Several bots are incorrectly adding the category bpy:সাকরামেনটো to the Sacramento, California page. The category belongs to a small South American province rather than the capitol of California. Multiple bots make this mistake - which is probably based on a bad data file. What can be done to get the data file repaired?-DevinCook (talk) 04:38, 15 April 2008 (UTC)

You are dealing with an error in interlanguage links. See User talk:Yurik/Interwiki Bot FAQ for information on how to fix problems with these links. To remove a bad interlanguage link for good, it must be removed from all Wikipedia languages which contain it. Therefore I have gone to all the interwiki links from Sacramento, California and removed links to bpy:সাকরামেনটো. I have also fixed all the interwiki links from bpy:সাকরামেনটো, which should have gone to Sacramento, Minas Gerais. Graham87 06:21, 15 April 2008 (UTC)
Thank you. Your help is greatly appreciated. -DevinCook (talk) 08:46, 15 April 2008 (UTC)

Something I noticed elsewere

While looking for the external link search page, I noticed this comment by User:TSamuel at Wikipedia_talk:Searching/External_search_engines#Improper_Google_Define_linking_to_Wikipedia. He asks about Google searches for "define:topic"'s which result in malformed links, where the topic includes spaces. The links on the define: searches include '+'s in place of ' 's. Currently define: searches don't seem to be a prominent part of the Google interface to most users, although they can be very useful when you a) know about them and b)want a definition very fast. Is there any plan to correct this, either by a (higher-level) request to Google that they correct their behaviour where possible, or by the technical solution TSamuel proposes? John Nevard (talk) 04:43, 15 April 2008 (UTC)

Perhaps I am missing what you are saying, but I just did a quick search for define:United Nations and define:president of the United States (both of which bring up a wikipedia entry, amongst others). I then click on the wikipedia link and go to the respective page. I checked these in IE6, FF3, Opera9, and Safari3. I had no problems in any of them. Is there an example of a malformed link that you can point to? - AWeenieMan (talk) 03:23, 16 April 2008 (UTC)

Linking to old versions of redirected articles

Something seems to be wrong when linking to old versions of redirected articles. I was trying to show someone what the pre-redirect version of Introduction to particles looked like, and when I linked to the pre-redirect version, it showed up as a redirect. See this for example. It is still possible to see what was there, but you have to click "edit". For example, click (but don't save!) this. Is there a reason for this change and was it discussed anywhere? Non-redirected articles work fine, as seen from an old version of subatomic particle compared to the current version. Carcharoth (talk) 06:12, 15 April 2008 (UTC)

This seems to be a bug, otherwise these links should work. עוד מישהו Od Mishehu 12:53, 15 April 2008 (UTC)
Introduction to particles works properly when it is wikilinked. Look at your first link:
[ Introduction to particles]
This is an external link, not an internal wikilink; the "redirect=no" means just that— do not process the redirect. --— Gadget850 (Ed) talk - 12:53, 15 April 2008 (UTC)
That's not what he means; the problem is that the second link should render an old version of the page before it was redirected, but instead it displays the current redirect (even though the #REDIRECT[[]] code is not present in the old version. Happymelon 13:19, 15 April 2008 (UTC)
I don't have a bugzilla account. Could someone flag this up? I'm also sure this didn't happen before, so it must have been caused by a fairly recent change. Carcharoth (talk) 13:29, 15 April 2008 (UTC)
This has happened with the history of Coachella, as well, so it's not an isolated problem. Probably it's an issue with redirects generally, but I,ve actually seen it in action with that (no longer redirected) page. Gavia immer (talk) 15:15, 15 April 2008 (UTC)
This was presumably an error in the fix for T12931, namely r33133. It should be fairly easy to fix. —Simetrical (talk • contribs) 16:16, 15 April 2008 (UTC)
Bugzilla:13754, should be cleared up now. --brion (talk) 18:35, 15 April 2008 (UTC)
Thanks! Carcharoth (talk) 23:51, 15 April 2008 (UTC)

Link to Arthroplasty

Hi, I was working on the Hip dysplasia (human) page and can't get one of my links to work. The frustrating thing is that if I copy the word and paste it into the search window it works. (To rule out typos) But in the text it ends me on the Edit page instead of the article. At another place in the article the link worked fine. What am I missing?? This is the text copied out:

Hip dysplasia is often cited as causing osteoarthritis of the hip at a comparatively young age. Dislocated load bearing surfaces lead to increased and unusual wear. Although there are studies that contradict these findings. (see [1] [2]) Subsequent treatment with total hip athroplasty (Hip replacement) is complicated by a need for revision surgery (replacing the artificial joint) due to skeletal changes as the body matures, loosening/wear or bone resorption.Osteotomies are either used in conjunction with arthroplasty or by themselves to correct misalignment. —Preceding unsigned comment added by Lisa4edit (talkcontribs) 11:18, 15 April 2008 (UTC)

Did this help? Carcharoth (talk) 11:43, 15 April 2008 (UTC)

Thanks I had tried to correct that typo but somehow my edit didn't take. The last time the word showed up correctly but still redirected wrong. I hope no one has blocked me from editing that page because of some stupid bot rule. I don't have all the references in yet and the paragraphs on "history" and "development" are still missing. There may also still be more articles I can link to. —Preceding unsigned comment added by Lisa4edit (talkcontribs) 11:59, 15 April 2008 (UTC)

1-click infobox conversion

At WP:SHIPS and WP:MILHIST we've developed and been using a fairly painless way to do mass infobox conversions. In a nutshell, you develop a template T which takes the input for template A and converts it to a call to template B. Then you go to pages that have template A on them and replace A with subst:T. You click preview, then click save. No fuss, no muss.

An example of one of these "doggy-door" converters is currently available at User:Haus/9.

Since other projects are presumably doing conversions by hand, I thought it might be nice to share this technique. So my questions are: 1) has this method been used/documented elsewhere, and, if not, 2) what would be an appropriate place to post an essay on it.

Cheers. HausTalk 13:06, 15 April 2008 (UTC)

We used something like that with {{Infobox Television episode}}. We took the old (show specific) box, filled it with a usage of the new box, but used the old param values as the new values. Then we subst all transclusions of the old box. Presto, you have the new infobox in place. Then delete or redirect the old one. --TheDJ (talkcontribs) 13:25, 15 April 2008 (UTC)

Adding a music sample

i was wondering how i can add a 30 second sample of a song to the arcicle on Page 44? I've seen this done on other articles and still havn't figured it out. LukeTheSpook (talk) 23:46, 15 April 2008 (UTC)

See Wikipedia:Music samples and Wikipedia:Creation and usage of media files#Audio. Graham87 04:34, 16 April 2008 (UTC)

Where to follow software/technical upgrades to Wikipedia?

I just logged in for the first time in a few weeks and I was surprised to see that the search results have a different format than I'm used to. The found terms used to be red and now they are bold black. I've been trying to figure out what update caused the change but haven't been able to find any info. What's the place to keep tracks of mediawiki and other software changes to wikipedia? Jason Quinn (talk) 01:00, 16 April 2008 (UTC)

To get an overview of bug fixes and new features that affect Wikipedia, check out the technology report in the Wikipedia Signpost. To follow MediaWiki development more closely, use the MediaWiki SVN repository. Important changes are often announced on this page as well. Graham87 04:26, 16 April 2008 (UTC)
And for more non-mediawiki updates, see the Server admin log. Mr.Z-man 04:27, 16 April 2008 (UTC)

Post swallowed by Wiki software ?

I replied to a talk page post here, it stills shows in the diff when the next post was made here (but there is something wrong with that diff, it repeats some other text), but the text I posted is gone from the page. I've seen strange edit conflicts before where text was deleted by subsequent posts because of an edit conflict, but in those cases, the text showed as deleted in the diff. In this case, the text shows as if it's there in the diff, but it's gone. SandyGeorgia (Talk) 06:14, 16 April 2008 (UTC)

You're response is there. The post which added duplicate text added the section as it was before you responded. Is it possible that's what you were looking at, and not the other section of the same name? Or am I missing the problem? Gimmetrow 06:20, 16 April 2008 (UTC)
I'm seeing it now (I think ?) ... the poster after me duplicated three sections above his, but chopped my last post from the second, so I need to revert to before his post and then restore his post, since he made several errors there. Is that what you see? SandyGeorgia (Talk) 06:25, 16 April 2008 (UTC)
Yes, that's it. What a strange thing he did. I reverted back to my last version and then re-added his post. Thanks, Gimmetrow. SandyGeorgia (Talk) 06:29, 16 April 2008 (UTC)

version of {{PAGENAME}} that shows redirects

It would be useful if there was a version of {{PAGENAME}} which gave the title of the page as based on the URL before any redirection was applied, i.e. in the case that you accessed the page via a redirect then it would give the name of the redirect itself rather than the page that it pointed to.

The reason for this is that templates could then be written which test for it, and produce different results depending on what title had been used to access the page. I am thinking in particular of use with some of the disambig link templates. To give just one example, template {{redirect}}, as used in article United Kingdom, will tell you:

"UK" redirects here. For other uses, see UK (disambiguation).

but this is essentially only relevant to you if you have used the title UK to access the page. If you have used United Kingdom in full, then you are very unlikely to be looking for other meanings for "UK". I am not very familiar with template syntax myself, but I presume that if the software were to provide a variable called {{PAGENAMEORIG}} or whatever, then it would be fairly trivial to modify templates such as {{redirect}} to make it so that these sort of disambig links only appear if you have actually gone via the redirect in the first place.

Many thanks, — Alan 16:20, 10 April 2008 (UTC)

This doesn't seem likely to be implemented. It would require fragmenting the parser cache unnecessarily. You'd have to cache the parsed text separately for every separate redirect, if someone used this, and inevitably it would get used everywhere. —Simetrical (talk • contribs) 14:20, 11 April 2008 (UTC)
Correct me if I am wrong, but the squids cache pages based on URLs, and mediawiki redirects are not HTML redirects, but content transclusions. So Oxford University (that redirects to University of Oxford), will have a URL of, while the destination page has a URL of There shouldn't be a squid-level problem with having separate versions of the page based on a magic word like {{NOREDIRECTPAGENAME}} (although... memcaches and other internal caches might be afllictified?). --Splarka (rant) 07:15, 12 April 2008 (UTC)
I said parser cache, not Squid cache. The Squid cache would, as you say, not be affected. I don't know how bad the parser cache fragmentation would be, but I doubt it would be considered warranted for a feature like this. —Simetrical (talk • contribs) 15:42, 13 April 2008 (UTC)
Okay, pity there had to be some issue like this. Many thanks for the feedback. — Alan 07:31, 16 April 2008 (UTC)

Section Shortcuts

Don't seem to be working properly. The Help desk has gotten a few questions over this recently. (Crossposting below).

Section shortcuts such as WP:IDHT, WP:DASH, WP:HEAD etc. don't seem to be working at the moment. They only take the user to the top of the relevant page, rather than to the appropriate section. Any ideas? Jayen466 12:13, 15 April 2008 (UTC)

Yikes, this is the third question on this topic today. See above. --Richardrj talk email 12:16, 15 April 2008 (UTC)
Tx, indeed. (Probably not the last either, unless someone manages to fix it. ;-) ) Jayen466 12:29, 15 April 2008 (UTC)

Any fixes / help? Best, --Bfigura (talk) 13:39, 15 April 2008 (UTC)

It appears to affect all redirects to sections, not just WP. For example, Infinitude of primes contains #redirect [[prime number#There are infinitely many prime numbers]], so it should go to Prime number#There are infinitely many prime numbers, but #There_are_infinitely_many_prime_numbers is not in the url generated by the redirect. PrimeHunter (talk) 15:15, 15 April 2008 (UTC)
This seems to have been broken for at least a year. A great nuisance as it is often better to redirect to a specific section of an article rather than to create a separate stub. —Ian Spackman (talk) 14:37, 16 April 2008 (UTC)

I believe these redirect issues were caused by revision 33133. It seems to be the most likely culprit. --- RockMFR 15:18, 15 April 2008 (UTC)

My first bug - I hope it's not a duplicate :S Happymelon 15:56, 15 April 2008 (UTC)
I'm guessing we all have to wait until tomorrow for it to be fixed? (talk) 17:54, 15 April 2008 (UTC)
Nope, was already fixed. --brion (talk) 18:35, 15 April 2008 (UTC)
And thanks for this as well! Impressive service and speedy fixes. Carcharoth (talk) 23:52, 15 April 2008 (UTC)

What Links Here

There was a recent vandal on the Heath Ledger page [3] and now if you go to the Special:Whatlinkshere page for a link that is on HL, the HL page appears at the bottom, now? What the..? Did something go wrong in the move revert or does it take awhile to change?  Chantessy  23:44, 15 April 2008 (UTC)

By the way, the user who did it looks like the same one behind User:Grawp, who has a history of vandal sprees on current event topics.  Chantessy  23:50, 15 April 2008 (UTC)
I don't understand what you refer to. Is it something in Special:WhatLinksHere/Heath Ledger? Please use the precise names. PrimeHunter (talk) 01:17, 16 April 2008 (UTC)

The results on the whatlinkshere page is ordered by the article's id number. If an article is deleted (as this one was while cleaning up the move), it is assigned a new id number, so it gets sent to the bottom of the list. --- RockMFR 03:04, 16 April 2008 (UTC)

Deleted articles have not been assigned new id's for something like a year, to the best of my knowledge. tables.sql says ar_page_id has existed since 1.11. —Simetrical (talk • contribs) 14:47, 16 April 2008 (UTC)

centrifugal fans

Dear Sirs

I am a design engineer and I want to know about centrfugal fans and its accesory parts like damper, couplings, vibration isolators etc.

plesae help me regrding these.

rgds Rakesh —Preceding unsigned comment added by Rcrakesh (talkcontribs) 07:16, 16 April 2008 (UTC)

I've answered on his talk page. —EncMstr 07:24, 16 April 2008 (UTC)

Blocks not being recorded in the block log...

I recently blocked this IP for 24 hours, and nothing showed up in there block log. I definitely blocked them and did not pull the stupid mistake of only leaving them a blocked template and not blocking because the the bot at AIV picked it up. I was looking through a few other recently blocked IP’s block logs and found that this IP also was blocked for (this time for one month) but again nothing was recorded in their block log. Tiptoety talk 22:22, 16 April 2008 (UTC)

This is an ancient bug, so ancient that I can't remember the reason for it. You can always find existing blocks by clicking on view existing blocks on the blocking page, at Special:IPBlockList, eg[4]. -- zzuuzz (talk) 00:46, 17 April 2008 (UTC)