Wikipedia:Village pump (technical)/Archive AI

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


Calculations with month names

I'm wanting to use an expression to dynamically give me the date of 3 months ago. I've tried using {{#expr:{{CURRENTMONTH}}-3}} {{CURRENTDAY}} but that yields -2 29. I know this is because CURRENTMONTH returns the month in numeric form (1 for January) and I need to use CURRENTMONTHNAME but that just results in an error. Is there a way to handle this or am I out of luck?↔NMajdantalk 17:59, 29 January 2007 (UTC)

Just put {{MONTHNAME|1}} in CURRENTMONTHNAME's place and change the number to one higher every month until next January, then make it 1 again. RyGuy Sign Here! My Journal 18:10, 29 January 2007 (UTC)
April 22 ({{#time:F d|-3 months}}). --ais523 18:18, 29 January 2007 (UTC)
That works, too. Thank you.↔NMajdantalk
Ok, thank you. I can actually take your suggestion and combine it with my previous attempt: {{MONTHNAME|{{#expr:{{CURRENTMONTH}}-3}} }} {{CURRENTDAY}} which gives me: April 22.↔NMajdantalk 18:19, 29 January 2007 (UTC)
SFriendly.gif What can I say? I like to help people. You're most welcome, RyGuy Sign Here! My Journal 18:29, 29 January 2007 (UTC)

Show preview and <ref>

Is there a way of having the show preview page have a <references/> section built into it so that we can see how the notes would display? Apologies if this is an old one. Steve block Talk 14:32, 29 January 2007 (UTC)

References do show up if you're editing the whole article and not just a section. So you may want to do that if you know you're going to want to preview the references.↔NMajdantalk 18:01, 29 January 2007 (UTC)
It'd be helpful in section editing because some of the larger articles really slow my machine down. Still, if there isn't a hack, so be it. Steve block Talk 18:19, 29 January 2007 (UTC)
For a section preview, add the <references/> tag somewhere after the last <ref>...</ref> piece. (Don't forget to remove it before pressing Save page.) —EncMstr 18:20, 29 January 2007 (UTC)
Ah, that's a good idea. Thanks, much appreciated. Wish I'd thought of that. :) Steve block Talk 00:46, 30 January 2007 (UTC)

Clearing seach box

I have searched Wikipedia on the home page that jumps up when you enter many times and the list of searches is now very long. How can I clear the list? —The preceding unsigned comment was added by (talk) 19:21, 28 January 2007 (UTC).

You can hit the DELETE key while scrolling through the list to delete a single item. Also, assuming you're using Internet Explorer, you can go to Tools --> Internet Options --> Content --> AutoComplete to clear your AutoComplete history completely. Just be aware that it will delete your AutoComplete history for everything, not just the Wikipedia search box. SubSeven 21:31, 28 January 2007 (UTC)
I am using IE but do not want to clear my AutoComplete history and deleting one by onedoes not work. Any other suggestions? —The preceding unsigned comment was added by (talkcontribs).
Make sure you have not selected the item when you try to delete it. Well, here's a way that it should definitely work for you. Click on the search box, hit the down arrow, and the AutoComplete entries should pop up. Move through the list with the up and down arrows (dont click anything with the mouse) and you should be able to hit DELETE to get rid of them at your leisure. SubSeven 04:32, 30 January 2007 (UTC)

Great! Got it. I was selecting the entry I wanted to delete.

Extremely slow

I find that any link to has an extremely slow response time, I think minutes to tens of minutes. (I haven't measured it lately; it takes too long to bother waiting.) This didn't matter so much when it was just Kate's tool, but now the geographic coordinate links have moved to t.w.d as well.

It seems to me that others must be getting reasonable response times, or no one would have accepted the migration. Can anyone explain why t.w.d is responding so slowly for me? I'm using Mozilla on Debian stable. I recall once long ago getting a reasonable response time from Kate's tool on t.w.d, when accessing it from Windows. --Trovatore 20:05, 27 January 2007 (UTC)

It seems fine to me now. In fact, it's really really fast right now. My article count (3600+) took less than 5 seconds, both on SeaMonkey on Windows XP, and Firefox on Linux Fedora. —EncMstr 06:59, 28 January 2007 (UTC)
OK, I clicked on that link, and it took more than three minutes. Better than the fifteen I've seen in the past, but still hardly acceptable. Could you try it from a Linux machine and see if you still get a fast response time? Another possibility--I'm using fairly tight packet filtering (which I'm not willing to turn off; that would be foolhardy--if that's the reason, it's's fault, not mine). --07:10, 28 January 2007 (UTC)
Sorry, I guess you did try it from a Linux machine, hadn't seen that. Still, there's clearly a problem. I don't get slow responses from websites in general, so I have to think it's the server's problem and not mine. Can anyone else confirm having seen this? Maybe we can try to figure out the cause and let the wikimedia people know. --Trovatore 07:20, 28 January 2007 (UTC)
On both machines it remains quite peppy: 4 to 9 seconds, with 5 tests on each machine. The equivalent query for your account (7800+ entries) varies from 9 to 13 seconds. —EncMstr 07:50, 28 January 2007 (UTC)
Are you using iptables on the Fedora machine? (And if not, why not? but that's beside the point here.) --Trovatore 08:34, 28 January 2007 (UTC)
It's using the installation defaults, which is 15 rules like ACCEPT all -- anywhere anywhere. —EncMstr 14:27, 28 January 2007 (UTC)
If you are using too tight packet filtering rules, yes, it could be your fault. There's one kind of ICMP packet which should never be blocked: the destination unreachable - fragmentation needed and DF set. If that kind of ICMP packet is blocked, PMTUD stops working, with the usual symptom being either extremly slow page load times or pages starting to load and then stopping, but only when your packets have to cross a link with a smaller MTU. On the other hand, it's also possible that it's the toolserver that has the bogus packet filtering rules, and you are the only one using PPPoE without MSS clamping (are you?). If you find out that's the case, please tell us so we can ask the server admin to fix it. --cesarb 13:20, 28 January 2007 (UTC)
I don't know what MSS clamping is, so I don't know if I'm using it. I open the pppoe channel with "pon" from the pppd distribution for Debian. --Trovatore 00:24, 29 January 2007 (UTC)
If you don't know and are using the Debian pppoe, then you are using it (it's the default on Debian). Which means that the problem probably isn't PMTUD. Did you try using network diagnostic tools to try to find the problem (ping, traceroute, tracepath, tcptraceroute, tcpdump, etc)? --cesarb 00:56, 29 January 2007 (UTC)
Nothing from ping or traceroute looks unusual. The other tools you mention I don't have installed, or know anything about. Here's my iptables; maybe you can tell me a rule for the ICMP packets that I can try.
:OUTPUT ACCEPT [644:61225]
-A INPUT -i lo -j ACCEPT
-A INPUT -p tcp -m multiport --dports 135,137,139,445 -j DROP
Thanks, --Trovatore 04:42, 29 January 2007 (UTC)
Update: Based on your comments and the PMTUD article, I tried loosening my ICMP rules slightly, to the following:
-A INPUT -i lo -j ACCEPT
-A INPUT -p icmp -m icmp --icmp-type 3 -j ACCEPT
-A INPUT -p tcp -m multiport --dports 135,137,139,445 -j DROP
-A FORWARD -p icmp -m icmp --icmp-type 3 -j ACCEPT
It made no difference. Thanks, --Trovatore 05:11, 29 January 2007 (UTC)
Tool1 / wannabe_kate has to load all of the Special:Contributions for a given user, and so is going to be very slow for users with large edit counts. Now that the database is replicating, it's probably a good idea to switch back to Tool3, which doesn't depend on the output of Special:Contributions. (as long as 13 days of replication lag is acceptable) --Interiot 14:34, 29 January 2007 (UTC)
Thanks, Interiot, but I'm afraid your tool is not the issue. It takes me three minutes just to load the URL , not trying to run any tool at all. --Trovatore 18:54, 29 January 2007 (UTC)
That would seem to eliminate many of the possible answers to the question. His Tool3 page loads here in less than a second. Maybe your systems don't like German-accented URIs?  :-) How long does a single element from the page take to load? How about the powered by tag? Also, what does it say at the bottom of the Tool3 page? Mine says Edit count is powered by Linux and MySQL 5.0.27-max-log. Runtime: 0.24 secondsEncMstr 19:05, 29 January 2007 (UTC)
It doesn't seem to be a per-element thing. It says "connecting to" in the status bar for about three minutes, then the whole page loads at once, more or less independently of which page I try to load. Then if I click on any links there, it's another three minutes.
Here's what it alleges at the bottom of the Tool3 page:
Edit count is powered by Linux and MySQL 5.0.27-max-log. Runtime: 0.03 seconds.
Of course it took three minutes or so for me to see it. --Trovatore 06:06, 30 January 2007 (UTC)

Strange edit conflict

Image:Editdiffs.png - the software apparently converted the signature tildes into text already, so I got an "edit conflict" when I tried to save the page. Strange, never happened before, even when using show preview a few times. Is this just a random glitch? Hbdragon88 04:39, 1 February 2007 (UTC)

Happens to me if I double-hit "save page" or sometimes when I re-POST after a Squid error. My edit was already saved and the ~~~~ already converted, so a second save with the same text with the un-converted sig results in a conflict. —bbatsell ¿? 04:41, 1 February 2007 (UTC)

Special:Upload should link to the deletion log

I want to suggest that the Special:Upload page have a link to the deletion log. That way, when clicking on the red link of an image that used to be there, it's easier to see what happened to it. Where, if not here, can I suggest this? — coelacan talk — 02:23, 1 February 2007 (UTC)

I agree. I specifically remember looking for a deleted image the other day and couldn't find out what happened to it. However, this might be better off at WP:VPR. PTO 02:25, 1 February 2007 (UTC)
Aha! Thanks! I overlooked that. — coelacan talk — 02:32, 1 February 2007 (UTC)

Cannot log into my account

Hi. I am the user Hoot. As of today, I can no longer log into my account, because my former password, which I have used for the past 4 or 5 years, no longer works for some reason. I have tried to get the server to e-mail me a new password, but that didn't work either, as there no longer seems to be an e-mail address associated with my account.

Perhaps someone has kidnapped my account, but this seems strange to me, as I have never been targeted by phishing attempts (let alone submitted to them).

In any case, I'd like to have my account back. Any help would be greatly appreciated.


I think this might be a database problem, requiring developer attention. There is an e-mail address associated with the account Hoot (talk · contribs) because I can see an 'e-mail this user' link but when clicking it, I am told that there is no e-mail enabled. I think this might have happened before with another user a few weeks ago. Tra (Talk) 23:05, 30 January 2007 (UTC)
Your account last made an edit on the tenth of this month. See your contributions here to see if there's a sign of someone else using your account. Indeed, your contribution history is "spacey"—that is, a few days use, and then many months without a single edit. Not insult intended, but have you checked the correct setting of Caps Lock and Num Lock? Try typing your password into a clear text field (like notepad) to insure it shows up correctly. Is it possible your keyboard is in a strange state or defective, like a broken key—or a sticking shift key? —EncMstr 23:07, 30 January 2007 (UTC)

Request for assistance

Hello, if anyone who understands esoteric template syntax could help me out at User:Angr/Template request, I'd be very grateful! Thanks. —Angr 21:32, 30 January 2007 (UTC)

I've made a template at wikt:Template:Irish noun m1 vowel. Tra (Talk) 00:09, 31 January 2007 (UTC)

What links here

Does anyone know of a script or tool that will crawl a category and create a list based on the "What links here" count. I'm looking for something to guide priority assessment in a Wikiproject (not that it would be the basis for priority - just a helpfull tool). Morphh (talk) 20:06, 30 January 2007 (UTC)

You can always ask for one at WP:BOTREQ.↔NMajdantalk 20:15, 30 January 2007 (UTC)

Marking other users as "trusted" to help with watchlists

I keep coming back to the thought that whenever I see a user I trust make an edit to something that pops up in my watchlist, I think "I don't really need to know this, as I trust their judgment". It would be nice to filter out watchlist edits by users I trust. Of course, this would require a new mechanism to mark other users as "trusted". Has anyone else had this same thought, and would this be difficult to implement? Stevie is the man! TalkWork 19:56, 30 January 2007 (UTC)

Isn't this similar to patrolled editing? The problem is, it's not turned on here on English Wikipedia, probably due to lack of consensus. Flyingtoaster1337 02:13, 31 January 2007 (UTC)
I think patrolled editing is more about verified article versions rather than trusted users, but I suppose both approaches could be used to reduce what pops up in a watchlist. Stevie is the man! TalkWork 02:26, 31 January 2007 (UTC)
Patrolled edits are, at least for the software, "Someone saw this, he said it is good, so let's mark it as such so other people don't have to check it again." Titoxd(?!?) 03:47, 31 January 2007 (UTC)
That's what I thought. If we combined both of the concepts we're discussing here, it would be possible to exclude from our watchlists either edits or edit patrols by a trusted user, thus saving a lot of time and making a lot of us more productive. Stevie is the man! TalkWork 04:02, 31 January 2007 (UTC)
Due to recent changes to the patrolling software, now edits by any user capable of patrolling an edit are excluded in situations where patrolled edits are excluded, so turning on patrolling would just what Stevietheman just described. It would be necessary to think carefully about what level of users we allowed to patrol things (possibly even inventing a new 'patroller' access level); the problem here is that enwiki has changes so often that we'd need a lot of RC patrollers to mark edits as patrolled faster than they were made. --ais523 11:18, 31 January 2007 (UTC)
Also, patrolling an edit is a log action; see Special:Log/patrol. Titoxd(?!?) 00:12, 1 February 2007 (UTC)

Site leeching live content (

The site at is leeching live content direct from - see here, for example. Is this the right place to report it for blocking? CiaranG 15:48, 30 January 2007 (UTC)

I suggest sending a mail to the enwiki mailing list. Or log into #wikimedia-tech at and leave a message. Check this also. -- ReyBrujo 13:18, 31 January 2007 (UTC)
I reported it to wikitech-l in the end, which is apparently the right place. Still live as of now though. Cheers, CiaranG 13:36, 31 January 2007 (UTC)

user not getting orange bar for new messages

User:TomStar81 says,

This is the second day in a row that I have logged on to wikipedia, had new messages waiting on my talk page, and not gotten that orange/yellowisk "you have new messages" bar. Why is that?[1]

Thought I'd bring that up here. I'll ask the user to watch this page and section for replies. — coel</span>acan talk — 05:26, 29 January 2007 (UTC)

Which page does he visit first when he logs in? If he has a "remember me" cookie set, and visits his own talk page first, before doing anything else, he won't get the bar. Titoxd(?!?) 03:45, 31 January 2007 (UTC)

I do not have the "remeber me" cookie set since I do not want people to gain access to my computer and use my Wikipedia account for thier own purposes. Usually the bar appears with the screen announcing a sucsessful login, but for the last three days I have logged in and it has not been there. Since the seconf page I check after logging in is my watchlist, and my userpage and user talkpage are both on it, I can tell if someone has left me a message by doing that, which is how the problem came to my attention. Incidentally, when I was on earlier from my home computer the isp adress I was using had a new message, but after reading the talk page the bar at the top refused to go away. I have no idea if they are related, but it does bug me. I am at a loss to explain this, so any suggestions would be nice. TomStar81 (Talk) 07:58, 31 January 2007 (UTC)

Is Inline Citation Style preferred over Footnote3 Citation Style

I first point you to this brief conversation. I cite sources a lot but I've always been using Footnote3 style. Before I change the way I operate, and without me preaching why I think Footnote3 is better, is it indeed true that Wikipedia recommends Inline style over Footnote3 style? Thanks. Fife Club 14:23, 30 January 2007 (UTC)

"Inline citations" is a fairly generic term, at least as it is used in WP:WIAFA. The Footnote3 system (ref/note) and Cite.php (ref/references) are two technologies to implement "inline citations". Per an ArbCom case (see WP:FN), It's inappropriate to change from one technology to the other without consensus among the editors. Gimmetrow 15:20, 30 January 2007 (UTC)

Template needed for image that should be JPEG

After watching Image:HRE 11Jh.PNG load, I ended up concluding that a progressive JPEG might compress the image better than PNG for that image. We have {{ShouldBeJPEG}}, but that is intended for GIF images. This image doesn't suffer from the 256 color problem metioned in {{ShouldBeJPEG}}. Do we have a template for Image:HRE 11Jh.PNG? Will (Talk - contribs) 05:21, 2 February 2007 (UTC)

I don't agree. True if you save it as a JPEG with 0 compression the quality loss is almost not noticable on the full sized image and you save a megabyte or two on the image size, but it will also make the image harder to edit. Any "non expert" who edit the image and save the result with a default JPEG compression setting would interoduce compression artifacts and thus degrade the quality of subsequent revisions. The thumbnail quality will also be noticable worse for a JPEG version of the image, idealy maps should be in SVG format, unless the "terain texture" is important (hard to do in SVG without embedding a raster image in it wich kind of defeat much of the benefit of using SVG). --Sherool (talk) 09:58, 2 February 2007 (UTC)

White Letters On A Table

Hi, I tried what it said on the Table Help page, but I was unable to get white letters in a table. I was wondering if someone could assist, please let me know on my talk page. Just H 04:42, 2 February 2007 (UTC)

Signature failing

Recently, my signature just went crazy. Previously I had a Script MT Bold design and switch to design like:

William Pembroke

but when I use my signature (~~~~) it turns out like this:

<font face="Edwardian Script ITC" size="5">[[User:William Pembroke|William [[User talk:William Pembroke|Pembroke]]</font>]]

Why is it changing it from my original

 <font face="Edwardian Script ITC" size="4">[[User:William Pembroke|William]] [[User talk:William Pembroke|Pembroke]]</font>  


<font face="Edwardian Script ITC" size="5">[[User:William Pembroke|William [[User talk:William Pembroke|Pembroke]]</font>]]

Thanks to whoever can fix this. William Pembroke 20:45, 1 February 2007 (UTC)

Scratch that, problem has been fixed. William Pembroke 21:11, 1 February 2007 (UTC)
Per WP:SIG it is strongly preferred that you not have signatures that are of a larger size that the surrounding font. It can be very distracting and can take up a lot of excess space on a page.↔NMajdantalk 22:47, 1 February 2007 (UTC)

Media on pages loads incorrectly

Since last night, I've been having problems with various articles. Images generally don't load, and I just get the alt text. TeX images don't load either, and display just as ASCII TeX markup. What's going on? --Smack (talk) 19:51, 1 February 2007 (UTC)

SVG thumbs

Why does MediaWiki use PNG to thumb a SVG? It doesn't make any sense to me. SVG is by definition scalable. It can serve as its own thumb. It is my experience that some SVGs (national flags) load faster than their thumb. The only time using PNG for a SVG thumb might make sense is if the thumb is really tiny and the SVG is complex. Will (Talk - contribs) 07:22, 1 February 2007 (UTC)

Direct viewing of SVG files is not supported in most people's browser configurations. Mike Dillon 07:51, 1 February 2007 (UTC)

Then what happens when I view the actual SVG? I use Firefox and it appears to work fine. Will (Talk - contribs) 08:00, 1 February 2007 (UTC)

Most people don't use Firefox. Also, I don't think that was even enabled in default Firefox builds on all platforms until Firefox 2.0. I guess you're one of the lucky ones, but most people use Internet Explorer which doesn't support SVG viewing without a plugin. Mike Dillon 08:03, 1 February 2007 (UTC)
Should also be noted that firefox's implimentation of SVG is far from flawless.Geni 13:46, 1 February 2007 (UTC)
SVG's can also contain external referenes, scripts and all sorts of other fun stuff that creative vandals could make use of. "Direct" SVG support is on the "wish list" though, see meta:SVG image support for some details. --Sherool (talk) 08:52, 1 February 2007 (UTC)

What were the scripts intended for? Animation? Will (Talk - contribs) 04:43, 2 February 2007 (UTC)

Animation, games, menus and all sorts of Ajaxy stuff you name it. Anyting you can do with JavaScript and XML can be done in a SVG, here is a tetris game in SVG format for example : --Sherool (talk) 10:20, 2 February 2007 (UTC)

Use external editor by default

Anyone know where I can get an external editor to use the "Use external editor by default" thing? I'd prefer one that's not a firefox plugin and not just like Javascript only.

When I try it without one, it just downloads index.php and in the file is:

Type=Edit text
Special namespace=Special


SakotGrimshine 06:00, 1 February 2007 (UTC)

Check here. -- ReyBrujo 21:04, 1 February 2007 (UTC)

Also how do I get external diff by default to work? SakotGrimshine 06:01, 1 February 2007 (UTC)

I checked meta:Help:External editors ages ago. It only has one editor that doesn't require linux, and that's "Mime Type Editor". It gives a link to a site that no longer has it. A google search for "Mime Type Editor" finds only pages of people asking where they can get it and others responding they have no idea, or else dead links. SakotGrimshine 05:05, 2 February 2007 (UTC)


Is it possible to define the width of a column? If so, how is that done? Specifically I want to set the width of the second column to 410 px and let the first column float with the screen resolution. Thanks for your help.
Frank van Mierlo 21:08, 31 January 2007 (UTC)

If you mean a column of a table, see Help:Table#Width.2C_height. If you mean a <div class="references-2column"> tag, then I don't know. —EncMstr 04:48, 1 February 2007 (UTC)
I mean the latter and like you I don't know how to do it. :) Frank van Mierlo 16:32, 1 February 2007 (UTC)
The following template allows you to define the width: Template:Reflist.↔NMajdantalk 16:45, 1 February 2007 (UTC)
Sorry for being so dense. I mean the {{col-begin}} {{col-2}} TEXT FOR FIRST COLUMN {{col-2}} TEXT FOR SECOND COLUMN sequense. I would like to set the width of the second column and let the first column float with the screen resolution. Is this possible? Frank van Mierlo 22:37, 1 February 2007 (UTC)
Oh, I thought you said above it was in reference to a 2 column layout for references. If you're wanting a dynamic first column and a static second column, then you are going to have to resort to CSS.↔NMajdantalk 22:49, 1 February 2007 (UTC)

References and footnotes

Some editors are collaborating on the "Hawker Hurricane" page, and we would like some information on how to list references and footnotes. A lot of the references are from one book and form a long list, which probably does not look right. What style of footnote and reference listing should be used here? We are aiming to make the article a good article, which should be referenced correctly. Snowman 19:20, 31 January 2007 (UTC)

That shouldn't be a problem. Just use Cite.php, and make 20 calls to the same reference, if necessary. Titoxd(?!?) 00:11, 1 February 2007 (UTC)

Editing problem at Template:DMoz/doc

I was going to add the following to the Usage topic at Template:Dmoz/doc:
- - - - - - - - - -

Include the trailing slash on the category name. Most browsers will still function if the / is excluded, but a nonstandard URL will result.

Although the second attribute is optional, it may be desired when the DMoz category has two or more words. The dmoz template defaults to using Wiki capitalization defaults (only the first word is capitalized), whereas at DMoz, multi-word category names follow more traditional capitalization rules (all words capitalized except leading words and connectives). So if "House of Wax" is the text you want users to see, include it as the second attribute. Otherwise:

{{dmoz|Arts/Movies/Titles/H/House_of_Wax_-_2005/}} results in

Village pump (technical)/Archive AI at DMOZ

- - - - - - - - - -
A couple of weirdnesses result on preview.

1) The nowiki sectioin is not placed in a blockquote box as is the pre-existing nowiki text just above what I was going to add.

2) The text displayed by the dmoz template is Dmoz/doc instead of the dmoz category name. Similarly I notice as I preview this very session at the Village pump that "Village pump (technical)" is displayed. This doesn't happen when I use the dmoz template on a user article page. (I decided to update the dmoz template documentation based on what happened when using dmoz in the Workers' compensation article. There were no weirdnesses there; I just learned about the trailing slash and capitalization issues.) – 2*6 04:10, 31 January 2007 (UTC)

As noted on template:dmoz that template has two parameters. If you don't specify the second param, the template takes the name of the page where the call is placed instead (this is expressed in the template code by using {{PAGENAME}} = "Village pump (technical)/Archive AI", see m:Help:Variable). --Ligulem 19:30, 31 January 2007 (UTC)
Ah. I interpreted "name of the page" to be the name of the DMoz page, not the name of the Wikipedia article. Main weirdness solved. It still doesn't explain why the pre-existing text uses nowiki and gets a block quote but when I tried to use it, no blockquote occurred. Oh well, this is enough information to update the DMoz/doc page using workarounds. – 2*6 23:12, 31 January 2007 (UTC)
I'm not sure I understand what exactly your remaining problem is but note that if a line entered in the edit window starts with a space, the remaining text is formatted as if using <pre>...</pre>:
lorem ipsum dolor
This feature is described at Help:Wiki_markup_examples#Just_show_what_I_typed (third row). --Ligulem 23:28, 31 January 2007 (UTC)
Thanks; I had missed that leading space trick. I've updated Template:Dmoz/doc; take a look, folks, and see if I made it clearer. – 2*6 05:50, 1 February 2007 (UTC)

No links in headers of sortable tables?

Looks like declaring a wikitable "sortable" converts links in the header row to be plain-text.

letters numbers[2]
C 1
B 3
A 2

Why can't I have it both ways? Looks like a rendering/js bug (I'm using default monobook): if my machine sluggish, the links appear link-colored briefly, until the "sort this column" box appears, then they become plain-text colored. DMacks 03:51, 30 January 2007 (UTC)

I don't know what the intent of this behavior was, but I can confirm that it is done intentionally by sorttable.js. It's part of the ts_getInnerText function. Mike Dillon 06:26, 30 January 2007 (UTC)
Seems to be working fine for me. Perhaps it's been fixed since you posted? If it's still not allowing links in the header row, what browser are you using? -–RHolton– 19:28, 30 January 2007 (UTC)
This was a bug in sorttable.js, and has been fixed. If it's not working for you, refresh your browser cache. —Ilmari Karonen (talk) 23:19, 30 January 2007 (UTC)
Works now. Thanks! DMacks 17:40, 31 January 2007 (UTC)

Equation class

Hello. I was wondering if someone could point me to the class for laying out an inline equation. I know that it's <img class="tex" .... >, but I can't seem to locate it in either Common.css or Monobook.css. Can anyone point it out to me? Thanks, 00:24, 30 January 2007 (UTC)

Try looking here: m:Math.↔NMajdantalk 00:40, 30 January 2007 (UTC)
Thanks, that pointed me to Help:User_style#CSS_selectors which I hadn't thought of, but unfortunately that just tells me it's img.tex, which I already had available from the tag. I cannot find img.tex in either of Mediawiki:Monobook.css or Mediawiki:Common.css. Is there another one lurking? - 01:10, 30 January 2007 (UTC)
Are you trying to do something else other than display an equation? Such as use the css class for another purpose? It shouldn't be in monobook, since that can be changed by the user (by selecting a different style). --MECUtalk 01:17, 30 January 2007 (UTC)
Why are you trying to do <img class="tex" ...>? That's generated by TeX, but you shouldn't be writing it. You should be writing <math>...</math>. — Randall Bart 23:45, 1 February 2007 (UTC)

Strange redirect

I updated the redirect at 10q to point at SEC filing instead of List of Internet slang phrases. 10q does show the new value when edited, but when I visit 10q I am still taken to the old redirect target (List of Internet slang phrases)? Thue | talk 10:46, 27 January 2007 (UTC)

It's working for me, and by now probably for you too. I believe this generally happens when the slave database servers lag behind the master, causing very recent changes not to show up on the actual page. The fact that, in this case, the change was to a redirect is just an extra twist on a common problem. If you experience similar problems in the future, appending ?action=purge to the URL of the page should show the very latest version, and will also clear any other server side caches that might cause old versions to be shown. Or you could just wait a while and it should sort itself out on its own. —Ilmari Karonen (talk) 12:02, 27 January 2007 (UTC)
It is actually still redirecting wrongly for me. And I actually did try purging before I posted the initial post. Thue | talk 12:16, 27 January 2007 (UTC)
It's working fine for me as well, so it's likely a browser caching issue. See another discussion dealing with the same thing Wikipedia:Village_pump_(technical)#Redirects_to_anchors. Jayden54 12:28, 27 January 2007 (UTC)
Browser cacheing... of course, thanks :). Clearing the cache worked. I had thought Firefox would be better than that at checking for updated versions... Thue 01:14, 28 January 2007 (UTC)
This is a caching issue, but split between browser and the Wiki servers. The Wiki side lag is especially bad for redirects, because somehow they store redirectness separately from page content. Whenever that redirectness changes, whether changing content to redirect, creating a redirect, or changing redirect to content, there is a lag where it does not work. Worse, in the process of changing a redirect from one target to another it sometimes loses the redirectness for a while. Usually the lag is only a few seconds, so reloading the page will clear it. Sometimes, however, the lag runs several minutes and it can be maddening the first time you run into it. — Randall Bart 20:55, 1 February 2007 (UTC)

Is there some central list of open proxies detected by Wikipedia admins?

I'd like to add them to the block lists on a site I maintain. Thanks! --BenBurch 05:45, 3 February 2007 (UTC)

Check Wikipedia:WikiProject on open proxies. -- ReyBrujo 06:14, 3 February 2007 (UTC)
Thanks! That's it. --BenBurch 06:16, 3 February 2007 (UTC)


Is Wikipedia on Firefox incompatible with WindowBlinds or something? Whenever I use a skin other than XP or Classic, the text, and just the text, shift up and down. It only moves a little bit, but it's continuous and really annoying. Any help? Zeratul En Taro Adun!So be it. 18:09, 2 February 2007 (UTC)


Wikipedia, along with all of the other WikiMedia sites were just down from 19:44 to 19:58 GMT. Is there any reason for this recent downtime? That was very sudden, and I was wondering if it's possible to provide an explanation. Thanks!
--Soliloquial 20:05, 29 January 2007 (UTC)

You'll find the server admin log here. -- John Broughton (☎☎) 03:17, 3 February 2007 (UTC)

What, exactly, is wrong with double, triple, quadruple, etc. redirects?

"Move page" says to "be sure to check for double redirects (using "What links here") after the move. You are responsible for making sure that links continue to point where they are supposed to go."

  • What, exactly, is wrong with double, triple, quadruple, etc. redirects? The software resolves them properly, and we are told that "redirects are cheap," so what's the problem? What exactly is gained when we replace [[MIT]] with [[Massachusetts Institute of Technology|MIT]]? The end-user hardly tell the difference between MIT and MIT.
  • If it is a problem, the replacement of redirects with direct references would seem to be a completely mechanical procedure; why can't it be automated?

Dpbsmith (talk) 22:41, 28 January 2007 (UTC)

Double redirects are redirect pages that redirect to another redirect page, not what you described above. The point of checking for double redirects is stressed because the MediaWiki software does not allow consecutive redirects (that is, a page redirecting to a page that redirects to another). --TMF Let's Go Mets - Stats 22:48, 28 January 2007 (UTC)
(edit conflict) ::The problem with a double (or more) redirect is that it creates the potential for an infinite loop—page A redirects to page B, but page B redirects to page A. The server would jump back and forth between the two pages, theoretically forever, and from the reader's point of view nothing would happen. For this reason, double redirects have to be disabled and not lead anywhere.
There are bots and users who go ahead and fix double redirects, but this takes time and effort, plus first they have to find them. Therefore, it's easier and more courteous for someone who moves or renames a page to deal with them at the time. Newyorkbrad 22:49, 28 January 2007 (UTC)
I wonder, is it too hard to create a bot that would correct double/triple/etc redirects? -- ReyBrujo 22:51, 28 January 2007 (UTC)
There've been bots that have done so in the past. Presumeably, they no longer run because Special:DoubleRedirects no longer gets updated, but it would reasonable to have them watch through the move log (and perhaps Recent Changes' RSS or IRC feeds, looking for WP:AES redirects) and check those. There's also User:Cryptic/dblredir.js, which nearly automates the process for human editors post-move. —Cryptic 23:18, 28 January 2007 (UTC)
The WikiMedia software should at least resolve double redirects. I have had legitimate uses, eg, a misspelling to a redirect with possibilities. I could probably cook up a theoretically legitimate use for a triple. It wouldn't be hard to allow unlimited redirects (with a circular reference trap). — Randall Bart 21:42, 1 February 2007 (UTC)
How's this legitimate? Why not just redirect the misspelling to the actual article? What do you mean by "with possibilities"? --Golbez 11:38, 2 February 2007 (UTC)
{{R with possibilities}}. Basically, it's a redirect that serves to make a link blue, but where the redirect could conceivably be turned into an article that was separate from its target (for instance, C99 standard redirects to C (programming language), but they could conceivably be separate articles). In such a case, C99 standart (assuming, for now, that it's desirable for that to be blue) should redirect to C (programming language) via C99 standard (so that if C99 standard is turned into an article, it goes to the right place), but with double-redirects not working that can't be done at present. --ais523 13:09, 2 February 2007 (UTC)
Ah, I see what you mean. --Golbez 07:32, 3 February 2007 (UTC)
Example (not necessarily a good one): [[No pair]] redirects to [[Hand rankings]]. If I redirect [[No pear]] (a misspelling) to [[Hand rankings]], and later [[No pair]] is split off into a separate article, the redirect is now wrong. If I could direct [[No pear]] to [[No pair]], I could head off that problem. No, I'm not going to add [[no pear]]. Redirects like that are beyond the pail.— Randall Bart 05:41, 3 February 2007 (UTC)

Cross-wiki links

I know this is stupid, and I've looked everywhere for it... but... if I wanted to link to, say, an article in wikinews, is there any way to do it besides []? Because that looks really bad in a signature when editing. My wikinews sig currently looks like this:

[[User:GofG]]<sup>[[User talk:GofG Talk]]</sup><sub>[ WP]

Is there anyway I could do that without the external link? I assume something like [[wn:User:Gofg]] or something? GofG ||| Contribs 16:02, 25 January 2007 (UTC)

n:User:GofG, which can be piped in the usual fashion. --ais523 16:46, 25 January 2007 (UTC)
And to link from Wikinews to Wikipedia, you can write w:User:GofG. --ais523 18:50, 25 January 2007 (UTC)
See Help:Interwiki linking. --Teratornis 18:56, 2 February 2007 (UTC)

Portuguese Ceylon

In the article Portuguese Ceylon there is an (apparently )correctly typed link to Twelve Years' Truce that shows up as red-linked. I can't fix it. Help! Iwalters 04:29, 4 February 2007 (UTC)

Fixed. That is the good thing about knowing Spanish, you can tell the difference between ` (ì), ' (í) and ' :-) -- ReyBrujo 04:36, 4 February 2007 (UTC)

Help with a query for domains

I am trying to create a list of the most linked sites, and although I got a relatively fast query, it has a drawback. The query right now looks like:

SELECT COUNT(el_to) AS total, SUBSTRING_INDEX(el_to, '/', 3) AS search
FROM externallinks, page
WHERE page_id = el_from AND page_namespace = 0
GROUP BY search

(although I believe Vim sorts faster than MySQL in some cases :-)). Anyways, the query works, but it is not able to group subdomains as one (in example, and are considered two different sites). And although I should use el_index, my guess is that it is impossible to do it automatically. I am just hoping someone had created a Perl or Python script that groups the subdomains as one that I could borrow without having to write it myself :-) -- ReyBrujo 03:23, 4 February 2007 (UTC)

I think you're right the el_index can't be used easily because you'd have to split gTLDs from ccTLDs and the rules for how many components a subdomain has varies by country. It would be a pretty hairy case statement with a bunch of LIKEs and SUBSTRING_INDEXes. I took a quick look through CPAN, but I didn't find a module that handles the ccTLD differences. Mike Dillon 03:59, 4 February 2007 (UTC)
Talk about standards :-( -- ReyBrujo 04:37, 4 February 2007 (UTC)

Personal Mirror

Is there a way that I can have a personal copy of the wiki on my own computer, possibly using a database dump? I looked, and couldn't find anything. It would be nice if I did not have to run a web server, since I can't seem to get MySQL and PHP to work with apache. ffm yes? 22:25, 3 February 2007 (UTC)

For a static dump, you can use the HTML dumps. —Simetrical (talk • contribs) 05:42, 4 February 2007 (UTC)
Actually it can be trivially easy to run your own workstation server with a middleware stack such as XAMPP. On a crummy old Windows 98 SE computer, all I had to do was un-zip XAMPP and install MediaWiki to get my own personal wiki to run. I did not try importing the entire Wikipedia database (it's big), but I imported a couple of small company wikis. For a few brief clues, see the section XAMPP#Wiki on a stick, and Personal wiki#MediaWiki as a personal wiki. --Teratornis 08:47, 4 February 2007 (UTC)

Moth Section

I tried to add a link and have now lost the main heading to 'links'. Please can some one put it back and I swear not to play around again :( —The preceding unsigned comment was added by Dormouse7 (talkcontribs) 20:55, 3 February 2007 (UTC).

The edit was made by (talk · contribs · WHOIS) to the Moth article, which I presume must have been you. I have now fixed it. I think the mistake you made was you tried to sign an edit to an article. Signatures should only be used in either talk pages or discussion pages such as this one. Don't worry, feel free to carry on editing as since this is a wiki, any mistakes you make can be fixed by you or anyone else. I have added a welcome message to your talk page which should contain some useful links. Tra (Talk) 21:50, 3 February 2007 (UTC)

Rollback question

How do you use rollback?? When I used it on my local test copy of MediaWiki, I only managed to revert back to the earliest version of the article. If anyone's got a screenshot, post it on here!

Thanks, --sunstar nettalk 13:51, 3 February 2007 (UTC)

It rolls back all of the most recent editor's edits. So, if the edits look like:
Vandalism from
Vandalism from
Legitimate edit from Foobar
and you click Rollback, it will remove all of's edits and take you back to the last version by Foobar. It probably went to the earliest version because you're the only person editing your test wiki, and therefore it has no previous user to revert to. --Golbez 15:56, 3 February 2007 (UTC)
Rollback is used when there's a large page history then?? --sunstar nettalk 16:17, 3 February 2007 (UTC)
Yes, rollback can be used for pages with a large page history. However, on Wikipedia, the pages are often edited by more than one user, so MediaWiki does not need to go very far back before it finds an revision made by a different user that it can revert to. If your test wiki is public, it might help if you post a link here to the page you're dealing with to make it easier to explain. Tra (Talk) 16:35, 3 February 2007 (UTC)

Default heading styles for columns

I have created some tables that I wanted the same formatting for every cell in a column. MS Word provides a way to do that quickly. However, both the XHTML and Wiki table syntaxes force me to repeat the formatting for every cell.

I would like to suggest that that the Wiki syntax be extended to use carets (^) to solve this problem. When used before the first "|-" row or "!" cell, a "^" would cause the text that follows to be repeated for each cell in that column. For consistency, columns after the first column would be proceeded with "^^". This would include header cells, only the ^ style would be overridden so that the text was made bold. If a style is provided at the cell, that would override the ^ style.


^style="border:1px solid blue;"^^style="border:2px dashed brown;"
!Column 1!!Column 2
|Text in Column 1||Text in Column 2
|style="border:none;"|No border here overrides ^ style||More text in column 2

That would be equivalent to what is shown below.

^style="border:1px solid blue;"^^style="border:2px dashed brown;"
!Column 1!!Column 2
|style="border:1px solid blue;"|Text in Column 1||style="border:2px dashed brown;"|Text in Column 2
|No border here overrides ^ style||style="border:2px dashed brown;"|More text in column 2

Will (Talk - contribs) 09:26, 2 February 2007 (UTC)

The XHTML syntax does have a way to do it, with the col element; unfortunately, it is filtered out by MediaWiki's sanitizer (you would have to request that the developers add it to the list of allowed elements).

<table style="border-collapse: collapse">
<col style="border:1px solid blue"/><col style="border:2px dashed brown"/>
<tr><th>Column 1</th><th>Column 2</th></tr>
<tr><td>Text in Column 1</td><td>Text in Column 2</td></tr>
<tr><td style="border:none">No border here overrides ^ style</td><td>More text in column 2</td></tr>

(see CSS 2.1 17.3 for why you need border-collapse: collapse)

--cesarb 16:17, 2 February 2007 (UTC)

Note, though, that the CSS2 column styling is depressingly limited. —Simetrical (talk • contribs) 19:52, 2 February 2007 (UTC)

My CSS knowledge is very limited. (I never used it at all until a few months ago.) My XHTML knowledge is mainly limited to HTML with some (ironically) XHTML table syntax mixed in. So please bear with me. I learn quick though.

Why does the software filter out column styling? Is that somehow a security risk? If so, in what way? Will (Talk - contribs) 01:57, 3 February 2007 (UTC)

The software does not filter out anything (blacklisting); doing so would be a known security risk (you always miss something). Instead, it filters in everything which should be allowed (whitelisting), which is much less riskier (if you forget something with blacklisting, you potentially have a security hole in your hands; on the other hand, if you forget something with whitelisting, it simply doesn't work). What is probably the issue is that, since col and colgroup is rarely used, it was forgotten while making the whitelist. --cesarb 11:39, 3 February 2007 (UTC)
They're rarely used because browser support for the styling aspect is abyssmal, as Simetrical noted. Enabling them could lead to the false impression that they will work across different browsers. Mike Dillon 15:44, 3 February 2007 (UTC)
It's not that browser support is abysmal, it's that the standards are abysmal on this point. —Simetrical (talk • contribs) 05:31, 4 February 2007 (UTC)
Some HTML elements perceived as not very useful (e.g., <abbr>) are deliberately left out to keep the markup relatively simple, as I understand it. Probably columns should be allowed; I've been meaning to ask Brion. —Simetrical (talk • contribs) 05:31, 4 February 2007 (UTC)

user's autogenerated password doesn't work to log in

User: says, "I appear to have forgotten that I even had an account here. I had the wikipedia system email a new password for the name I would doubtless have picked. And I received that email, so it looks like I do in fact have an account. But the new password I was given does not work with the username used to send me the email. Now the system won't try to send another password without waiting, I don't know, a day or something. I think I need an actual person to help with this. Something appears to not be working correctly somewhere. Thank you."[3] Seems like a WP:VPT question so I'm bringing it up here.

coelacan talk ý 21:57, 29 January 2007 (UTC)

Tried to get an autogenerated password again, today. Received no email at all this time even after several hours.

If someone at Wikipedia HQ needs to see the email received yesterday, or my email address, or Wikipedia username, etc.. please give me info about whom to contact.

Thank you. 19:32, 30 January 2007 (UTC)

There is no one at "Wikipedia HQ" who handles password problems. Please read Wikipedia:Contact us/login problems, which essentially says to ask at Wikipedia:Help desk, where you're likely to be told the same thing. Basically, you should just start another account, per Wikipedia:Why create an account?; it's no big deal. -- John Broughton (☎☎) 03:09, 3 February 2007 (UTC)
Not true. Server admins can help with this. Ask, for instance, User:Hashar. Also, if the password isn't accepted it might mean a bug, or at least some interface issue. —Simetrical (talk • contribs) 05:27, 4 February 2007 (UTC)

Design flaw in the login page?

On the login page, there is a "Remember me" check box. Even if I've never checked that box, the login page always remembers my login name. The behavior is contradictary to what the box says. When the users choose "not to be remembered", total amnesia is expected. I prefer to stay anonymous, but my identity is revealed to the next person who uses the computer. This behavior is unfriendly to users who need to use shared computers in public library. Will that be fixed soon? Kowloonese 22:44, 26 January 2007 (UTC)

BTW. Is this the right place to post this? I couldn't find the link to the bug report page anymore. I remember the bug report used to be tracked in an external system. Where is it now? Kowloonese 22:47, 26 January 2007 (UTC)
It's your browser that is remembering/showing your username, not Wikipedia. The "Remember me" checkbox is essentially a "keep me logged in box". Check your browser's configuration to see how to disable automatic filling in of forms. --TheParanoidOne 23:21, 26 January 2007 (UTC)
It's a browser thing. If you check the box, you get a different type of cookie that logs you in transparently every time you visit the site, so you wouldn't even get the login screen. Which browser are you using?
And yes, this is the correct place to ask about this. The bug reporter,, would have pointed you back to here, as it isn't a bug with MediaWiki per se. Titoxd(?!?) 23:38, 26 January 2007 (UTC)
As others said, the "remember me" box sets a cookie which automatically logs you in. The login name is always remembered in another cookie; simply clear your cookies and it'll be gone. --cesarb 00:51, 28 January 2007 (UTC)
I see the same problem in IE 6 and Safari. I don't agree this problem is not a MediaWiki bug because I don't see this kind of symptoms with any other website. For example, in, in etc., as long as I don't check the remember me checkbox, the login name is not remembered. Note that the browser settings are the same for these websites too. Even though it were the browser that was keeping the cookie, wikipedia is lacking a certain step to clear it when you log out. If other websites can clear it, why can't wikipedia? Hence it is a bug by my book. Kowloonese 21:26, 29 January 2007 (UTC)
It's not that Wikipedia doesn't clear it; instead, it (on purpose) retains it after you log out. It's not a bug, it's a feature. --cesarb 07:31, 4 February 2007 (UTC)


There appears to be a problem with displaying the accessyear for citations in the featured article B-17 Flying Fortress. Sometimes it works and sometimes it does not. For example reference 13:- Tate, Dr. James P. (June 1998), shows the accessdate but not the accessyear. It is formatted wrong on the page or is there a bug. Snowman 11:20, 5 February 2007 (UTC)

That ref was made by calling:
{{cite book |last= Tate |first= Dr. James P. |title= The Army and Its Air Corps: Army Policy toward Aviation 1919–1941 |origdate= |origyear= |origmonth= |url= 1-42891-257-6&id=pZyLTfJFaEgC&dq=The+Army+and+Its+Air+Corps+Army+Policy+toward+Aviation |accessdate= 16 January |accessyear=2006|edition= |year= 1998 |month= June |publisher= Air University Press |location= Maxwell Air Force Base, Alabama |isbn= 1428912576 |pages= p. 164 |quote= }}
which gives (template call expanded):
Tate, Dr. James P. (June 1998). 1-42891-257-6&id=pZyLTfJFaEgC&dq=The+Army+and+Its+Air+Corps+Army+Policy+toward+Aviation The Army and Its Air Corps: Army Policy toward Aviation 1919–1941. Maxwell Air Force Base, Alabama: Air University Press, p. 164. ISBN 1428912576. Retrieved on 16 January. 
which is a violation of the spec of template:cite book as it uses both accessdate and accessyear. The spec of cite book says:
  • accessdate: Full date when url was accessed, in ISO 8601 YYYY-MM-DD format, eg. 2006-02-17. Required when url field is used. Must not be wikilinked.
  • OR: accessyear: Year when item was accessed, and accessmonth: Month when item was accessed. If you also have the day, use accessdate instead. Must not be wikilinked.
So, either give the full date on accessdate or use accessyear/ accessmonth. On another note, there is something borked with the value given for the url param as well (I haven't analyzed that, but the space in the url smacks strange to me). I doubt though whether a google search link is a good idea for a reference. --Ligulem 12:06, 5 February 2007 (UTC)
The link for the url should probably be given as:
The fixed template call might then be:
{{cite book |last= Tate |first= Dr. James P. |title= The Army and Its Air Corps: Army Policy toward Aviation 1919–1941 |origdate= |origyear= |origmonth= |url= |accessdate= 2006-01-16 |edition= |year= 1998 |month= June |publisher= Air University Press |location= Maxwell Air Force Base, Alabama |isbn= 1428912576 |pages= p. 164 |quote= }}
which expands to:
Tate, Dr. James P. (June 1998). The Army and Its Air Corps: Army Policy toward Aviation 1919–1941. Maxwell Air Force Base, Alabama: Air University Press, p. 164. ISBN 1428912576. Retrieved on 2006-01-16. 
--Ligulem 12:15, 5 February 2007 (UTC)

Your analysis is appreciated. Some references on the page allow both accessdate and accessyear, but "cite" is not consistent with this trend, which was unexpected. It would be good if all the accessdates could be quoted in the same formats including those in "cite". Snowman 13:24, 5 February 2007 (UTC)

Talk:Criticisms of Communist party rule

Maybe it's just me. I did a history merge of the article and talk page, undeleted the talk, but now I can't find the history. The deletion log says it's undeleted, but I get the "undelete _ edits" when looking through the talk page. Then when I click it says "no edits in archive". Is the old history in server limbo or something? Or is this just happening on my end? -Royalguard11(Talk·Desk·Review Me!) 23:12, 4 February 2007 (UTC)

This is a caching problem, and a known bug. A hard-reload when viewing the histories should clear it; does this work? 02:52, 5 February 2007 (UTC)
Agree with the above. I have done several page history merges; the history view always needs a forced reload (Shift-Reload on Firefox) before it updates (yes, it's a known bug — MediaWiki simply doesn't invalidate the cache for the history page when undeleting). --cesarb 03:23, 5 February 2007 (UTC)

Thumbnails/resized images in personal wiki...

My personal wiki doesn't seem to be displaying thumbnails properly... in fact, it spits out this error:

Incomplete GD library configuration: missing function imagecreatefromjpeg

Any comments/opinions? I think it may have something to do with PHP, as opposed to a setting on the wiki, but I can't seem to figure it out... Kareeser|Talk! 21:28, 4 February 2007 (UTC)

Assuming it's a machine you control you are best off installing imagemagick and configuring mediawiki to use that instead. Plugwash 21:56, 4 February 2007 (UTC)
I'm interested, but I'm not sure how I would go about installing that... and you assume correctly, it is a machine I control :)
I'm also wondering why that error was there in the first place... Kareeser|Talk! 22:16, 4 February 2007 (UTC)
That error suggests that PHP doesn't have access to version 2 of the GD Graphics Library. Dragons flight 22:23, 4 February 2007 (UTC)
Installing ImageMagick is easy, simply download it, install the default package, and alter a few lines in LocalSettings.php to tell Mediawiki to use it and how to find it. Dragons flight 22:36, 4 February 2007 (UTC)
Thanks for everybody's help. Turns out I just didn't have the gd2 and exif dll uncommented... silly me... Wikipedia uses gd2, right? 01:00, 5 February 2007 (UTC)
No, Wikipedia and all Wikimedia sites use ImageMagick, which MediaWiki's installer will detect if possible during the setup process. This can be installed for most platforms, including Unix/Linux, Mac, and Windows. 02:54, 5 February 2007 (UTC)
Performance of Wikis differ from server to server due to different settings between different servers. If it is a problem on one's own server, the best solution is still trial and error. --Deryck C. 12:44, 5 February 2007 (UTC)

Unwatch issue

I have noticed that if, after I unwatch an article, I change my mind and click on "watch" on the resulting screen, then spaces in the article title are replaced by plus signs ('+'), and the desired article is not properly added to my watchlist. For example, with the Harriet Beecher Stowe article, taking the steps of unwatching, then clicking watch on the resulting screen leads to the removal of Harriet Beecher Stowe from my watchlist and the addition of the nonexistant Harriet+Beecher+Stowe to my watchlist. Is this a "bug"? I apologize if this is not the best place to bring this up or if this has been brought up before. Thanks in advance.--GregRM 20:54, 4 February 2007 (UTC)

Fixed by editing {{MediaWiki watch page}}. --TheParanoidOne 23:01, 4 February 2007 (UTC)
Great...Thank you very much!--GregRM 01:20, 5 February 2007 (UTC)


The wiki looks great from the bb browser. Good wook whoever set that up! It is shur a shame its such a pain to type the code to sign. --User:J.smith

If you get the little buttons above the editing box that we do in major browsers, one of them is a "Your signature with timestamp" button. If not, whoever is doing the code for BlackBerrys around here (is that WML? Just a custom stylesheet? I don't know at thing about BlackBerrys) can implement a short-cut like that for BlackBerry users, site-wide. — SMcCandlish [talk] [contrib] 11:18, 4 February 2007 (UTC)
BlackBerrys (at least the version I've used) don't support css or javascript (which makes it impossible to use the toolbar buttons). They see the same HTML, however, that is generated for everyone else. Tra (Talk) 12:01, 4 February 2007 (UTC)
Yeah, they don't like the JS links... Ummm are you sure about that Tra? My display is diffrent enough to make me wonder. Like, I get "Jump to" links near the top and a few other things. Oh well, I think I can setup an auto-text to do the ~~~~ for me. I'll need to look into that. ---J.S (T/C/WRE) 18:56, 4 February 2007 (UTC)
The 'jump to' links are sent to all users of the site, except that for most people, they are hidden by css. However, for users of BlackBerrys and screen readers, they are not hidden in this way, so they appear at the top of the page. They work by linking to an anchor lower down the page. Tra (Talk) 19:56, 4 February 2007 (UTC)
Ah, I guess that explains that. Well, either way it's nice to be able to confirm that this site works decently well on a blackberry. ---J.S (T/C/WRE) 02:22, 5 February 2007 (UTC)
AFAIK, there's a preference to disable these links, so some people (the ones who disabled it) might not receive them. --cesarb 14:45, 5 February 2007 (UTC)

Looking for a Complete List of tag types

Where can I find a list of tag types with curly brackets {{ }} e.g. stub Dhammapal 04:37, 3 February 2007 (UTC)

Wikipedia:Template messages is probably the closest thing. Curly brackets are used to call templates, of which there are a very large amount and a wide variety of different types. I'm not sure if there's a complete list of them though. Richard001 04:50, 3 February 2007 (UTC)
The complete list can be found using here using Special:Allpages. Mike Dillon 05:04, 3 February 2007 (UTC)
That's not the complete list of everything which can be used within {{ }}; there's also Help:Magic words. By the way, the complete list of stub tags is at Wikipedia:WikiProject Stub sorting/Stub types. --cesarb 11:34, 3 February 2007 (UTC)
Being even more pedantic: Custom parser hooks often use the curly bracket syntax, such as the m:ParserFunctions extension. To boost pedantry to even greater heights: any editable content page on Wikimedia (and any install of a recent version of MediaWiki allowing it) can be transcluded with {{ }}, you just specify the namespace to dis-assume Template:, for example: {{:Main_Page}} {{User:Splarka/neverforget}} {{Help:Magic words}}. And in case you are still reading, even some Special: namespace pages can be transcluded, like: {{Special:Allpages}} {{Special:Prefixindex/User:Splarka/}} {{Special:Newpages/5}} {{Special:Wantedpages/5}} {{Special:Recentchanges/5}}. And even then, this is not the complete list. --Splarka (rant) 13:35, 3 February 2007 (UTC)
Noted. Still, I assume the Special:Allpages link provides what the original poster was looking for. I'm well aware of the more general nature of that syntax. Mike Dillon 15:40, 3 February 2007 (UTC)

Detecting when text was added to an article

Is there any way to detect when a certain passage of text (e.g. vandalism, something needing a source) was added to an article? Currently I can only find this by manually looking through the article's histroy and very slowly trying to narrow it down. However, if the text was rare, e.g. those exact words had only ever been added once, surely there could be a search feature that could help editors track down specific text changes to combat vandalism, improve veracity etc.

My question is, does anything like this exist already? Has it been suggested before? And if not, could it be integrated into the MediaWiki software, or created as an add-on? My programming is limited to hello world in a couple of languages, but I'm sure it would be possible. Richard001 20:29, 2 February 2007 (UTC)

Usually it's found by doing (by hand) a binary search, which is very fast. Since it's so easy to do by hand, there isn't much a need for a special search feature for that. --cesarb 22:13, 2 February 2007 (UTC)
Yes, that's basically what I've been doing, but a manual binary search is hardly easy. In some pages you have to go through dozens of pages of history just to get to the starting point, and then look through pages upon pages of diffs to find the right target. Surely an automated process could do this a lot faster. Richard001 00:04, 3 February 2007 (UTC)
Well, this is just speculation, but one option might be to use Special:Export to dump up to 100 revisions at a time into a single document, which you can then easily search. (If you're searching, say, through 400 revisions, you could use a binary search to narrow the range to 100, then do a dump, I guess.) -- John Broughton (☎☎) 02:48, 3 February 2007 (UTC)
I think a better solution would be to use the MediaWiki API to fetch the revision ids for the entire article history and then fetch the content of only the revisions needed to do the binary search. Mike Dillon 03:00, 3 February 2007 (UTC)
This feature has been desired for some time now. The only problem is getting someone to code it. A recently-added committer is working on this right now, so we'll see where that goes. —Simetrical (talk • contribs) 05:35, 4 February 2007 (UTC)

By the way, a standard binary search doesn't actually work for this problem. A binary search requires a list where you can pick the "right" or "left" half of the list to continue searching. If there is no way to rule out one half of the data, this algorithm can't be used.

Let's say you start in the middle and don't find the text, which side do you pick? It's possible that some variant on the algorithm would work, but it's not immediately clear to me what that variant would look like. You would probably start by looking at the latest revision to see if the text is in it, but even then a binary search won't let you find either the first or the latest occurrence; it can only help you find that it was added at some point. If it was repeated added and removed (as can be the case with vandalism and unsourced statements), you may have to keep looking.

Finding the first occurrence would require an iterative binary search.

  1. Find a revision where the text was added using a binary search
  2. Keep binary searching between the first revision and a revision that is known to have the text until you can find no more additions

Another question I have is, would you want to find the first occurrence, or the most recent time it appeared in the article history? Finding the most recent addition is more expensive since you have to check every revision in the case where you find an addition and the text stays in every revision until the current one. Mike Dillon 03:38, 3 February 2007 (UTC)

The assumption underlying the binary-search approach is that there are no reversions: the content was added at some point and then stayed in until the present. Thus you would first examine a middle revision. If the text is present, you'd go earlier, and if it's not, you'd go later. This works in most cases, but if you need to take reversions into account you have to examine every revision in sequence until you find one with the desired trait. —Simetrical (talk • contribs) 05:35, 4 February 2007 (UTC)
Not really. It works quite well when there are reversions, if they are next to the edit they are reverting (the usual case) and have one of the several dozen standard revert/undo edit summaries (also very common). You simply skip the reverted edit in that case (pretending both the edit and its reversion are a single null edit). Of course, that cannot be automated. --cesarb 03:30, 5 February 2007 (UTC)

Purging categories of articles which should not be in it

The template populating Category:User que was recently changed to instead add articles to Category:User qya. Category:User que was then deleted, but still had articles in it which shouldn't have been in it (thus causing it to appear in Special:Wanted categories). It started with about 70 entries and has slowly dropped each day (at the time of this edit, it has 20 pages). Is there a way to clear these out quickly without bothering with null edits? (I'm going to go ahead and clear it out now, but I'd like to know if there is a better way) --- RockMFR 21:40, 28 January 2007 (UTC)

You could AWB the null edits. That makes it trivial. -Amark moo! 21:41, 28 January 2007 (UTC)

That's what the job queue was invented for. Just wait, and it'll be cleared out. 03:03, 5 February 2007 (UTC)

Section redirecting problem

Hi. Please try the re-direct Manseibashi Subway Station. In my browser (Konqueror), it adds two pound signs to the URL instead of just one, and I'm stuck in a room with R2 until I manually use an index.php link with redirect=no to see where the it was supposed to re-direct.-- 23:18, 5 February 2007 (UTC)

Seemed to work fine to me, redirected to Manseibashi Station, subsection "Manseibashi Subway Station" with no problem. No redundant ## were added. Using Firefox 2.00001. SkierRMH 02:18, 6 February 2007 (UTC)

Forcing template substitution

I hesitate to write a bug report for this until it's fully discussed, so here goes:

  1. A problem occasionally occurs when neither substituting a template nor transcluding it is required. Often, templates can't be dated, because m:ParserFunction dates (like {{#time:j}}) and magic word dates (like {{CURRENTDAY}}) require substitution. If a template like {{wikify}} needs to be transcluded, perhaps the two-tiered approach of {{prod}} would work. In this case, "wikify" would always be substituted, and other one would need to be transcluded, rather than having a bot do the dating. Automated warning of people for not substituting a template is peachy, but it's just easier to automatically substitute it. If people don't want to conform in the substitution of a template for whatever reason (not aware of how substitution works, find that typing out "subst:" is time-consuming), this would help them.
  2. There could be an mode, preferably set by an admin, for a particular template to substitute itself by default when it is called {{in this manner}}. This mode would be akin to page protection in its setting, just more dynamic and positive.
  3. Saving a page with {{uw-vandalism1}} would automatically transclude that template. Perhaps by writing {{trans:uw-vandalism1}} can such templates remain in transcluded form, and not be auto-subst'd.

I had an idea for this a while back, and was surprised to find that I'm not the only person to think of it; see Wikipedia talk:Proposed deletion#Prod for templates idea. This appears technically possible, but if no developers feel like doing it, that's fine. I submit this technical proposal to the community for review; any counterpoints would be helpful. GracenotesT § 19:23, 4 February 2007 (UTC)

Note: given this large a community, this has probably already been suggested, but a search did not yield much. GracenotesT § 19:57, 4 February 2007 (UTC)
Feel free to vote for bug 2003 —Simetrical (talk • contribs) 22:15, 4 February 2007 (UTC)
Lovely, thanks. GracenotesT § 22:36, 4 February 2007 (UTC)
Just an opinion. I think it is impractical (or not suitable, whatever) to separate templates into "autosubst" and "autotrans" divisions. This would make editing difficult and add more and more difficulties to new users. --Deryck C. 12:48, 5 February 2007 (UTC)
Difficult to edit what? Not sure about what you're saying. Just to note, a possible drawback of this proposal is that this flag must be checked every time any template is called, but I don't think that this will be so detrimental. GracenotesT § 21:21, 5 February 2007 (UTC)

Infobox format trouble?

On Singapore Changi Airport, it is claimed that the name in alternate languages in the infobox is unreadable. Is it true? can it be fixed? What causes it? I think that the infobox is a great place for long strings of information and lists, but it is a matter of concern if it is unreadable.--Filll 18:48, 4 February 2007 (UTC)

On default Firefox sizes, I agree the font is not very readable. It's quite possible IE is incapable of showing the characters at all and displays black boxes.Circeus 19:11, 4 February 2007 (UTC)

The claim is that with Firefox at 1024 resolution, the text is split on two lines. There is no way to fix this so it works properly for a wider range of browsers at a wider range of resolutions?--Filll 20:11, 4 February 2007 (UTC)

I can see them all using IE 7. Corvus cornix 17:49, 5 February 2007 (UTC)

Case sensitivity in headings & browser variation

If I put a link like such as [[Contract bridge glossary#c|Caddy]], (test: Caddy) it will work properly in Internet Explorer, but not in Firefox. The reason is that the heading is defined as a capital C on the target page and IE is not case-sensitive there, but Firefox is. Unfortunately, this means that I can make an edit and test it successfully, but it won't work for others.

It would be nice if WP would automatically change both the references and the definitions to a consistent case (e.g., always lower) when rendering the HTML, so this problem would not arise.

I'm hoping that somebody knows whether this bug/feature has already been logged in BugZilla and will take care of it if it's not there. I already waste enough time on WP without getting into BugZilla as well. Matchups 05:04, 3 February 2007 (UTC)

I would oppose that change. Currently we can simply (in most cases) copy the heading and use it as the anchor, which is a nice property; if the software changed it to lower case, it would no longer be a case of simple copy-and-paste. --cesarb 11:31, 3 February 2007 (UTC)
Please note my second paragraph--both sides would be changed to lower case, so you'd still be fine in the suggested example. Matchups 02:43, 4 February 2007 (UTC)
That would only work when the link is created within Wikipedia itself; the way it's currently done, I can do the same when creating a link from outside Wikipedia to within Wikipedia (unless the heading has special characters). If the software was changed like you propose, the behaviour for a simple anchor (without any special characters) would differ whether the link is a wiki link or an external link, which is surprising behaviour. --cesarb 07:28, 4 February 2007 (UTC)
*sigh* You're right. Anyone have a better idea how to deal with this? Matchups 03:22, 5 February 2007 (UTC)
Maybe when an internal link to a section of a page is made, MediaWiki should then check the contents of that page (or a separate table with a cached list of headings to optimise performance) and see if there are any headings on that page which are the same name as what is linked to but a different case. Then, when the HTML is rendered, the case could automatically be corrected to match the heading on the target page. e.g. with the example you gave:
  • [[Contract bridge glossary#c|Caddy]] is typed into a page
  • MediaWiki checks the headings in the page Contract bridge glossary
  • The heading 'C' is found which matches the link 'c' but with a different case
  • The link 'c' is automatically corrected to 'C' when the page is rendered
  • The link is rendered as <a href="/wiki/Contract_bridge_glossary#c" title="">Caddy</a>
This method would mean that the correct anchor link appearss in the address bar, and also that it is less likely for someone to make mistakes. However, it would impact performance and there may need to be alterations to the database design. Tra (Talk) 17:52, 5 February 2007 (UTC)

How do I get to the wikipedia chat room?

I know there are several wikipedia chat rooms. What's the biggest one, and how do I get there? —The preceding unsigned comment was added by Andrewdt85 (talkcontribs) 05:21, 7 February 2007 (UTC).

The Wikipedia "chat rooms" or channels, as we typically call them, are all located on the IRC network. See m:IRC for more information about how to download an IRC client, connect to the network, and join a Wikipedia-related channel. AmiDaniel (talk) 05:26, 7 February 2007 (UTC)

database corruption?

The whole content of Wikipedia:Articles for creation/Today has been mysteriously replicated at the top of Wikipedia:Articles for creation/2006-12-16 where a bunch of "7 February 2007" submissions have appeared. Since the archive's history [4] has no activity since 1 February 2007, I know it's not some manual error but a database problem at some level.

I dunno if this is some temporary glitch or a permanent corruption and a serious bug, so I'm reporting it to you. I found out about it while looking at the "What links here" of my suggested article [5] 05:17, 7 February 2007 (UTC)

I've fixed the duplication. The duplication was added in this edit. Perhaps the editor was confused about template syntax. The database was not corrupted in this case; the page Wikipedia:Articles for creation/Today was just inappropriately transcluded. Graham87 08:11, 7 February 2007 (UTC)

How to add pictures

How do I add pictures so they're smaller, but if clicked on, will open up in full size? Thank you in advance.

N734LQ 04:53, 7 February 2007 (UTC)

See Help:image especially for the description of the thumb parameter. Also, a look at the wikitext of an article with several thumbs might be informative. See Mount Hood. —EncMstr 04:58, 7 February 2007 (UTC)

Making table of contents

Please help. How do I make the table of contents that contain the links that take you to a certain section on the page? N734LQ 00:59, 7 February 2007 (UTC)

It should appear automatically when the page has enough headings and it should appear immediately before the first heading. If, however, you want to control when it appears and where it goes, you can put __Toc__ where you want the table of contents to go. Tra (Talk) 01:04, 7 February 2007 (UTC)

Category includes templates that include the category, resulting in the category including itself.

Please see Category_talk:Female_Wikipedians. Anyone know how to fix it? Thanks. Xiner (talk, email) 23:50, 6 February 2007 (UTC)

For when the categories are specified in the templates, I've changed the syntax from [[Category:Female Wikipedians|{{PAGENAME}}]] to {{{category|[[Category:Female Wikipedians|{{PAGENAME}}]]}}} so that the template can be called like this: {{user female2|category= }} if you want to disable the category for that use. Tra (Talk) 00:56, 7 February 2007 (UTC)

how to extract all Wikipedia Reference Desk content?

I would like to extract all the contents of the Wikipedia Reference Desk [6]. I could parse the pages, but it would be much easier I think to extract the data from Wikipedia's MySQL database. Which database snapshot should I download? Some database snapshots only contain articles. In which snapshot is the Reference Desk contained? I assume I can easily import the snapshot into a local MySQL database, and then run queries to extract the content? Jawed 23:21, 6 February 2007 (UTC)

Does Special:Export get you close? It's easy to limit to one article, and since there are only about 20 revisions, the 100 revision-at-a-time limit doesn't apply. —EncMstr 05:50, 7 February 2007 (UTC)

Proposal for new table class in .css

I made a proposal for changes to the CSS files at Wikipedia:Village pump (proposals)#Proposal for new table class in .css. —Doug Bell talk 22:47, 6 February 2007 (UTC)

Long articles causing Safari browser to crash after last Jan. 2007 MacOS X Security update?

When I try to do a history diff between two versions of a long article, such as Glossary of pool, billiards and snooker terms, Safari crashes on me. The only thing that has changed on my system (other than, of course, the creation of cache files and the like) between "when this didn't happen" and now are that I allowed MacOS's Software Update application to install the most recent Apple security patch for the OS, and reboot. Anyone else seeing this behavior? Either my system just suddenly has something inexplicably messed up, or the latest Apple patches inadvertently introduced what amounts to an "anti-Wikipedia bug". — SMcCandlish [talk] [contrib] 11:18, 4 February 2007 (UTC)

I'm not sure what the problem is, but perhaps you can turn off the preview when displaying the diff (under "my preferences" near the top). It makes viewing diffs a lot quicker anyway and it would reduce the size quite a bit. ---J.S (T/C/WRE) 18:59, 4 February 2007 (UTC)
If you mean "Don't show page content below diffs", I will try turning that off; I've never found it useful anyway! — SMcCandlish [talk] [contrib] 20:26, 6 February 2007 (UTC)
Did you submit a bug report to Apple? I am sure they will fix it if they know about it. --BenBurch 19:12, 4 February 2007 (UTC)
Not yet; wanted to bring it up here for a sanity check first. I note that no other Mac users have chimed in with a "me too!", so I'm still uncertain... — SMcCandlish [talk] [contrib] 20:26, 6 February 2007 (UTC)

Color coded Wiki Editing Text

Hey, does anyone know a way to make the Wiki editing text colored by default, like Dreamweaver HTML Color coding or other IDEs? Anyone interested in working on it? Let me know on my talk page if you are. Bmunden 22:49, 4 February 2007 (UTC)

"Syntax highlighting" for the Wikipedia editing text area already exists, all you have to do is to install wikEd. Cacycle 01:03, 7 February 2007 (UTC)

Edit section

I get a lot of times concurrent edit errors, and lose my added content, and the page diff shows no difference, just shows my last saved edit. When I look at the page history, it says noone edited it, besides me. (see Father's right movement around this cooment's date) (section was without time stamp - moved here by Cacycle 03:51, 7 February 2007 (UTC))

Edit Count

Where can one find a count of one's cotributions?--Elatanatari 01:21, 9 February 2007 (UTC)

See WP:COUNT. For example, this is one of the edit counters. Jacek Kendysz 01:32, 9 February 2007 (UTC)

User box can't get image

Could someone explain why this user box instance is expanding incorrectly. It is acting as though the image doesn't exist. However, it does. I uploaded it to Commons. You will need to get through several layers of template calls to see the entire thing.

{{User:Will Pittenger/User Boxes/User Works at Another Wiki|Wikicities:||logo=Logo for Wikia.png}}
File:Logo for Wikia.png This user works on another wiki:

Will (Talk - contribs) 22:51, 8 February 2007 (UTC)

I don't see the image on en.wikipedia, nor at commons. image:Logo for Wikia.png Are you sure that's the right name? —EncMstr 22:56, 8 February 2007 (UTC)
It was deleted at commons by User:Yann, as a copyvio. Tra (Talk) 23:16, 8 February 2007 (UTC)

I could have sworn it read GFDL. Must have had that backwards. Will (Talk - contribs) 06:17, 9 February 2007 (UTC)

Game-cover Merge again

Some may recall 2 months ago when I posted something about a merge proposal for game-related fair use templates. After waiting a while with no objections I performed the merge. It has now been reverted by someone who thinks I did not make enough of an effort to contact interested persons to obtain concensus. So here we go again. ANYONE INTERESTED IN Template:Game-cover, Template:Boardgamecover, OR Template:RPG-artwork IS INVITED TO JOIN A DISCUSSION AT Template_talk:Game-cover#Merge ABOUT MERGING THESE THREE TEMPLATES. I'm cross-posting this to all the Village pumps. ~ ONUnicorn(Talk|Contribs)problem solving 20:28, 8 February 2007 (UTC)

Chancellor, Denice Denton

Hi, the Denice Denton article looks corrupted now, in that everything below the bottom of the University of California, career section is truncated, at least on my computer--the text is there, but in edit mode only. Can this be fixed? bob -  uriel8  (talk) 17:06, 8 February 2007 (UTC)

There was an unclosed <ref> at the end of the section. I've fixed it. Corvus cornix 17:41, 8 February 2007 (UTC)

Searching problems

I've always occasionally seen a message telling me that there was a problem with my search, and to try again in a few minutes, but lately it's becoming more and more frequent. Running Firefox 2.0 on Mac OS X 10.4.8 behind a campus LAN. Anybody else have this issue? Atchius 09:07, 8 February 2007 (UTC)

There's been a lot of maintenance being done on the servers lately... in most cases it's only for a short period. --SB_Johnny|talk|books 17:52, 8 February 2007 (UTC)

Parameter ignored

I am having problems with one of the user boxes I host. {{User:Will Pittenger/User Boxes/MLB-Cardinals}} has a parameter called {{{No Categories}}}. It is supposed to prevent the page that you are transcluding the template to being added to Category:Wikipedian St. Louis Cardinals fans. However, that parameter is causing me problems. It is being ignored in two places. One is my list of custom user boxes, User:Will Pittenger/User Boxes. The other is Talk:St. Louis Cardinals.

Strangely, it works just fine on that page. My samples all skip the category. I have tried purging every page I could think of. Why is my parameter being ignored? Will (Talk - contribs) 07:42, 8 February 2007 (UTC)

The case on Talk:St. Louis Cardinals is not passing the "No Categories" parameter because there is no equals sign; it is passing effectively "1=No Categories". Your template MLB-Cardinals requires it to pass a non-blank value. On User:Will Pittenger/User Boxes, the problem is that User:Will Pittenger/templates/User Box List Row is passing "No Categories=True" as a third parameter to the {{#if:}}, not to the selected userbox itself. Mike Dillon 15:27, 8 February 2007 (UTC)

Duh. I must have been half asleep at the time. Sorry. Will (Talk - contribs) 21:12, 8 February 2007 (UTC)

User .js: getting blank lines in the toolbox

I use this script, which is modified version of this. I changed all the things I wanted to, but there's still one unsolvable problem for me. I'd like to make the lines which include '//empty line' empty (free of text). How to do it? I tried few things (see the history of the page), but none of them worked properly. Thanks in advance. Jacek Kendysz 23:57, 7 February 2007 (UTC)

I assume you want the script to be displayed properly in one box instead of broken up whenever there's a blank line? See how it looks now. (I apologize if that's not what you wanted ... it's easy enough to revert if necessary.) In order to get it to display nicely w/o breaks in the formatting when there's an empty line, you just need to make sure that there's at least one space character at the beginning of the empty line.
Another reading of your question could be how to get blank lines into the toolbox itself, the output of this code. That I don't know. Maybe wrap the next line with something like <span style="padding-top:1em;">...</span>? --Kbh3rdtalk 15:52, 8 February 2007 (UTC)
Your second reading of my question is exactly what I want. I tried what you wrote, but it still doesn't work... Why defining the element as space ( ) doesn't work (I tried it)? Thanks. Jacek Kendysz 23:10, 8 February 2007 (UTC)

OK, the problem is solved. The script works like I want to. Jacek Kendysz 01:29, 9 February 2007 (UTC)


Hi. Why can't Wikipedia implement ASCIIMath4Wiki immediately? If a new math tag like "asciimath" is introduced in addition to the usual "math", there will be no conflict with existing equations, and authors can choose to use either method of expressing mathematics. If the ASCIIMath language is adequate to express the desired equation, authors will probably choose to use that syntax rather than LaTeX syntax, because the language is easier and the resulting equation looks a hell of a lot better. I don't know anything about the technology behind Wikipedia, but if they have this working over at Wikimedia, why can't we get it working here (where it's needed!). Thanks. —Dfass 15:42, 7 February 2007 (UTC)

The way I see it, the obvious issue is browser compatibility. Having an engine convert math expressions into images solves browser compatibility, because nearly every browser can display images. Requiring the installation of fonts would be a burden to users who use Firefox; and Internet Explorer's inability to render MathML makes matters even worse, seeing that it is the predominant browser used. --MZMcBride 04:34, 8 February 2007 (UTC)
Thanks for the response. On my system, both Firefox and IE7 were able to display the math with no further installation of packages (that I can recall, although I do have LaTeX and Unicode fonts previously installed). Could there not be a user preference that specifies display properties? Or even a toggle button on the page itself. Also, is it not possible to detect the browser's capacity? Then, if there is no facility (or fonts) detected to render the math, default to JPEG. It would look so much better if it could be made to work... Thanks. —Dfass 14:50, 8 February 2007 (UTC)
Also see m:Blahtex. The big problem is that MathML requires all page to be outputed in proper XHTML with the correct content-type. This causes issues with older browsers, however. —Ruud 18:29, 8 February 2007 (UTC)

Steward extension on my version of MediaWiki


I set up MediaWiki 1.91 on my localhost, as shown above. When I tried to set user rights for User:Testuser@wiki1 like the user rights log on Meta it said the user did not exist, even though I have a database called "wiki1". What caused this?? How do I fix it?? --sunstar nettalk 00:42, 6 February 2007 (UTC)

According to your user list in the screen shot you uploaded, it looks like you have not created that user yet. What did you do to create that user? --Teratornis 04:58, 6 February 2007 (UTC)
You seem to be mistaking the MySQL user ("Testuser@wiki1") with the MediaWiki user. They are not the same. You need to create a user for MediaWiki via the Special:Login page. --TheParanoidOne 06:48, 6 February 2007 (UTC)
  • You need to create the user first (Special:Userlogin), then sysop him (Special:Userrights). Special:Makesysop is not enabled by default, as it is a separate extension.
  • That should do if you have one wiki. If you have more of them, and you want to change user rights from one place, then there's a whole lot of more things to do... Titoxd(?!?) 02:22, 8 February 2007 (UTC)
I enabled Special:Userrights on my wiki in LocalSettings.php, I have three wikis, all running MW 1.9 on my WAMPserver. Titoxd, any idea what to do?? --sunstar nettalk 16:12, 8 February 2007 (UTC)

Blank lines at the top of articles

My biggest pet peeve lately is blank space at the top of articles. What will usually happen is that there will be the code for an infobox at the beginning and then someone will put a blank line after it and then start the article, which creates some ugly blank space at the top of the page (like here). I've just been taking out that extra line, but it's on a LOT of articles. Couldn't there be some code that can tell when it's going to do that and then just not have that blank space at the top? Recury 14:34, 5 February 2007 (UTC)

Getting a new feature added to MediaWiki to handle this might be harder than requesting a bot edit. --Teratornis 04:56, 6 February 2007 (UTC)
True, but the extra blank line in the code makes the code easier to read and edit. So just removing it from the rendered version is a better solution from the point of view of usability. -Pgan002 22:46, 7 February 2007 (UTC)
Yeah and I'm sure that's what all the people who do it are thinking of when they do add the extra line, and a bot wouldn't really stop people from doing it in the future. That said, having a bot clean up all the articles that have the empty line up there now couldn't hurt, so I may give that a shot as well. Recury 17:08, 8 February 2007 (UTC)
Uh, actually it doesn't look like it's doing it any more. Is it possible that someone made a code change that quickly? Recury 17:13, 8 February 2007 (UTC)
It is possible that someone noticed long before you, so maybe the change had nothing to do with you :-) -Pgan002 23:31, 8 February 2007 (UTC)
I just came across an article that had two empty lines after the infobox and that had the empty space at the top still. Can we ever win? Sigh. Recury 14:37, 9 February 2007 (UTC)

Help to make the MediaWiki in-browser editor script wikEd work under Internet Explorer 7 and Opera 9

wikEd is a full-featured in-browser text editor that adds enhanced text processing functions to Wikipedia and other MediaWiki edit pages. It is a JavaScript monobook.js program. Features include:

  • Regular expression search and replace
  • Wikicode syntax highlighting
  • Fullscreen editing mode
  • Server-independent local Show preview and Show changes
  • Pasting formatted text, e.g. from MS-Word (including tables)
  • Conversion of formatted text to wikicode
  • Single-click fixing of common mistakes (including Unicode fixing)
  • History for summary, search, and replace fields

Currently it works only for Firefox and other Mozilla browsers. I am looking for people who can help to make the script work under Internet Explorer 7 and Opera 9. Cacycle 06:07, 16 January 2007 (UTC)

Sounds interesting, I will try it out since I can't make the external editors to work without a root account. -- ReyBrujo 06:12, 16 January 2007 (UTC)
The code is pure browser-independent JavaScript and does not use any browser-specific API functions. Therefore it shouldn't be too difficult to tweak the code for non-Mozilla browsers. Anybody? Cacycle 22:11, 21 January 2007 (UTC)
Moreover, the code is extensively commented and you could count on my help to setup a testing environment and to make the compatibility adjustments. Have a look at the wikEd script. Cacycle 18:42, 28 January 2007 (UTC)
Can somebody point me to another place to ask for help to make it IE compatible. Thanks, Cacycle 14:07, 6 February 2007 (UTC)
If it's the wysiwyg editing that is not supported by Internet Explorer, you might want to have a look at something like HTMLArea, which I know works in Internet Explorer. It's no longer available but I have a copy of it at where you could probably have a look at how it works and maybe base your script around it. Tra (Talk) 21:51, 6 February 2007 (UTC)
Thanks for your reply. wikEd is already fully functional. All it needs are more or less minor adjustments for IE or Opera specifics. Cacycle 03:45, 9 February 2007 (UTC)

EVO Electric Vehicles for Ohio

What is this EVO? Electric Vehicles for Ohio. —The preceding unsigned comment was added by Dougboncha (talkcontribs) 15:17, 8 February 2007.

Have you tried Wikipedia's Reference Desk? They specialize in knowledge questions, and will try to answer any question in the universe. Just follow the link, select the relevant section, and ask away. I hope this helps. Jacek Kendysz 15:22, 8 February 2007 (UTC)

Issue with new template

I created a series of new templates for use with college football and basketball recruits but I'm having some issues that I could use some help with. The template I'm having an issue with is {{College Athlete Recruit End}}. You'll notice I have three sources inside the table footer (produced by this template) along with two inline citations. The Sources stated in the footer appear fine, but the inline citations do not. The refs that are produced by the inline citations do not show any variables. There are two in both refs: year and accessdate.

Here is the code in question: '''Overall Recruiting Rankings:''' {{#if:{{{scout overall|}}}|Scout - ''{{{scout overall|}}}''<ref>{{cite web | Team Recruiting Rankings | | url={{{year|}}} | accessdate={{{accessdate|}}} }}</ref>}} {{#if:{{{rivals overall|}}}|Rivals - ''{{{rivals overall|}}}''<ref>{{cite web | title={{{year|}}} Team Ranking | | url={{{year|}}} | accessdate={{{accessdate|}}} }}</ref>}}

I believe the issue is because the refs are inside of an Contribs)problem solving 20:01, 9 February 2007 (UTC)

If you've provided an email address, you can click on "E-mail new password" button at Special:Userlogin. Then follow the instructions there, and viola! It should work... once again, assuming that you've provided Wikipedia with your email address. Or if you know your previous password and want to pick another one, go to Special:Preferences logged into your account, type in your old password, type in the new one in the textboxes labeled "New password" and "Retype new password", and click on "Save" at the bottom of the page. If none of these are the case, then you may wish to get a new account, unless someone else has any suggestions. GracenotesT § 20:17, 9 February 2007 (UTC)

Template question

A while ago I posted the following at Template talk:Test, posing a suggestion and asking if it was possible to do. There has not been any response, so I figured the technical-minded people who hang out here might be able to answer the "is it possible" question.

I'm wondering if there's any way to make it so that; when warning someone with {{subst:test-n|name of page}} we could make it work with multiple pages? See, for example User talk: Water is on my watchlist; so I saw the vandalism there and reverted it, then checked the user's contributions and reverted their earlier vandalism to the other two pages. I only wanted to use one warning template as they hadn't yet been warned and it seems silly to give them {{test}}, {{test2}}, and {{test3}} without giving them a chance to see {{test}} and respond. At the same time you want them to know that all their testing and vandalism has been reverted.

I don't know enough template syntax to know if it's even possible; but is there a way that we could specify with {{subst:test-n}} that there's more than one page, how many, and the names of all pages? Maybe something like {{subst:test-n|3|Water|Protein|Lion}}? Is that possible? Is it hard? Should we edit the template to use that? ~ ONUnicorn(Talk statements and such, you need to install the ParserFunctions extension available here. On that page, there's also information regarding how to use more advanced template syntax. --MZMcBride 04:02, 16 February 2007 (UTC)

Longitude/Latitude coordinates don't render correct on some city/town pages

On the top right of town and city pages it shows the latitude and longitude, like this: "Coordinates: 28.61° N 77.23° E". On some pages, it turns into an illegible hash, in both Firefox and IE. The markup in the infobox looks identical (to me) when you compare pages that work and pages that don't: Pages that don't work:

Pages that do work:

Any ideas? Thanks, ॐ Priyanath talk 02:15, 16 February 2007 (UTC)

Yeah. You've got two templates competing against each other for the same space. Basically, if you use Template:Infobox Indian urban area (which is actually Template:Infobox Indian Jurisdiction), the coordinates function is built-in. When you try to use Template:coor title dm with Template:Infobox Indian urban area, it becomes garbled. I put HTML comments around the "coor title dm" template, and everything seems to be working fine now. Cheers. --MZMcBride 04:10, 16 February 2007 (UTC)
Thank you!!! ॐ Priyanath talk 04:11, 16 February 2007 (UTC)

Ogg file conversion

Please, please help! How do I convert sound files (place names etc) into ogg format? Where can I get a reliable ogg encoder from? Anil 00:42, 16 February 2007 (UTC)

It really depends on what operating system you're using. See here for the different options. --MZMcBride 00:53, 16 February 2007 (UTC)


I need advice on the popups. Everytime I revert vandalism using the rv button, my edit summaries look like this...

(Revert to revision $1 dated $2 by $3 using popups)

Instead of this example...

(Revert to revision 98116240 dated 2007-01-03 05:47:13 by Hmains using popups)

Is there something wrong with the rv button or my Javascript? Thank you!--PrestonH | talk | contribs | editor review | 16:10, 15 February 2007 (UTC)

Please see User talk:Lupin#Getting "revision $1 dated $2 by $3" or User talk:Lupin/popups.js#revision $1 dated $2 by $3 --Van helsing 16:19, 15 February 2007 (UTC)

Distinguish "What links here" by namespace

Another useful aspect would be something that allows users, when viewing "What links here", to choose which namespaces that appear. Instead of a big, unsorted list of all namespaces, I might want to just see what User: spaces link to an article, for instance. − Twas Now ( talkcontribse-mail ) 09:15, 15 February 2007 (UTC)

Interesting. I think I would find that useful. Not sure the devs will want to spend time on that though... Grandmasterka 11:18, 15 February 2007 (UTC)
It's been a request for a while. See Bug 4624. –RHolton– 14:16, 15 February 2007 (UTC)
The disambiguation script provides some of those features. My latest version is at [[User:Barticus88/dpl.js]]. I was about to drop the WP:DPL focus and turn it into a general whatlinkshere enhancement script. Any feature requests? — Randall Bart 00:06, 16 February 2007 (UTC)

"Invert selection" option for watchlist

In Special:Recentchanges, there is an option to "invert" the selected namespace. I think this means it will show edits in all namespaces except the namespace selected, whereas it would normally show only the selected namespace.

If this is correct, I think it would be useful (thought not imperative) to have this option available in our watchlists. Is there a more appropriate place to bring this up? − Twas Now ( talkcontribse-mail ) 08:59, 15 February 2007 (UTC)

Feature requests, as well as bug reports, can be submitted via MediaZilla. –RHolton– 14:23, 15 February 2007 (UTC)