Wikipedia:Village pump (technical)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

• Table of contents
• First discussion
• End of page
• New post

The technical section of the village pump is used to discuss technical issues. Bugs and feature requests should be made at BugZilla since there is no guarantee developers will read this page.

FAQ: Intermittent database lags can make new articles take some minutes to appear, and cause the watchlist, contributions, and page history/old views sometimes not show the very latest changes. This is an ongoing issue we are working on.

Details about the occasional slow speeds and deadlock errors: here

Please sign and date your post (by typing ~~~~ or clicking the signature icon in the edit toolbar).

Please add new topics at the bottom of the page.

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

Technical archive

Discussions older than 7 days (date of last made comment) are moved here. These discussions will be kept archived for 7 more days. During this period the discussion can be moved to a relevant talk page if appropriate. After 7 days the discussion will be permanently removed.

Template for rollback summary?

The admin rollback feature leaves an edit summary of "Reverted edits by user contributions page to last version by User. Is this an already-existing template that normal users could use for their rv summaries? Boojum 14:49, 5 September 2005 (UTC)

I don't think you can use templates in edit summaries. ~~ N (t/c) 16:16, 5 September 2005 (UTC)

I'd like this, however. mrholybrain 15:31, September 10, 2005 (UTC)

• There is a user script (godmode-light.js) that gives non-administrators the "rollback" button. However, some people seem to have trouble getting it to work. --Ixfd64 23:22, 16 September 2005 (UTC)

  Work fine for me in Firefox. --Celestianpower ^hab 20:49, 24 September 2005 (UTC)

godmode-light.js only works with that lame Internet Explorer browser, and only on that lame Windows operating system. ;-) In any case, I actually find it confusing when a non-admin uses edit summaries that have the appearance of the admin-only rollback feature, as I'm sure other users do as well. Func( t, c, @, ) 21:14, 20 September 2005 (UTC)

• As a user of the "lame Internet Explorer," what am I doing wrong? I can't get the script to work in IE. I tried it on a friend's computer with Firefox and it worked. I would appreciate any advise. Thanks. Psy guy ^(talk) 04:07, 23 September 2005 (UTC)

Ah, my apoligies...I had only glanced over godmode-light, saw a lot of ActiveXObject-type code, and made some poor assumptions. It looks like the code is written to work in any modern browser, with appropriate branching for IE and Mozilla/FireFox/Netscape, etc. In any case, this is User:Sam Hocevar's talk page. Func( t, c, @, ) 04:44, 23 September 2005 (UTC)

Newuserlog

I've installed my extension that logs new user creations at Special:Log/newusers, it's neat for finding new users to welcome, usernames to ban and for catching vandals. —Ævar Arnfjörð Bjarmason 22:20:00, 2005-09-07 (UTC)

Neat. It would be handy if each log line had a "contribs" link too. -- Finlay McWalter | Talk 22:30, September 7, 2005 (UTC)

Very nice, contribs would be nice. vandals life just got harder. Martin - The non-blue non-moose 22:37, 7 September 2005 (UTC)

Thanks for adding the contribs link. Would it be easy to add a "block" link for admins? --MarkSweep✍ 06:13, 8 September 2005 (UTC)

Nice work. Interesting to see just how many in the short time it's been up, too. Mindspillage (spill yours?) 23:04, 7 September 2005 (UTC)

Well done indeed! --Canderson7 23:16, September 7, 2005 (UTC)

That's wonderful - thanks! Shimgray 23:20, 7 September 2005 (UTC)

Excellent idea, but could we remove it from Recent changes? Zoe 23:23, September 7, 2005 (UTC)

Thank you! I love this; it should make the life a bit easier for vandal-fighters. Antandrus (talk) 01:54, 8 September 2005 (UTC)

Compugeeks like Ævar Arnfjörð Bjarmason never cease to amaze me. Well done that man! Moriori 03:45, September 8, 2005 (UTC)

Thank you. It will help prevent the inappropriate user names. But I would like to remind everyone that this new feature will not totally stop the long term vandals because I fully expect that they will start to create usernames that look legit at first... until they start editing. And so I hope we can get a feature to help curb the massive page move vandalism. Still, the new feature is a step in the right direction. Zzyzx11 (Talk) 05:12, 8 September 2005 (UTC)

A bot solution to the problem of WoW-style page move vandalism has been deployed and was discussed on WP:AN or WP:ANI fairly recently (see the recent archives). --MarkSweep✍ 06:12, 8 September 2005 (UTC)

Excellent work, thanks a lot. the wub "?/!" 13:12, 8 September 2005 (UTC)

I've created a monobook.js script that adds additional links to the Newusers log: User:Func/wpfunc/nupatrol.js Func( t, c, @, ) 06:05, 15 September 2005 (UTC)

Go logs

I was thinking about people coming to Wikipedia and being unable to find the information they need. I realise that there are privacy concerns, but would it be a good idea to make available unsuccessful "Go" logs? If we could see what people are looking for articles on, especially if it's sorted by frequency like wanted pages, then we might be better able to plan ways to help users. Bovlb 01:58:15, 2005-09-11 (UTC)

I imagine the response to this will be that it can't be done due to "performance issues", which I hae asked about below. --Commander Keane 06:43, September 11, 2005 (UTC)

I can't imagine that to be the reason. The list should be easy to generate in a batch job going through the webserver-logs once a day (along with other webserver stats I'm sure are being run on the logs on a daily basis). That it's a "go search", along with the actual search string, is right there in the HTTP-request being logged. I think it's a very good idea. Shanes 07:12, 11 September 2005 (UTC)

Sorry about that then. It seemed similar to the request about finding out which are the most frequented articles, which is usually responded to with "performance reasons". --Commander Keane 08:02, September 11, 2005 (UTC)

The fundamental difference is that page hit counting is an existing feature of the mediawiki software, but it is implemented in a way that would require disabling the front-end squid cache servers that respond to most of the page requests (which would have a huge performance impact). I suspect "Go" misses won't be implemented any time soon either, but likely because the developers have other more urgent issues (like performance improvements). Scraping the logs is perhaps possible, but even this would take developer involvement (and is not as easy as it sounds, since there are multiple web servers involved). The place to request this as a feature is Wikipedia:Feature request, or more directly by entering a feature request via MediaZilla. -- Rick Block (talk) 18:37, September 11, 2005 (UTC)

A grep for "Go" misses in the logs wouldn't have to be done on all the logs to be usefull. I'm sure the ranking of most missed search words aren't very different if you run it on just a few gigs of logs on one server. The missed go-list doesn't have to be very accurate, just something giving us a hint about potentially missing articles, misspellings or redirects. Shanes 21:13, 11 September 2005 (UTC)

File this as a feature request for Logwood statistics at Bugzilla. I would describe it as a "most wanted articles" list. Then it depends on if Kate thinks it's a good idea =) --Alterego 19:03, September 11, 2005 (UTC)

This used to be a feature years ago. I remember it when I first came here. And the searches even appeared on Recent changes. I'm sure it's a performance issue. User:Zoe|^(talk) 05:07, September 12, 2005 (UTC)

Submitted as bug #3437. Bovlb 05:20:27, 2005-09-12 (UTC)

It's not a performance issue at all, it's just a matter of adding a hook to special:log that gets run whenever the "No page with that title exists" message gets displayed, I might implement this. —Ævar Arnfjörð Bjarmason 17:37, 16 September 2005 (UTC)

Note that there is just a lot of stuff that would come into a log like that, and it's really really annoying to sort through. It clogged up Recent Changes badly back three years ago; imagine how much worse it would be now! Note that most such failures will be a) misspelled, b) miscapitalized, and/or c) mispunctuated, making them unsuitable for direct clicking as titles. Also logging with the username/ip of the searching visitor is totally unexpected and a serious privacy issue. (Additionally note that Special:Log currently has no infrastructure for logging events not attached to a user account.) --Brion 21:33, 17 September 2005 (UTC)

Again, I suggest that we sort the list by descending frequency. We could even omit those with frequency one. I don't think that listing the usernames would be appropriate. Bovlb 22:46, 17 September 2005 (UTC)

As an experiment I've installed an extension to log terms that do not directly match any article in a log on the Wikimedia cluster, nothing else is being logged. Brion was right, it's going through *fast*, currently ~3 entries per second. —Ævar Arnfjörð Bjarmason 16:20, 19 September 2005 (UTC)

• Thanks. How many repetitions are there? Maybe it would be simpler to collate the most frequent N for each day. Bovlb 18:55, 19 September 2005 (UTC)

Image caption instead of the rendered picture

After uploading Image:Parfleche,FieldMuseum.jpg into the Parfleche article I clicked on the image itself. After that the thumbnail in the Parfleche did not render to the picture, but merely to the words in the caption. Ancheta Wis 02:13, 11 September 2005 (UTC)

looks fine to me now presumablly a temporary glitch. Plugwash 03:22, 11 September 2005 (UTC)

hmm it worked fine for me the first time i visited but doesn't anymore Plugwash 03:24, 11 September 2005 (UTC)

• I believe this sometimes happens when the servers are under very heavy load. Bovlb 06:06:06, 2005-09-11 (UTC)

English Wikisource messed up?

I know this is probably the wrong place to post this, but has anyone noticed that the English Wikisource is acting strangely? First of all, the main page there shows that there are -1 documents. Also, all the usernames and IPs in the block-list seemed to have disappeared into thin air.

What's going on? --Ixfd64 23:46, 2005 September 11 (UTC)

English Wikisource is brand new, it never existed before last night. --Brion 23:55, September 12, 2005 (UTC)

• I distinctively remember visiting their main page ages ago. Was it located on a different address? - Mgm|^(talk) 10:43, 14 September 2005 (UTC)

  Until very recently there was only a multilingual Wikisource (wikisource.org) and a Hebrew edition (he.wikisource.org). It's recently been split into multiple language subdomains, with pages in various languages copied out to the subdomains. See eg [1] --Brion 11:08, 14 September 2005 (UTC)

ERROR: The URL could not be retrieved

Why do I always get this error when I save my changes? It unfortunately stops any changes from going through most of the time, it's frustrating. Anybody else get this error? --Revolución (talk) 23:37, 12 September 2005 (UTC)

I just got this error while posting this! --Revolución (talk) 23:47, 12 September 2005 (UTC)

Hopefully you shouldn't be getting this anymore. I'll be replacing the search update system with one that doesn't hang when it breaks. --Brion 00:03, September 13, 2005 (UTC)

I know, it's terrible! Please get it fixed! --Mb1000 19:35, 17 September 2005 (UTC)

redirect breaking out of frames

I wrote a utility for my own personal use that makes use of frames. In the top frame it allows me to save links, and in the bottom main frame, it loads a random or specific link that was saved earlier. The interface is similar to that of google images.

When any wikipedia page is loaded within the frame though, it appears there is some sort of redirect which causes it to break out of the frame. The culprit appears to be the code on line 16:

<script type="text/javascript" src="/w/index.php?title=-&action=raw&smaxage=0&gen=js"></script>

I don't know if this is an intentional attempt to keep people from viewing Wikipedia from inside of external frames, or if the breaking out is just a side effect of some other purpose. Whatever the case may be, I would like to have the option as the end user, to choose how I want to view the page. I was hoping that perhaps there was a preference I could set, or a flag that would prevent it from redirecting.

Any help would be greatly appreciated.

I would like to second the above. The "line 16" you are referring to is actually served to us from http://en.wikipedia.org/skins-1.5/common/wikibits.js and the problem bit is this:

// Un-trap us from framesets
if( window.top != window ) window.top.location = window.location;

I feel that it should be changed to something like the following:

if ( ( window.top.protocol != 'file:' ) &&
     ( window.top.location.hostname != window.location.hostname )
       window.top.location = window.location;

This would allow someone to create their own script-based facilities on their own computer such as the above poster would like, and would monobook.js users to do things like create dynamic iframes that could load 2 or more pages into one window; as an example: loading Special:Ipblocklist into an iframe on Special:Blockip. Func( t, c, @, ) 06:31, 15 September 2005 (UTC)

Arabic/Hebrew: a proposed solution to Unicode bidirectional algorithm woes in the text editor

In a nutshell, Latin and Cyrillic and Greek etc. letters are strongly directional (left-to-right) and Arabic and Hebrew letters are strongly directional (right-to-left), but punctuation is neutral. See [2] [3]

A problem arises if punctuation is sandwiched between two strongly-directional characters of opposite directionality... it doesn't know which one to attach to.

A common example is an Arabic or Hebrew interwiki link with disambiguation parentheses, of the form:

(xxx) xxxxxxx

Pretend the x's are Arabic or Hebrew letters
and they're read ← right-to-left.

So the part in (  ) comes last and represents a disambiguation.
The final character is the closing parenthesis at extreme left
(it's a "mirrored" closing parenthesis that looks like an open parenthesis)

However, in the text editor, this will display as:

[[ar: xxx) xxxxxxxx) ]]

That's because the closing parenthesis of the interwiki link text is sandwiched between the xxx right-to-left Arabic text and the "left-to-rightedness" of the remainder of the document, and the browser text editor decides that the closing parenthesis belongs to and should attach to the latter rather than the former (applying the hideously complicated Unicode bidirectional algorithm.

One workaround is to use U+200E left-to-right mark and U+200F right-to-left mark, but these are zero-width invisible characters, so you can accidentally delete them in the course of editing and never even notice. Also, you can't use these within a wikilink because they're Unicode characters in their own right and adding them will turn a bluelink red.

A visible way to create a left-to-right mark is just to embed a Latin letter within an HTML comment: <!--B-->, and similarly a right-to-left mark can be simulated by embedding an Arabic or Hebrew letter within an HTML comment. An added benefit is that these can be used within wikilinks too.

The interwiki link problem described earlier can be solved if we do:

[[ar: <!--LA--> xxxxxxxxxxxx <!--AL--> ]]

here, L is a single Latin letter,
A is a single Arabic (or Hebrew) letter,
and xxxxxxxxxxxx is Arabic or Hebrew text.

The key thing is, the Arabic or Hebrew text will always display correctly, even if it begins or ends with an apostrophe or parentheses or other punctuation or directionally-neutral characters.

For examples see:

fa: interwiki link: before after (closing parenthesis placement)

he: interwiki link: before after (apostrophe placement)

Note that due to bidirectional algorith display issues, it looks as if the HTML comments are wrong. It looks as if it's:

<!--LA--!> xxxxxxxxxxxx <--LA-->

But that's only a display issue... in fact, all the bytes are in the right order and both HTML comments are well-formed. The important thing is that the xxxxxxxxxxxx Arabic or Hebrew (or Farsi, etc) text in the middle displays correctly.

Using this workaround, Arabic or Hebrew speaking editors can get the benefits of editing Hebrew or Arabic interwiki links as text (rather than &# gobbledygook), but don't run into annoying bidirectional display issues.

-- Curps 11:54, 13 September 2005 (UTC)

As an aside, would the following work? Say we coded the "save page" logic in mediawiki to search for runs of (2 or more, I guess) characters in the hebrew and arabic ranges. It would then make sure these runs are properly embraced between U+200E and U+200F marks before commiting the page to the database. So even if someone does erase the marks, they're automatically reinstated. It strikes me that R-L characters in L-R wikis (and vice versa, I guess) are common and important enough to justify a code improvement. -- Finlay McWalter | Talk 12:28, 13 September 2005 (UTC)

I think there are a couple of problems with trying to make it completely tranparent.

The first problem is, if the marks get erased, the display in the browser edit can get scrambled (the display only; the underlying byte stream does not get scrambled). An example is the following paragraph (from User:Grenavitar):

Abu 'Abd Allah ash-Shafi'i (Arabic: إمام الشافعى ) (767 [150 AH] - January 20, 820 [204 AH]) commonly called Imam Shafi'i (Arabic: ابو عبد الله الشافعى ) or fully, Muhammad ibn Idris ibn al ‘Abbas ibn ‘Uthman ibn Shafi’i ibn al Sa’ib ibn ‘Ubayd ibn Abd Yazid ibn Hashim ibn al Muttalib ibn ‘Abd Manaf.

If you edit the above paragraph and delete the "c" within <!-- c -->, the text will suddenly "jump"... the "767" will pop to the other side. Type it back in and it returns to the previous appearance. Repeatedly typing and deleting "c" toggles it. Now imagine that instead of the visible <!-- c --> we have an invisible zero-width left-to-right mark or right-to-left mark, and you unknowingly delete it.

A more serious problem is that sometimes you want punctuation to attach to the Arabic text. In the example I gave above, you want the closing parenthesis in the interwiki link to attach to the Arabic text, and it would be a mistake to automatically place LTR and RTL marks only around the Arabic lette characters. In general there's no way to do this automatically and get it right, without a human understanding of the author's intentions... if there was, the Unicode bidirectional algorithm would get it right in the first place.

Another problem is that you can't use literal LTR and RTL marks within a wikilink... they'll be interpreted as Unicode characters and instead of a bluelink like Wikipedia, you'll get a &lrm;&rlm;Wikipedia&rlm;&lrm; redlink. So this won't work for interwiki links. HTML comments don't have that problem.

It should be emphasized by the way that the above problems are relatively rare. The great majority of cases where Arabic or Hebrew text is embedded within an English article pose no problem at all, and the small minority that do can be worked around using HTML comments as described above, in a visually explicit way. Probably we should write a Wikipedia: namespace article about these bidirectional issues and the workarounds.

-- Curps 12:44, 13 September 2005 (UTC)

I tried to use the comment method you described in the Abu Hanifa an-Nu'man article. Maybe I was using it incorrectly, or expected it to behave differently, but the text in the edit box still got messed up. The browser I'm using is Firefox 1.0.6 on Windows. Could this be a problem? I ended up using the L-T-R mark which, troublesome as it may be, seems to be the only thing that works in my environment. --Yodakii 18:21, 15 September 2005 (UTC)

Extreme server slowness

I've recently run into severe delays in accessing Wikipedia. Average page load times for me have now risen to ~14 seconds per page. That's just to load a page. To edit a page or save changes can take as long as several minutes. This has become completely disruptive of my editing. I was running a high average of edits per day, but that average has dropped significantly over the last several days due to this severe lag. I'm only doing bare minimums now; closing out work I had in progress, etc. I've found some potential answers here, but that is dated from January of this year; 8 months ago. I'm presuming it's not terribly relevant now.

I did a traceroute to en.wikipedia.org to see if there's a delay in the hops somewhere. I had the following results (ran three times, all with similar results):

traceroute: Warning: rr.pmtpa.wikimedia.org has multiple addresses; using 207.142.131.214
traceroute: Warning: Multiple interfaces found; using 129.79.6.186 @ dmfe0
traceroute to rr.pmtpa.wikimedia.org (207.142.131.214), 30 hops max, 40 byte packets
 1  default-gw5.ucs.indiana.edu (129.79.5.254)  1.193 ms  0.891 ms  0.339 ms
 2  ibr-ibs1-pb.noc.iu.edu (149.166.2.52)  1.489 ms  1.505 ms  1.501 ms
 3  149.165.254.229 (149.165.254.229)  3.111 ms  1.504 ms  1.606 ms
 4  192.12.206.241 (192.12.206.241)  1.533 ms  2.240 ms  1.602 ms
 5  so-2-3-0-0.gar2.Chicago1.Level3.net (67.72.124.9)  6.019 ms  6.102 ms  6.036 ms
 6  ae-1-56.bbr2.Chicago1.Level3.net (4.68.101.161)  6.078 ms ae-1-54.bbr2.Chicago1.Level3.net (4.68.101.97)  6.033 ms  5.980 ms
 7  as-0-0.mp2.Tampa1.Level3.net (209.247.11.198)  42.252 ms  42.034 ms  41.948 ms
 8  ae-23-54.car3.Tampa1.Level3.net (4.68.104.107)  42.265 ms ae-13-55.car3.Tampa1.Level3.net (4.68.104.139)  42.135 ms ae-23-52.car3.Tampa1.Level3.net (4.68.104.43)  42.070 ms
 9  4.71.0.10 (4.71.0.10)  42.225 ms  42.094 ms  42.262 ms
10  84.40.24.26 (84.40.24.26)  42.100 ms  42.449 ms  42.147 ms
11  64.156.25.242 (64.156.25.242)  42.446 ms  42.453 ms  48.852 ms
12  * * *
13  * * *
14  * * *
15  * * *
16  * * *
17  * * *
18  * * *
19  * * *
20  * * *
21  * * *
22  * * *
23  * * *
24  * * *
25  * * *
26  * * *
27  * * *
28  * * *
29  * * *
30  * * *

Is there any explanation for the severe lag I am receiving? Is there any prognosis for this to ever recover or should I just hang up being a significant editor of Wikipedia? --Durin 17:53, 13 September 2005 (UTC)

I hope someone investigates this: the Dutch Wikipedia is almost unreachable, and I'm getting seriously annoyed: it takes up to 5 minutes to load a page! --Tuvic 18:55, 13 September 2005 (UTC)

The slowness is probably within the servers. Wikipedia tends to have a cyclic nature: new hardware make it faster, then more people start editing and it gets progressively slower, until more new hardware is added. So, it will recover when more hardware is added and brought online. --cesarb 19:13, 13 September 2005 (UTC)

I know this is probably a bit far fetched; but is it possible that the new firefox (1.5 beta) will reduce server load, as it now caches pages so that when you use the back and foward buttons it doesnt re-load the page, saving time and bandwidth, especially as firefox is quite popular with wikipedia editors. Probably just me being over-optimistic, but i wonder what effect it will have. Martin 23:34, 13 September 2005 (UTC)

Firefox already caches pages from Wikipedia very well. The fastback feature caches the rendering of the page, not the actual data (which is already cached). It's a very cool feature though. — Ambush Commander^(Talk) 23:44, 13 September 2005 (UTC)

I expect the problems are related to one of the main database servers crashing a few days ago (see the server admin log) and the new Yahoo servers being set up over the last few days. Angela. 09:07, 14 September 2005 (UTC)

Our primary database server had a disk fail a couple days ago, so we're limping along on fewer resources. (It's being reinstalled as I type, and should be back in service within a day, if nothing else goes drastically wrong.) --Brion 09:21, 14 September 2005 (UTC)

• I'm wondering if something else went drastically wrong. I know it's not been a day, but pages are now taking minutes to load. I just tried to load an admin nominee's voting page...not to edit, just view. It took 3 minutes and 17 seconds to finally load. It is virtually impossible to work with such delays. I know this is a volunteer organization operating on volunteered/donated resources...but certainly there's got to be a structure in place to be a backup for the primary database server losing a single disk? I suspect there's more wrong than just 1 disk in a (what I imagine to be) RAID farm with probably hot swappable drives. Can you let us know when this disk has been replaced and re-striped? I'll update here when I hear of that, and let you know if the delays are still large. --Durin 17:24, 14 September 2005 (UTC)

Not to strike a carping note, but considering that Wikipedia is (and quite loudly prides itself on being) one of the biggest and best sites on the net, one would expect it to provide faster service for its users, at least as fast as that provided by any of the top 10-20 sites. --Peripatetic 19:49, 14 September 2005 (UTC)

I came here to post the same question and while I was surfing around things got lightning fast: I have never seen Wiki so fast! I hope Yahoo! servers started working :) Renata3 03:28, 16 September 2005 (UTC)

My interpretation of "lightning fast" (LF) is that the (LF) requests had finally been cached. But the "extreme slowness" (XS) is the overloaded servers still had not completed processing (XS) requests. If I understand Brion's comment, when the crashed disk has fully recovered, we should be back to normal response times, even with the heavier load due to being in the 30s for Alex rankings. Ancheta Wis 07:26, 16 September 2005 (UTC)

It has been slower before, but now that the Yahoo servers have been added, and the traffic is up, I am experiencing the slow times cited above (15 sec just to see the page to edit this section); came to VP to note the slowness. Am currently not experiencing lightning fast response. Thank you everyone, Ancheta Wis 07:14, 16 September 2005 (UTC)

The situation is now sometimes better, and sometimes worse. Just before coming here, it took several minutes to load an AfD page just to view, much less edit. Requesting to edit this page too ~10 seconds. Not too bad. But, the speed this took to load is in the minority. My traceroute results are still the same; the delay is at Wikipedia. Is there any prognosis for when the speed problem is going to be resolved? When observing some increases, I began to work on pages again...and ran into huge delay problems again. It got so bad with opening of ~20 tabs that I lost track of what I was doing. The lack of free flow movement now makes it impossible for me to do any signficant editing. One page at a time, and wait for minutes to respond. I can't work this way. I suspect many others can not as well. --Durin 17:33, 16 September 2005 (UTC)

The situation is so bad and never seems to improve for very long, I think Wikipedia has become a waste of my time and money. Every effort should now be going into making the technical side of WP usable, forget everything else or the whole project will fail Bob Palin 00:31, 17 September 2005 (UTC)

The trouble is that whenever wikipedia is fast its userbase grows whenever wikipedias userbase grows it gets slow again (both through resource exaustion and issues that only become issues as the system scales up). I strongly suspect a long period of slowness will stunt growth at worst and not result in any significan't downterm. Plugwash 01:10, 17 September 2005 (UTC)

• At one point I was editing over 25 pages a day, now I'm down to about 5 or so a day (sometimes even less) becuase of this EXTREME slowness. This needs to get fixed, and fast! --Mb1000 19:31, 17 September 2005 (UTC)

• And just an hour ago, the Wiki was completely dead. I would try to access the main page through http://en.wikipedia.org/ and it would send me immediately to the "Wikimedia servers have a problem" page. Does anyone know what's going on, and why? Titoxd 02:13, 19 September 2005 (UTC)

  A tech accidentally rebooted our primary file server during maintenance. We're still more dependent on it than we should be, and there seems to be a problem with it coming back online very slowly. (The wikis' contents are not on that server; it carries configuration files, logs, utility programs, etc and is the main entry point and operations for working in our Florida cluster. With better configuration our web servers should be able to keep running when it's offline, but currently they require access to some files on it and everything hangs if it goes down unexpectedly. See wikitech-l recent post.) --Brion 09:25, 19 September 2005 (UTC)

It was previously stated (around January) that one of the key bottlenecks was a problem with a poorly configured load balancing server. Was this ever resolved, or is this still one of the sources of slowness? If it is, what can we as a community do to recruit the appropriate human/hardware resources to fix it? --Dhartung | Talk 03:16, 19 September 2005 (UTC)

We use perlbal for HTTP load balancing (between the proxies and the web servers). Under some failure conditions it explodes in horrible ways, so it's not perfect yet. If you know a good replacement, or can help fix issues with it ;) that would be of course super. --Brion 09:25, 19 September 2005 (UTC)

Correcting Misspellings (One-word Edits)

Often when I'm reading through an article I will find the odd misspelling or two, but I'm too lazy to click the Edit button, change the text and save it, just to correct one word, so the typo stays there until someone more motivated than me comes along.

Now, in DOM compliant browsers, it should be possible to make it so that, e.g., when you double-click on a word, it puts that word in an edit box and lets you edit just that one word which then gets saved when you press ENTER, transparently to the user. I got the idea from the way Flickr lets you change the title, description, etc., of your own photos.--81.42.154.112 19:36, 13 September 2005 (UTC)

If it can be done, that seems like a great idea. -- NatsukiGirl\^talk 19:54, 13 September 2005 (UTC)

Sounds like a wonderful idea! I'm all in favour! Grutness...wha? 01:08, 14 September 2005 (UTC)

It'll need some MediaWiki hooks to work though... how about posting a feature request at bugzilla and posing it to wikitech-l? — Ambush Commander^(Talk) 01:15, 14 September 2005 (UTC)

The potential for abuse of this would be real, but: if we could fix typos this way this would be all that and a bag of chips. I'd love it. Nandesuka 04:19, 16 September 2005 (UTC)

I dislike this idea. I often will highlight and double click words as I'm reading as a guide (to keep my place and as a highlighter for key words). I'd suggest a highlight plus a key combo or a middle click (mouse wheel). This link is Broken 00:12, 19 September 2005 (UTC)

OP here. I agree with the above. A user preference and/or toggle option should be added so that when activated double-click causes an edit, and when unactivated double-click events are unbound (i.e., act as normal). Please note that when using a toggle type option (accesskey and/or clickable link on the page) no reload needs to take place, as the required behaviour (bind/unbind an event handler) can be taken care of entirely by client-side scripting.

Of course, non-ECMAScript enabled browsers should not see this toggle option as they wouldn't be implementing one-word edits anyway.--81.42.154.225 13:28, 20 September 2005 (UTC)

Balearic_Islands entry is goofed up?

Hey -

when I try to look at http://en.wikipedia.org/wiki/Balearic_Islands, I get a download prompt (rather than an html page that displays). Downloading the file (7.06 KB) is unenlightening; it appears to me as binary gibberish.

Am I doing something wrong? Has the page been vandalized? If so I'd like to revert it but I can't even access it; maybe someone with more technical knowledge or greater access could take a look?

FWIW I am running IE 6.0 and have no problem viewing any other wikipedia page I know of.

Thanks!

• Did you accidentally click on "Save this page as" or a similar option in your right-click menu for the link. It appears perfectly normal to me. - Mgm|^(talk) 10:51, 14 September 2005 (UTC)

garbage pages

hello. there was a brief problem which may have produced some garbage pages (random binary data). this should be fixed now but bad data may be left in the cache. sorry. kate.

Unicode chart templates

I have generated a complete series of Unicode charts for each code range. The names are {{Unicode chart <name>}}, where <name> is the official name used by Unicode.org. The only exceptions are surrogate and private use areas and specials, for which there are no glyphs defined.

These templates are used, for example, at Wikipedia:Naming conventions (Unicode) (draft) to show which Unicode characters are displayable using the default font and which ones are displayable using "Unicode fonts" (as defined by Template:Unicode fonts).

Unfortunately, a template can't take the value of another template as an argument. Thus to display a template using a particular font, you should use HTML to set the font in a <div> or table element <td> surrounding the template, as shown below:

Default font		Unicode font
Latin Extended Additional[1] Official Unicode Consortium code chart (PDF)   0 1 2 3 4 5 6 7 8 9 A B C D E F U+1E0x Ḁ ḁ Ḃ ḃ Ḅ ḅ Ḇ ḇ Ḉ ḉ Ḋ ḋ Ḍ ḍ Ḏ ḏ U+1E1x Ḑ ḑ Ḓ ḓ Ḕ ḕ Ḗ ḗ Ḙ ḙ Ḛ ḛ Ḝ ḝ Ḟ ḟ U+1E2x Ḡ ḡ Ḣ ḣ Ḥ ḥ Ḧ ḧ Ḩ ḩ Ḫ ḫ Ḭ ḭ Ḯ ḯ U+1E3x Ḱ ḱ Ḳ ḳ Ḵ ḵ Ḷ ḷ Ḹ ḹ Ḻ ḻ Ḽ ḽ Ḿ ḿ U+1E4x Ṁ ṁ Ṃ ṃ Ṅ ṅ Ṇ ṇ Ṉ ṉ Ṋ ṋ Ṍ ṍ Ṏ ṏ U+1E5x Ṑ ṑ Ṓ ṓ Ṕ ṕ Ṗ ṗ Ṙ ṙ Ṛ ṛ Ṝ ṝ Ṟ ṟ U+1E6x Ṡ ṡ Ṣ ṣ Ṥ ṥ Ṧ ṧ Ṩ ṩ Ṫ ṫ Ṭ ṭ Ṯ ṯ U+1E7x Ṱ ṱ Ṳ ṳ Ṵ ṵ Ṷ ṷ Ṹ ṹ Ṻ ṻ Ṽ ṽ Ṿ ṿ U+1E8x Ẁ ẁ Ẃ ẃ Ẅ ẅ Ẇ ẇ Ẉ ẉ Ẋ ẋ Ẍ ẍ Ẏ ẏ U+1E9x Ẑ ẑ Ẓ ẓ Ẕ ẕ ẖ ẗ ẘ ẙ ẚ ẛ ẜ ẝ ẞ ẟ U+1EAx Ạ ạ Ả ả Ấ ấ Ầ ầ Ẩ ẩ Ẫ ẫ Ậ ậ Ắ ắ U+1EBx Ằ ằ Ẳ ẳ Ẵ ẵ Ặ ặ Ẹ ẹ Ẻ ẻ Ẽ ẽ Ế ế U+1ECx Ề ề Ể ể Ễ ễ Ệ ệ Ỉ ỉ Ị ị Ọ ọ Ỏ ỏ U+1EDx Ố ố Ồ ồ Ổ ổ Ỗ ỗ Ộ ộ Ớ ớ Ờ ờ Ở ở U+1EEx Ỡ ỡ Ợ ợ Ụ ụ Ủ ủ Ứ ứ Ừ ừ Ử ử Ữ ữ U+1EFx Ự ự Ỳ ỳ Ỵ ỵ Ỷ ỷ Ỹ ỹ Ỻ ỻ Ỽ ỽ Ỿ ỿ Notes 1.^ As of Unicode version 15.1		Latin Extended Additional[1] Official Unicode Consortium code chart (PDF)   0 1 2 3 4 5 6 7 8 9 A B C D E F U+1E0x Ḁ ḁ Ḃ ḃ Ḅ ḅ Ḇ ḇ Ḉ ḉ Ḋ ḋ Ḍ ḍ Ḏ ḏ U+1E1x Ḑ ḑ Ḓ ḓ Ḕ ḕ Ḗ ḗ Ḙ ḙ Ḛ ḛ Ḝ ḝ Ḟ ḟ U+1E2x Ḡ ḡ Ḣ ḣ Ḥ ḥ Ḧ ḧ Ḩ ḩ Ḫ ḫ Ḭ ḭ Ḯ ḯ U+1E3x Ḱ ḱ Ḳ ḳ Ḵ ḵ Ḷ ḷ Ḹ ḹ Ḻ ḻ Ḽ ḽ Ḿ ḿ U+1E4x Ṁ ṁ Ṃ ṃ Ṅ ṅ Ṇ ṇ Ṉ ṉ Ṋ ṋ Ṍ ṍ Ṏ ṏ U+1E5x Ṑ ṑ Ṓ ṓ Ṕ ṕ Ṗ ṗ Ṙ ṙ Ṛ ṛ Ṝ ṝ Ṟ ṟ U+1E6x Ṡ ṡ Ṣ ṣ Ṥ ṥ Ṧ ṧ Ṩ ṩ Ṫ ṫ Ṭ ṭ Ṯ ṯ U+1E7x Ṱ ṱ Ṳ ṳ Ṵ ṵ Ṷ ṷ Ṹ ṹ Ṻ ṻ Ṽ ṽ Ṿ ṿ U+1E8x Ẁ ẁ Ẃ ẃ Ẅ ẅ Ẇ ẇ Ẉ ẉ Ẋ ẋ Ẍ ẍ Ẏ ẏ U+1E9x Ẑ ẑ Ẓ ẓ Ẕ ẕ ẖ ẗ ẘ ẙ ẚ ẛ ẜ ẝ ẞ ẟ U+1EAx Ạ ạ Ả ả Ấ ấ Ầ ầ Ẩ ẩ Ẫ ẫ Ậ ậ Ắ ắ U+1EBx Ằ ằ Ẳ ẳ Ẵ ẵ Ặ ặ Ẹ ẹ Ẻ ẻ Ẽ ẽ Ế ế U+1ECx Ề ề Ể ể Ễ ễ Ệ ệ Ỉ ỉ Ị ị Ọ ọ Ỏ ỏ U+1EDx Ố ố Ồ ồ Ổ ổ Ỗ ỗ Ộ ộ Ớ ớ Ờ ờ Ở ở U+1EEx Ỡ ỡ Ợ ợ Ụ ụ Ủ ủ Ứ ứ Ừ ừ Ử ử Ữ ữ U+1EFx Ự ự Ỳ ỳ Ỵ ỵ Ỷ ỷ Ỹ ỹ Ỻ ỻ Ỽ ỽ Ỿ ỿ Notes 1.^ As of Unicode version 15.1

The full set is listed below:

• {{Unicode chart Alphabetic Presentation Forms}}
• {{Unicode chart Arabic}}
• {{Unicode chart Arabic Presentation Forms-A}}
• {{Unicode chart Arabic Presentation Forms-B}}
• {{Unicode chart Arabic Supplement}}
• {{Unicode chart Armenian}}
• {{Unicode chart Arrows}}
• {{Unicode chart Bengali}}
• {{Unicode chart Block Elements}}
• {{Unicode chart Bopomofo}}
• {{Unicode chart Bopomofo Extended}}
• {{Unicode chart Box Drawing}}
• {{Unicode chart Braille Patterns}}
• {{Unicode chart Buginese}}
• {{Unicode chart Buhid}}
• {{Unicode chart C0 Controls and Basic Latin}}
• {{Unicode chart C1 Controls and Latin-1 Supplement}}
• {{Unicode chart CJK Compatibility}}
• {{Unicode chart CJK Compatibility Forms}}
• {{Unicode chart CJK Compatibility Ideographs}}
• {{Unicode chart CJK Radicals Supplement}}
• {{Unicode chart CJK Strokes}}
• {{Unicode chart CJK Symbols and Punctuation}}
• {{Unicode chart CJK Unified Ideographs}}
• {{Unicode chart CJK Unified Ideographs Extension A}}
• {{Unicode chart Cherokee}}
• {{Unicode chart Combining Diacritical Marks}}
• {{Unicode chart Combining Diacritical Marks Supplement}}
• {{Unicode chart Combining Diacritical Marks for Symbols}}
• {{Unicode chart Combining Half Marks}}
• {{Unicode chart Control Pictures}}
• {{Unicode chart Coptic}}
• {{Unicode chart Currency Symbols}}
• {{Unicode chart Cyrillic}}
• {{Unicode chart Cyrillic Supplement}}
• {{Unicode chart Devanagari}}
• {{Unicode chart Dingbats}}
• {{Unicode chart Enclosed Alphanumerics}}
• {{Unicode chart Enclosed CJK Letters and Months}}
• {{Unicode chart Ethiopic}}
• {{Unicode chart Ethiopic Extended}}
• {{Unicode chart Ethiopic Supplement}}
• {{Unicode chart General Punctuation}}
• {{Unicode chart Geometric Shapes}}
• {{Unicode chart Georgian}}
• {{Unicode chart Georgian Supplement}}
• {{Unicode chart Glagolitic}}
• {{Unicode chart Greek Extended}}
• {{Unicode chart Greek and Coptic}}
• {{Unicode chart Gujarati}}
• {{Unicode chart Gurmukhi}}
• {{Unicode chart Halfwidth and Fullwidth Forms}}
• {{Unicode chart Hangul Compatibility Jamo}}
• {{Unicode chart Hangul Jamo}}
• {{Unicode chart Hangul Syllables}}
• {{Unicode chart Hanunoo}}
• {{Unicode chart Hebrew}}
• {{Unicode chart High Surrogate Area}}
• {{Unicode chart Hiragana}}
• {{Unicode chart IPA Extensions}}
• {{Unicode chart Ideographic Description Characters}}
• {{Unicode chart Kanbun}}
• {{Unicode chart Kangxi Radicals}}
• {{Unicode chart Kannada}}
• {{Unicode chart Katakana}}
• {{Unicode chart Katakana Phonetic Extensions}}
• {{Unicode chart Khmer}}
• {{Unicode chart Khmer Symbols}}
• {{Unicode chart Lao}}
• {{Unicode chart Latin Extended Additional}}
• {{Unicode chart Latin Extended-A}}
• {{Unicode chart Latin Extended-B}}
• {{Unicode chart Letterlike Symbols}}
• {{Unicode chart Limbu}}
• {{Unicode chart Low Surrogate Area}}
• {{Unicode chart Malayalam}}
• {{Unicode chart Mathematical Operators}}
• {{Unicode chart Miscellaneous Mathematical Symbols-A}}
• {{Unicode chart Miscellaneous Mathematical Symbols-B}}
• {{Unicode chart Miscellaneous Symbols}}
• {{Unicode chart Miscellaneous Symbols and Arrows}}
• {{Unicode chart Miscellaneous Technical}}
• {{Unicode chart Modifier Tone Letters}}
• {{Unicode chart Mongolian}}
• {{Unicode chart Myanmar}}
• {{Unicode chart New Tai Lue}}
• {{Unicode chart Number Forms}}
• {{Unicode chart Ogham}}
• {{Unicode chart Optical Character Recognition}}
• {{Unicode chart Oriya}}
• {{Unicode chart Phonetic Extensions}}
• {{Unicode chart Phonetic Extensions Supplement}}
• {{Unicode chart Private Use Area}}
• {{Unicode chart Runic}}
• {{Unicode chart Sinhala}}
• {{Unicode chart Small Form Variants}}
• {{Unicode chart Spacing Modifier Letters}}
• {{Unicode chart Specials}}
• {{Unicode chart Superscripts and Subscripts}}
• {{Unicode chart Supplemental Arrows-A}}
• {{Unicode chart Supplemental Arrows-B}}
• {{Unicode chart Supplemental Mathematical Operators}}
• {{Unicode chart Supplemental Punctuation}}
• {{Unicode chart Syloti Nagri}}
• {{Unicode chart Syriac}}
• {{Unicode chart Tagalog}}
• {{Unicode chart Tagbanwa}}
• {{Unicode chart Tai Le}}
• {{Unicode chart Tamil}}
• {{Unicode chart Telugu}}
• {{Unicode chart Thaana}}
• {{Unicode chart Thai}}
• {{Unicode chart Tibetan}}
• {{Unicode chart Tifinagh}}
• {{Unicode chart Unified Canadian Aboriginal Syllabics}}
• {{Unicode chart Variation Selectors}}
• {{Unicode chart Vertical Forms}}
• {{Unicode chart Yi Radicals}}
• {{Unicode chart Yi Syllables}}
• {{Unicode chart Yijing Hexagram Symbols}}

-- Curps 00:09, 14 September 2005 (UTC)

"What links here" ambiguity

Yesterday Producer had over 500 links leading to it. A couple of hours a go I did a "What links here" for Producer and got about 100 links. This corresponded with the a dump generated count of 101 that I saw here. However now, when I check, there are over 500 links. What is causing the discrepancy? --Commander Keane 12:45, 14 September 2005 (UTC)

Slow Going?

Is it me or is the server running very slowly lately? --Doug O'Connell 14:45, 14 September 2005 (UTC)

No it's not. See discussion above at Extreme server slowness -- Ian ≡ talk 14:55, 14 September 2005 (UTC)

Wiki access via PDA?

Hello, Would this be a place to ask about what developer options are available for building solutions to accessing Wikipedia via a handheld PDA. Options like XML (XML-RPC)?

Thanks, Eric

User:Erik Zachte has done some work on Wikipedia for PDA's. His user page has some useful links. --Commander Keane 05:43, 15 September 2005 (UTC)

Linking to an article version as of a particular date

I note that Wikipedia assigns a unique ID to each revision of an article. The URL to a particular revision is http://en.wikipedia.org/w/index.php followed by a query string with the article's title and revision ID. Example: http://en.wikipedia.org/w/index.php?title=Aristotle&oldid=22376284. This gives me the article Aristotle as at 05:00, 2 September 2005.

Is there any similar way of linking to an article as it appeared at a certain date and time, given only the article title and the date/time? For example, is there some query string like http://en.wikipedia.org/w/index.php?title=Aristotle&revdate=2005-09-02-0530, which will resolve to the same link as above? (Since Aristotle was not modified between 05:00 and 05:30 on 2 September 2005, the link should resolve to revision 22376284.) —Psychonaut 00:18, 15 September 2005 (UTC)

No, I don't believe so. — Knowledge Seeker দ 05:52, 15 September 2005 (UTC)

I recently posted a MediaWiki bug/wishlist report for this feature. Someone responded and apparently, it's already semi-possible to do this with the following URL syntax: http://en.wikipedia.org/w/index.php?title=Aristotle&offset=20050902053000&action=history. The first link returned is the right one; a simple page scraper can extract this link. —Psychonaut 18:50, 17 September 2005 (UTC)

Editing on a PC?

Is it possible to use Mediawiki on a PC for off-line editing? I am involved with Wikisource, editing for the 1911 Encyclopaedia Britannica project. Tbis involves some heavy-duty coding to get accents and foreign characters included, and it would be useful to do all this on my PC before uploading it. This save me the hassle of dialup costs and also delays when Wikisource is slow to respond. If this is not possible, can anyone suggest a suitable download. What I am after is to be able to view the results of my work before uploading in case I made a mistake. Thanks Apwoolrich 10:30, 15 September 2005 (UTC)

It's straightforward to run a personal mediawiki installation on your PC. You can make your edits there and later cut'n'paste into live wikisource edit windows when you're online. -- Finlay McWalter | Talk 14:05, 15 September 2005 (UTC)

New contributions

How do you view a users contributions, and filter the list so you only see what New pages they've created? - Xed 15:31, 15 September 2005 (UTC)

• Special:Contributions/Xed - simply change the username as appropriate. However, I don't believe there is any way to filter out new pages; sorry... Shimgray 16:13, 15 September 2005 (UTC)

linking/accessing articles from software applications

Hey guys,

I am a software developer currently working on a new project. I'm writing an index card application that is to help students learn and prepare exams with the help of the computer.

From my own experience I know that while studying it can be very handy to access reference works quickly to look up technical words or refer to a lexicon entry. For that matter, I'd like to implement a function that allows the user to click a word in a textbox that they want to look up in wikipedia. So I get the keyword from the user, but what wikipedia link do I have to use to call up any article available in your free encyclopedia? I want to hand the keyword over as a link parameter in the browser's address bar. It might be a good idea if wikipedia could handle those keywords that are not listed yet and offer a list of similar or related keywords. I need to look up lexical entries in English and German.

I'd appreciate if you could give me some information on how to implement this function.

Best regards, Sebastian Felling

That's pretty simple; just fill in http://en.wikipedia.org/wiki/ with the search string appended (with spaces converted to underscores, and other non-alphanumeric characters encoded under UTF-8 using percent signs and hex digits). (de.wikipedia.org for the German version, and you can also do something similar for Wiktionary if you want dictionary definitions.) *Dan T.* 12:12, 16 September 2005 (UTC)

Logo disappearance

Some time ago the Wikipedia logo in the upper-left corner disappeared... it used to be there just fine, but now it's gone. Has anyone changed (css/html?) code related to the logo? A few tests show that IE seems to have no problem displaying .png's with the background-image attribute, so that can't be it. Any other idea's on what causes this? Are my settings wrong, or has anyone changed some code? Is the logo affected by script in some way that may cause it not to display? Shinobu 14:48, 16 September 2005 (UTC)

It seems fine for me (in Mozilla Application Suite). Maybe you're running some kind of proxy filter that blocks images it thinks are ads or something? *Dan T.* 00:04, 17 September 2005 (UTC)

Try clearing your browser cache. --Brion 21:29, 17 September 2005 (UTC)

It seems to be caused by the IE 5.5 PNG alpha fix. Hm...

When I do the following, the logo reappears (modulo a layout problem; this single line is just a test and I don't pretend it undoes the fix fully):

javascript:alert(document.body.children(0).children(1).children(2).style.backgroundImage="url(http://upload.wikimedia.org/wikipedia/en/b/bc/Wiki.png)");

The fix uses:

FILTER: progid:DXImageTransform.Microsoft.AlphaImageLoader(src=http://upload.wikimedia.org/wikipedia/en/b/bc/Wiki.png);

and that doesn't seem to work. Shinobu 02:48, 21 September 2005 (UTC)

unbearable lag

Has anyone been experiencing severe lag lately? Sometimes, it takes several minutes for a single page to load, and it's extremely frustrating. --Ixfd64 23:26, 16 September 2005 (UTC)

YES. YES. ~~ N (t/c) 23:44, 16 September 2005 (UTC)

See above. --cesarb 01:23, 17 September 2005 (UTC)

External links in the status bar

Not sure why it's been changed, but external link URLs are no longer shown in the status bar (at least on IE). We now have to wait until the tooltip appears in order to see the link, which takes too long for us impatient types (and those trying to check for improper links). Anything that can be done to fix this? violet/riga (t) 08:09, 17 September 2005 (UTC)

Yeah. Switch to Firefox. OK sorry, I know that was very unhelpful. It still works in Firefox though; I'm not sure what's going on. — Knowledge Seeker দ 13:35, 17 September 2005 (UTC)

If only. They refuse to install Firefox on the computers as it's "non-standard and too much work". violet/riga (t) 17:59, 18 September 2005 (UTC)

I have the same problem at work, however I installed Firefox anyway (and the user agent spoofer so the proxy doesn't complain about a lack of NTLM authorisation). Two different techs have seen it on my machine and not commented, despite us not being allowed to install any software on the machines. One of the other people in the office (who is at the same level as my boss) even commented that he wasn't aware it worked - so I explained about the user agent spoofer. Whether he installed that or not I don't know. I still have to use IE for some pages on the intranet that just don't work at all with firefox, though. Thryduulf 18:26, 18 September 2005 (UTC)

Works for me (IE 6.0 on XP SP2) --Brion 21:28, 17 September 2005 (UTC)

IE 6.0 on XP SP1. Oddly the links in the header template at the top of this page (post) show it fine, but when it's any other part of the page it just flashes up briefly before returning to "Done". It does the same on the link I just wrote even as I'm previewing this section edit. violet/riga (t) 17:59, 18 September 2005 (UTC)

Works OK in Mozilla (Seamonkey) Suite. *Dan T.* 18:36, 18 September 2005 (UTC)

Attempt to create link using '[[' makes text disappear

I attempted to create a link using '[['. Instead of the text turning into a link, the text disappeared. See the attempt at [4] What did I do wrong? Bobblewik 11:37, 17 September 2005 (UTC)

When linking to categories, you have to prefix it with a ":", ie [[:Category:Whatever]]. Otherwise you just add the article to the category gkhan 12:41, 17 September 2005 (UTC)

That is a surprising requirement of the software, I would not have guessed it. Many thanks for letting me know. The request for comment now reads correctly as:

"Articles in Category:New York City Subway passenger equipment. How should specifications be displayed. Should they include spelling errors, non-standard capitals and tautology such as Length <xx> long. Editors appear to be engaged in a slow revert war with edits from anonymous addresses across all the articles in the category."

Bobblewik 14:39, 17 September 2005 (UTC)

error on image load

Whenever I try to load the following image page:
http://en.wikipedia.org/wiki/Image:Parthenonnashville1.jpg
I get this error:
Error in numRows(): Lock wait timeout exceeded; Try restarting transaction
or the image page with a blank image, or occassionally it will load fine.

This image page used to work, but then it was vandalized and reverted and has been acting wonky ever since. For example, even when the image page loads fine, if you click on the initial revision, it shows the vandalized version instead. How can I get this image back to normal? Kaldari 20:21, 17 September 2005 (UTC)

It works fine for me. Perhaps bypassing your cache will help. -- Finlay McWalter | Talk 20:23, 17 September 2005 (UTC)

I ended up just deleting the image completely and reuploading since it was so messed up. I tried reverting it first, but then it gave me the correct image at the resolution of the vandalized version (and file size of the vandalized version). Something was definitely screwed up. Kaldari 20:48, 17 September 2005 (UTC)

There were a lot of problems with this image yesterday. Somehow, somebody managed to upload an obscene image over it, and even when reverted, the image kept showing up, despite several people's attempts to correct it. When you went straight to the image page, the image looked good, but when you went to the Parthenon page, the obscene image showed up. Finally a new version of the image was uploaded and, I believe, an image was uploaded to Commons. See WP:ANI#Image vandalism for the discussion on this. User:Zoe|^(talk) 23:28, 17 September 2005 (UTC)

As a reminder: we have the capability to disable overwriting of existing images with new versions. If the community would like to enable this as a measure against image-vandalry we could do that. --Brion 04:53, 18 September 2005 (UTC)

Brion, if that feature is enabled, does it still allow the original uploader to overwrite the image? -- Finlay McWalter | Talk 18:13, 18 September 2005 (UTC)

An "original uploader" has no special privileges anywhere in the software, no. A fundamental part of the wiki way is that being there first doesn't give you special rights over those who come next. --Brion 20:39, 18 September 2005 (UTC)

Would the feature work inter-project - i.e. could I upload an image:BrionIsCool.jpg to Commons if such an image existed here (or vice versa)? Thryduulf 22:22, 18 September 2005 (UTC)

Currently you cannot upload a new local image if there's a conflicting image on Commons. --Brion 01:52, 19 September 2005 (UTC)

Another possible solution to image vandalism would be to change the software so that uploading new versions of an image automatically made a trivial edit to articles and templates which are directly linked to it (ie not articles which only include the image through a template). Then people with that page in their watchlist could check the change just like normal changes/vandalism. I don't know if this is technically feasible, though. Lupin|talk|popups 12:30, 19 September 2005 (UTC)

Thumbnail sizes

120

I don't know how long this has been going on for but I just noticed today that in special:preferences you can specify your preferred size of thumbnail. You can choose from 120px, 150px 180px, 200px, 250px, or 300px. If an image width is less than thumb size the in your settings, it is displayed at 100% resolution. This said, I think it important now to go through and remove size specifications from most of the thumbnails around so that users can define their own. I've got mine set on the largest. Dunc|☺ 19:49, 19 September 2005 (UTC)

What articles have I started?

Is there any way I can find a list of all the articles I have started? It would be nice if we could list all new articles on the user contributions page, is that something that could be implemented easily? thanks - Martin 20:47, 19 September 2005 (UTC)

I asked for roughly the samne thing above ("New contributions"). The answer seems to be No. A bit of an oversight - Xed 20:53, 19 September 2005 (UTC)

I think the answer to this lies a couple of comments above, at Error On Image Load, specifically the bit that says being there first does not give a user special privileges or something to the effect. In any case, this does not require software modifications, as it can be solved by the user bookmarking any pages he starts.--81.42.154.225 13:43, 20 September 2005 (UTC)

Twould be nice, I agree. But that's half the reason some people keep lists of contributions on their user page (or a subpage). I've been using a standard edit summary "create article" to make it easier to spot on my contribs list, but a simple list'd be nicer. — Catherine\^talk 02:14, 22 September 2005 (UTC)

Displaying certain (special) characters

My browser (IE6, SP2) does not display certain characters from Wikipedia's pages, such as á, é and € etc. - in case these do not display now, they are a and e with acute accents and the Euro symbol. I have set my encoding to Unicode UTF-8, in the view menu - all the other options seem to diminish other characters. Can you advise? 213.94.144.109 00:58, 20 September 2005 (UTC) 01:58 BST 20 September 2005

Use Mozilla Firefox. — Ambush Commander^(Talk) 01:30, 20 September 2005 (UTC)

@AC: Don't be silly. This is not an IE problem. I'm currently using IE and perfectly able to view those characters. Whether people should use Firefox or not is a decision they should make for themselves, and that should be based on different arguments.

That being said, UTF-8 is the correct encoding, so something else must be wrong. Have you checked your fonts?

Select some text and paste this in the address bar:

javascript:alert(document.selection.createRange().parentElement().currentStyle.fontFamily);

It should say "sans-serif" if you're using monobook.

Have you got a sans serif font installed that contains accented characters? If not, using another browser is not going to help you.

If you have fonts installed that do show accented characters, try fiddling around with your IE settings. Especially your "font" and "language" settings may be worth having a look at, although on my system the default setup seems to work just fine. As a reference, check whether the following are displayed properly: a (a, without accent), ä (a umlaut), á (a aigu), € (euro), ÿ (y diaresis), もののけ姫 (Princess Mononoke in Japanese), 新幹線 (Shinkansen in Japanese).

If you continue having problems, feel free to drop me a line. Shinobu 13:24, 20 September 2005 (UTC)

Hehe, after reading everything involved, it sounds to me like switching to FireFox is the far less silly option. ;-) (Oh, and get a Mac while you're at it...life becomes much simpler and pleasant) :) Func( t, c, @, ) 21:24, 20 September 2005 (UTC)

Again, this is not a browser problem. I'm using IE while I'm typing this post, and all characters are displaying properly. The first step checks the font, so you can see if it's a problem with your "wikiskin" (e.g. your monobook.css). If it is, switching browsers won't help, editing your "wikiskin" will. If you have (or someone else has) changed your IE fonts, yes then switching to Firefox will help. If you fuck it up in Firefox too, switching to yet another browser will help again, and so forth.

The reason you shouldn't switch because of this, is that there are real differences between these browsers that need consideration. Shinobu 02:20, 21 September 2005 (UTC)

Yeah, there are... IE sucks and Firefox doesn't! :-) *Dan T.* 02:37, 21 September 2005 (UTC)

It's not that b&w and you know it. Ask anyone who likes WindowBlinds and/or Visual Styles. Or indeed ask any Chrome hater. "Product Xyzzy sucks" never really was a convincing argument, and never will be. Firefox and IE are different and which one you use depends on your personal needs/preferences. That being said, I'm here to solve problems, not to debate Firefox/IE. And when the "s-word" is being used, that fact alone indicates it's a good time to unwatch a page. *unwatches page* Shinobu 03:03, 22 September 2005 (UTC)

Now that's a bit extreme, methinks (after all, there are other conversations here). Whatever. I don't even know if the anon is listening. Anyway, it doesn't hurt to try an alternate browser. The main thing is "switching to another browser" vs. "fiddle with the wikiskin and your fonts" (by the way, anonymous monobook.css/monobook.js files don't have any special effect) is that one is a clear path, while two is a fuzzy one. We need more diagnostic information to help. By the way, Firefox is very good about mixing up fonts to get as complete a Unicode character set as possible. I'm not sure if IE does the same, but it could be that you're right: it is a font problem: and you need a browser that is smart about this sort of stuff to work. Once again, this is all theorization.

In related news, Opera is now offered for free without ads! If you get a chance, try Opera too, and see if it breaks there. — Ambush Commander^(Talk) 03:25, 22 September 2005 (UTC)

thumbnail problems

I'm getting broken images instead of thumbnails for any new versions of images I upload. If I try to pull up the thumbnail URLs that are generated, I get 404 errors. For example, this thumbnail. I've noticed this problem several times today. Is this just database slowness that will fix itself? Kaldari 04:31, 20 September 2005 (UTC)

It seems the thumbnails eventually fix themselves and get new URLs. Perhaps this is simply an issue of server side caching on the article pages. Kaldari 04:58, 20 September 2005 (UTC)

As I recall, the new thumbs will be generated when the pages containing them are revisited. --Brion 05:50, 20 September 2005 (UTC)

I've encountered this a few times in the past week or so. Purging the cache of the affected page seems to clear it up. -- Norvy (talk) 06:11, 20 September 2005 (UTC)

Ah now I am wiser! "&action=purge"! Kaldari 15:13, 20 September 2005 (UTC)

This should be fixed now -- Tim Starling 10:47, 22 September 2005 (UTC)

Difference tool

Has there been a recent change to the "difference" tool? It seems to be working much less well than a couple of days ago, failing even to "notice" that the first paragraph of an article is unchanged. -- Jmabel | Talk 05:04, 20 September 2005 (UTC)

Yes, we've switched to a faster, but somewhat buggier, version. Hopefully the bugs will get worked out. --Brion 05:32, 20 September 2005 (UTC)

Hope so, because if it stays this buggy, I don't care how fast it is! -- Jmabel | Talk 06:18, 20 September 2005 (UTC)

I had the same problem two days ago at en:, nn: and no:. Now it seems to have resolved at nn: and no: (e.g. [5], [6]), and it has improved at en: but not completely (e.g. [7]). The latter didn't make any sense to begin with, but now at least the first few sentences are recognised. Hopefully it will continue to improve. --Eddi (Talk) 10:59, 21 September 2005 (UTC)

Un-nomination for deletion

What are the technical steps, if any, necessary to remove an article from the AfD list? Can someone just delete it from the page? My stub on Sophia Michahelles of Superior Concept Monsters no longer merits deletion -- (if this needs to run its course, find -- no big deal). I'm really just interested in the process for future reference. paul klenk 06:28, 20 September 2005 (UTC) Secchi

1. Click 'edit this page'
2. Make change
3. Click 'Save page'
   • There might be silly folks insisting that you follow some social procedure, but there's nothing technical about a list on a wiki page. :) --Brion 09:56, 20 September 2005 (UTC)

Thanks -- nothing technical about it, evidently. Answers my question. paul klenk 11:27, 20 September 2005 (UTC)

• Closing AFD discussions is often done by administrators, unless they're clearly without a doubt a consensus to keep. Visit the discussion page and add {{subst:at}} to the top of the article and {{subst:ab}} to the bottom to close the article. Then remove the tag from the article and make a note on the talk page that it was up for deletion. See Wikipedia:Deletion process. - Mgm|^(talk) 11:51, 21 September 2005 (UTC)
  • Thanks, Mmg. paul klenk ^talk

Context menu extension for Firefox/IE

Is there a context menu extension for Firefox/IE available. The scenario I have in mind is that a user will select a word in their browser and then right click. A popup menu will appear with Wikipedia Search as the title. The user will then click the Wikipedia Search menuitem and a new Browser session will start with the selected word as the search term in Wikipedia and the relevant page displayed (Google has something similar.) There could also be a Wiktionary link in the popup menu as well. Anyway if none exists I am happy to write an Internet Explorer extension. Perhaps someone can volunteer for the Mozilla extension version. I have written the IE version today. If anyone is interested I can forward the info required to download. -- Ctownshend

• I believe there's already such an extension, but without the machine I've got Firefox installed on, it's hard to say. Email me and I'll doublecheck. - Mgm|^(talk) 11:55, 21 September 2005 (UTC)
• Mgm is correct, there is already such an extension. You can get it here and add Wikipedia to the list of search engines by following the link from here. --GraemeL ^(talk) 13:21, 21 September 2005 (UTC)

Updating the search index

The search index does not appear to have been updated for some time. As time goes by, the number of false positives and misses in search results is increasing.

I remember we discussed this before. I can't find a record of the discussion but my recollection is that improvements were being made to search. I thought that one of those improvements involved automating re-indexing. I may have misunderstood.

Would it be possible to discuss this issue again? Bobblewik 16:02, 20 September 2005 (UTC)

Still not done. --Brion 21:00, 20 September 2005 (UTC)

OK. Thanks for your response. I appreciate all your good work. Bobblewik 00:40, 21 September 2005 (UTC)

Default arguments for template paramaters

I've commited code by myself and Thomas Kjosmoen to the parser which enables default arguments for template paramaters in the form of {{{argument|default}}} I've made a sample template to demonstrate it at Template:Default arguments, which contains the following wikicode:

* <nowiki>{{{argument}}}</nowiki>: {{{argument}}}
* <nowiki>{{{argument|default}}}</nowiki>: {{{argument|default}}}

if you call it like {{Default arguments}} it will yield:

but call it like {{Default arguments|argument = hello world}} and it will yield:

—Ævar Arnfjörð Bjarmason 19:45, 20 September 2005 (UTC)

Image:Wheat close-up.JPG

Does this have anything to do with the fact that the license templates on my images seem to not be working now, or just a coincidence? Martin 19:51, 20 September 2005 (UTC)

I think that may have to do with the fact that your templates use 5 braces around the arguments instead of 3. ~~ N (t/c) 20:04, 20 September 2005 (UTC)

I didnt make the template, it was automatically made by the license selector thingy. Martin 20:09, 20 September 2005 (UTC)

This might break templates which refer to other templates, such as Template:If equal, Template:Tl, and Template:Unsigned. Testing Tl: {{unsigned}}. (SEWilco 20:00, 20 September 2005 (UTC))

Why? ~~ N (t/c) 20:04, 20 September 2005 (UTC)

Oh, I misread the code. The phrasing above is for referring to parameters, not templates. Previous syntax only allowed the value of a parameter to exist within triple-braces. (SEWilco 20:08, 20 September 2005 (UTC))

It broke half of the infobox templates at Polish Wikipedia... Ausir

Reverted live copy due to reports of massive template breakage. --Brion 20:17, 20 September 2005 (UTC)

Reverted in CVS as well. Added a test case to the parser suite as well; the code seemed to get confused on template nesting such as {{{{{1}}}|{{{2}}}}} --Brion 20:59, 20 September 2005 (UTC)

oops -Ævar Arnfjörð Bjarmason 15:24, 21 September 2005 (UTC)

Taxonomy and Navigation System

I was wondering if there are any good examples or ideas on a better taxonomy system within wikipedia and it's related sister projects. I am currently building my own personal wiki site to categorize images and items within those images for reference material. Problem is that there are at lease 4 cross search sections that need to be taken into account and an unlimited possibility of searchable results. I also would like to dynimicly create navigational menus that reference the tagged metadata...

example: Image of a American WWII solider... In this pic we would want to describe and search on his..

Rank Uniform Timeline of image equipment weapon

These cats would also have subs like...

weapon

  Hard Gun
  Rifle
  Blade

These cats would have cats of there own...

  Hand Gun
   44
   38
   45
   9mm

Then finally the type

   Standard issue sidearm (etc)

So now I want to search for a American Solider + private + first-class + 9mm

I'm going to get every image that matches 1 of these words and not all of these words

Also, I would have to create the navigation menus by hand... That would take a really long time and updates to the navigation would effect almost all of the pages in my project.

So to the meat... What would be the best way to dynamically create menus with some sort of tag other then a category?

Are there any good taxonomy systems on wikimedia that I can review before I strike out on my own? Are the dynamic?

Is there a way that I can search that requires all works in the search?

Thanks FreRange 23:27, 20 September 2005 (UTC)

Where can I find the developers?

I'd like to throw a technical feature past some people who know how the software works here. Where is a good place to start? Thanks. paul klenk 01:23, 21 September 2005 (UTC)

Try the mailing list wikitech-l or Bugzilla. — Ambush Commander^(Talk) 01:25, 21 September 2005 (UTC)

Is it just me, or is search no longer giving summaries?

When I used to search, it would show the sentence containing the keyword. My preferences are set to:

• Lines per hit: 5000
• Context per line 5000

I have not changed them. However, I recently noticed that I am no longer getting a summary below the page title.

For example, a search for 'kalahari' gives Hukuntsi as the 5th result. It contains only one sentence: Hukuntsi is a village in Kgalagadi (Kalahari) District in Botswana.. Yet that sentence is not shown in the search result below the title as I would expect.

Is it just me, or is search no longer giving summaries?

I disabled it as an emergency optimisation measure, two out of three of our major database servers are currently offline and site performance was suffering as a result. Despite the low hit rate, context display accounted for about half of the load on the master database server. It'll be re-enabled when the hardware is back online, I don't know when that will be. The change was announced on wikitech-l, as usual. -- Tim Starling 06:19, 21 September 2005 (UTC)

Thanks for the response. I appreciate your efforts. I do not subscribe to wikitech-l. However I have been known to look at Wikipedia:Software status and/or meta:Wikimedia hardware status. I imagine that that is the way for technical issues to be communicated to me (unless there is a better single status page). I don't always understand the technical language but they do help. Is this issue the sort of thing that might be worth putting there? Bobblewik 13:55, 21 September 2005 (UTC)

Alt-D and admin powers

Alt-D is the keyboard shortcut in Firefox for "Go to the location bar", which I use frequently. Unfortunately, Alt-D is also the "Delete this page" shortcut in the Monobook skin, and ever since I became an admin, I've been accidentally going to the "Confirm delete" screen a lot. Is there any way to disable this one shortcut while retaining the others? android79 19:09, 21 September 2005 (UTC)

Use Ctrl-L. --Golbez 19:16, 21 September 2005 (UTC)

That's a nice workaround, but it won't change my habits. Alt-D is too far ingrained. :-) android79 19:22, 21 September 2005 (UTC)

I dunno; Ctrl-O was ingrained for me from using IE, but I shifted to Ctrl-L rather quick. You might be able to do it by editing your CSS. --Golbez 19:34, 21 September 2005 (UTC)

You can solve this in javascript. Something like this in your monobook.js should do the job:

function removeAccessKeys(keylist) {
  var t=document.getElementsByTagName('A'), u=document.getElementsByTagName('input');
  for (var i=0; i<t.length+u.length; ++i) {
    var s=((i<t.length) ? t[i] : u[i-t.length]);
    for (var j=0; j<keylist.length; ++j) if (s.accessKey==keylist[j]) s.accessKey='';
  }
};

function removeDeleteKey() {removeAccessKeys(['d']);};

if (window.addEventListener) window.addEventListener("load",removeDeleteKey,false);
else if (window.attachEvent) window.attachEvent("onload",removeDeleteKey);

Lupin|talk|popups 01:28, 22 September 2005 (UTC)

Hi, Lupin. It turns out to be simpler than the above: there is a global array 'ta' that holds access-key info and popup-help messages. The access keys don't get set on links until after page load, so we can unset them before page load:

 ta[ 'ca-delete'   ][ 0 ] = '';
 ta[ 'ca-undelete' ][ 0 ] = '';

Here is the complete list:

 /* tooltips and access keys */
 ta = new Object();
 ta['pt-userpage'] = new Array('.','My user page'); 
 ta['pt-anonuserpage'] = new Array('.','The user page for the ip you\'re editing as'); 
 ta['pt-mytalk'] = new Array('n','My talk page'); 
 ta['pt-anontalk'] = new Array('n','Discussion about edits from this ip address'); 
 ta['pt-preferences'] = new Array('','My preferences'); 
 ta['pt-watchlist'] = new Array('l','The list of pages you\'re monitoring for changes.'); 
 ta['pt-mycontris'] = new Array('y','List of my contributions'); 
 ta['pt-login'] = new Array('o','Logging in is not required, but brings many benefits.'); 
 ta['pt-anonlogin'] = new Array('o','Logging in is not required, but brings many benefits.'); 
 ta['pt-logout'] = new Array('o','Log out'); 
 ta['ca-talk'] = new Array('t','Discussion about the content page'); 
 ta['ca-edit'] = new Array('e','You can edit this page. Please use the preview button before saving.'); 
 ta['ca-addsection'] = new Array('+','Start a new discussion.'); 
 ta['ca-viewsource'] = new Array('e','This page is protected. You can view its source.'); 
 ta['ca-history'] = new Array('h','Past versions of this page.'); 
 ta['ca-protect'] = new Array('=','Protect this page'); 
 ta['ca-delete'] = new Array('d','Delete this page'); 
 ta['ca-undelete'] = new Array('d','Restore the edits done to this page before it was deleted'); 
 ta['ca-move'] = new Array('m','Move this page'); 
 ta['ca-nomove'] = new Array('','You don\'t have the permissions to move this page'); 
 ta['ca-watch'] = new Array('w','Add this page to your watchlist'); 
 ta['ca-unwatch'] = new Array('w','Remove this page from your watchlist'); 
 ta['search'] = new Array('f','Search Wikipedia'); 
 ta['p-logo'] = new Array('','Main Page'); 
 ta['n-mainpage'] = new Array('z','Visit the Main Page'); 
 ta['n-portal'] = new Array('','About the project, what you can do, where to find things'); 
 ta['n-currentevents'] = new Array('','Find background information on current events'); 
 ta['n-recentchanges'] = new Array('r','List of recent changes in Wikipedia'); 
 ta['n-randompage'] = new Array('x','Load a random page'); 
 ta['n-help'] = new Array('','The place to find out.'); 
 ta['n-sitesupport'] = new Array('','Support us'); 
 ta['t-whatlinkshere'] = new Array('j','List of all wiki pages that link here'); 
 ta['t-recentchangeslinked'] = new Array('k','Recent changes in pages linked from this page'); 
 ta['feed-rss'] = new Array('','RSS feed for this page'); 
 ta['feed-atom'] = new Array('','Atom feed for this page'); 
 ta['t-contributions'] = new Array('','View the list of contributions of this user'); 
 ta['t-emailuser'] = new Array('','Send a mail to this user'); 
 ta['t-upload'] = new Array('u','Upload images or media files'); 
 ta['t-specialpages'] = new Array('q','List of all special pages'); 
 ta['ca-nstab-main'] = new Array('c','View the content page'); 
 ta['ca-nstab-user'] = new Array('c','View the user page'); 
 ta['ca-nstab-media'] = new Array('c','View the media page'); 
 ta['ca-nstab-special'] = new Array('','This is a special page; you can\'t edit the page itself.'); 
 ta['ca-nstab-wp'] = new Array('c','View the project page'); 
 ta['ca-nstab-image'] = new Array('c','View the image page'); 
 ta['ca-nstab-mediawiki'] = new Array('c','View the system message'); 
 ta['ca-nstab-template'] = new Array('c','View the template'); 
 ta['ca-nstab-help'] = new Array('c','View the help page'); 
 ta['ca-nstab-category'] = new Array('c','View the category page');

Any access key can be removed or changed by adding this to your monobook.js:

 ta[ 'n-randompage' ][ 0 ] = 'z'; // now control-z rather than control-x

Func( t, c, @, ) 04:17, 23 September 2005 (UTC)

Wow, thanks, guys! BTW, Lupin, your admin-tools script works wonderfully as well. (Note to self: learn JavaScript at some point.) android79 15:12, 23 September 2005 (UTC)

Getting logged out every few minutes

Starting about an hour ago, I've been getting kicked out of my account every 3 or 4 minutes. I've never had this problem before, and it's very frustrating. Has this been happening here? Is there anything I can do about it? --Blackcap | talk 20:58, 21 September 2005 (UTC)

I have heard that usually, when that happens, it's a session problem. It's been reported that deleting all your wikipedia.org cookies makes that kind of problem go away. --cesarb 02:29, 22 September 2005 (UTC)

MDAC not showing diagram image

There is a diagram image on MDAC, but it is not displaying! Any issues going on I should be aware of? - 203.134.166.99 03:54, 22 September 2005 (UTC)

It looks OK to me. Perhaps the servers were slow when you viewed it and the image fetch timed out? Please try the suggestions at Wikipedia:Troubleshooting if simply reloading the page doesn't fix it. And, if those don't either, please come back and let us know the specific image you're talking about. -- Rick Block (talk) 04:04, 22 September 2005 (UTC)

There are apparently some oddities with our current image thumbnailing setup which may cause thumbnail images to temporarily vanish after an image is modified, depending on cache interaction.

Fiendish plans are being made to fix this. For now, if you notice it happening stick "?action=purge" onto the page URL (or do an edit preview) to force it to re-render, and the image should pop right back up. --Brion 09:00, 22 September 2005 (UTC)

Fiendish plans? Is Darth Vader a developer or something? gkhan 09:26, 22 September 2005 (UTC)

This should be fixed now -- Tim Starling 10:49, 22 September 2005 (UTC)

MediaWiki:Bad image list protection

The page is still protected for non-administrators (as it is supposed to be for the MediaWiki namespace), but when I try to edit the page, I do not get the "protected page" message. The protection button also says "Protect this page" instead of "Remove protection", as if it were unprotected. Why is this? --Ixfd64 08:56, 22 September 2005 (UTC)

edit: Several other MediaWiki pages seem to do the same thing. --Ixfd64 09:04, 22 September 2005 (UTC)

edit: All right, I did a test, and I found out that MediaWiki pages are indeed protected, even without the "protected" status. However, the "protected page" message only appears if the page has a proper "protected" status. --Ixfd64 09:14, 22 September 2005 (UTC)

XML parsing

I tried to parse the 20050909 Wikipedia dump (pages_current.xml) with Expat (through Python 2.4), and it failed with exception "reference to invalid character number" on the Chinese for Bohrium, which was written with an XML surrogate (&#xD801;). What should I do? Where is the failure in standards compliance, in Expat or the Wikipedia dump (or is it possible that I messed up something?)? What parser do you use/recommend for parsing the dump?

Thank you in advance for any help and I'm sorry if this is not the appropriate place/way to ask this question, I'm rather new to Wikipedia. --DanielDarabos 09:37, 22 September 2005 (UTC)

That entity is invalid, the question is how did it get there, maybe there is a design fault in the xml dump process.

P.S. does anyone know where the source for the tool used to make the XML dumps is located? Plugwash 10:29, 22 September 2005 (UTC)

I was thinking about getting rid of these surrogate codes via preprocessing, and that is most likely what I will do in the end, but they are not appearing exclusively in Chinese, but in Gothic language as well. This means that some useful content (not that Chinese texts are useless generally, but for me they are now :)) will be lost. I'll check out the Perl converter by Erik Zachte and see how he deals with it. --DanielDarabos 12:01, 22 September 2005 (UTC)

I vaguely recall there was a discussion on the topic of invalid characters (in the german dump?) on the wikitech mailing list not that long ago. Have a look there. Dmharvey File:User dmharvey sig.png Talk 12:16, 22 September 2005 (UTC)

I've found the discussion here. It includes tips on how to work around the invalid characters and also tells me that the bug has been fixed, next dump will be fine. Thank you! --DanielDarabos 13:24, 22 September 2005 (UTC)

The Unicode character in question looks like (金+波) and the code point is U+28A0F = &#166415; — obviously not a first-plane 16-bit character, it's in CJK Unified Ideographs Extension B (13 MB link). Here's the history of the Bohrium page's interwiki link to Chinese:

• Before April 30 2005, this was a numeric character reference [[zh:&#166415;]] (not a problem for the parser).
• Between April 30 and August 30 2005, FlaBot's edit changed this to a bogus interwiki link [[zh:&#55394;&#56847;]] that didn't even show up in the left-hand menu, but on the other hand this also was not a problem for the parser, which just sees literal ASCII & # 5 5 etc.
• After August 30 2005, YurikBot's edit changed this to a valid interwiki link of the form [[zh:�]], where we pretend that � represents literal U+28A0F. Works great when you click on it to go to the Chinese article, but maybe causes a problem for various database dump or parsing tools that might not understand it (surrogate pair encoding).
• September 9: the date of your 20050909 database dump, you encountered the problem

I have now changed it to a &#x28a0f; numeric character reference in the zh: interwiki link Bohrium article, but the various database dump and parse tools should nevertheless handle the literal Unicode character correctly, if for no other reason than we want them to work for the Chinese wikipedia, and the Chinese wikipedia does use this literal Unicode character as the title of its Bohrium article. -- Curps 09:57, 24 September 2005 (UTC)

Other Chinese wikipedia articles affected by this are Hassium, Seaborgium, Dubnium. -- Curps 10:26, 24 September 2005 (UTC)

Wikimedia namepsace in Wikipedia

There are two articles appearing on Special:Shortpages that have been created in Wikimedia namespace within Wikipedia. One is Wikimedia:Fundraising, the other Wikimedia:Wikimedia needs your help. There may be more, these are the only two I have noticed. The problem is because they are in an unknown namespace in Wikipedia I can't look at them to see what they are or what to do with them. Is anyone able to see what they are? -- Francs2000 File:Uk flag large.png 10:52, 22 September 2005 (UTC)

(As I guess you already know) you can't see the articles' contents normally, because they're overlaid by the Wikimedia entry in m:Interwiki map. You should be able to view their contents using Special:Export. I can get it to work for the former article, but not the latter (perhaps because it has spaces in its name) - all I get is a contentless XML page. If you're feeling brave, you could remove the Wikimedia line from the interwiki map, view the articles, and immediately restore the line in interwiki map - but you're bound to get yelled at for doing so. -- Finlay McWalter | Talk 12:38, 22 September 2005 (UTC)

Hmm, on looking at it further, the former query just gives me the contents of Fundraising. So you either need to do the interwiki map, or ask someone with SQL access to query it for you. -- Finlay McWalter | Talk 12:51, 22 September 2005 (UTC)

It was me who created these two pages. I did it because somehow the interwiki of Wikimedia: failed while both links were in the sitenotice (during the fundrasing drive). After the interwiki was fixed, both became inacessible. Their contents are just "See http://wikimediafoundation.org/wiki/Fundraising" and "See http://wikimediafoundation.org/wiki/Wikimedia_needs_your_help" (give or take a bit of punctuation). If you want both gone, ask Brion to rename them in the database (as he did with a lot of broken title articles a while ago), and then you can simply delete both. --cesarb 21:11, 22 September 2005 (UTC)

"Broken" template display in firefox

I don't know if this is the best place for this question, but I'll try anyway. Does anyone have a clue of why the Article Improvement Drive template shows a broken image as in this screenshot? It seems to happen with Firefox 1.0.6 and 1.0.7 in both Win XP Home and Pro. -- Rune Welsh ταλκ 12:57, 22 September 2005 (UTC)

Ragdoll

I don't know if this is the right place to take this, but the picture on the [Ragdoll] page has been mysteriously deleted. Is there anyway to figure out what happened to it, who did it and why? Thanks.Gator1 21:03, 22 September 2005 (UTC)

It was deleted by user:JesseW as part of an ongoing programme to delete images which don't have the proper source information, and which consequently may pose a legal threat to Wikipedia (by dint of being violations of someone else's copyright). -- Finlay McWalter | Talk 21:08, 22 September 2005 (UTC)

You just have to use the delete log to know all these pieces of information you asked. --cesarb 21:14, 22 September 2005 (UTC)

Hiding image metadata

I've been searching in vain for any info regarding the new visibility of metadata tables on Image:, pages. Can I choose to hide them somehow on the image pages I've uploaded? This foul and pointless annoyance new feature doesn't seem to fit well with the layout of existing image pages on either IE or Firefox and overlaps (i.e. hides) the information I carefully researched and provided with the images I've uploaded. ~ Veledan • Talk ^{+ new} 23:54, 22 September 2005 (UTC)

I believe you cannot selectively turn off these tables. However, if they are causing layout problems, the CSS should be edited to fix them. --cesarb 00:13, 23 September 2005 (UTC)

Troubles with links

Hi! I had problems with links in pages who have simbols like á, à, ã, ç, and similar letters. they snd me to an error page, with stranges associations with these letters (an Cádiz turn on C!:ediz, for example.) This trouble is on the pages, or in my PC? How to fix it? And, if this is not the right place to post this issue, where i can post it?

What operating system and browser are you using? Susvolans ⇔ 15:00, 23 September 2005 (UTC)

Redlinks

Am I the only one who sees redlinks as being striked through? Why was this introduced? Sjakkalle (Check!) 06:26, 23 September 2005 (UTC)

It could be a bug in your browser that's positioning underlines incorrectly. If you give more details about your browser and system it may be easier for people to help. --Brion 07:52, 23 September 2005 (UTC)

Perhaps, but it appears to be gone now. Thanks, BTW. Sjakkalle (Check!) 08:27, 23 September 2005 (UTC)

Any way to check a user's "non-reverted contributions"?

I'm wondering whether there's any way to check all the articles that have a given user as their last editor.

If there's no way to do this, please consider this a feature request. This would, I think, be a great tool in fighting persistent vandals; it would be possible to see which articles still have the vandal as the last editor (and would thus still need to be checked for vandalism).

Thank you! --Ashenai 14:46, 23 September 2005 (UTC)

Clarification: I know about the (top) thing; basically, I'm wondering if there's a way to only show the entries with (top). --Ashenai 14:55, 23 September 2005 (UTC)

custom monobook.js previewing broken?

It seems that the previewing of user monobook.js changes is broken at the moment. If I try it, all I get is the script's content dumped out again on top of the page. This is rather annoying, even though it boosts my edit count :/ Does anyone else have the same problems? -- grm_wnr Esc 22:42, 23 September 2005 (UTC)

I've found previewing and diffing javascript with the monobook skin is broken. It seems to work pretty well in the Classic skin, though. Lupin|talk|popups 02:18, 24 September 2005 (UTC)

Zomg...yeah, it's very seriously broken. Func( t, c, @, ) 02:34, 24 September 2005 (UTC)

This is probably because you guys have "</script>" appear in your scripts. I'll see about a workaround. --Brion 02:00, 25 September 2005 (UTC)

OK, if that's the issue, then it's no big deal to break up the string. Still, this is something that just started happening, so far as I know. Func( t, c, @, ) 03:24, 25 September 2005 (UTC)

Wronly sorted User Contributions

I noticed this during the adaptation of the new MediaWiki software time, and I thought it was fixed. But then since few days ago, edits on the User Contributions are sorted wrong again. For example, see this, the last few edits were on same page, but the last edit is below the first edit. Notice the (top) is below the non-top... Thanks, -- pmam21^talk_articles 02:31, 24 September 2005 (UTC)

Looks like clock synchronization wasn't working right on some of the new servers; I think I've got them re-synched now. --Brion 08:15, 24 September 2005 (UTC)

Still doing the same thing... -- pmam21^talk_articles 08:48, 24 September 2005 (UTC)

They already are recorded, they wouldn't have changed. --Brion 18:20, 24 September 2005 (UTC)

OK. Thanks, -- pmam21^talk_articles 21:04, 24 September 2005 (UTC)

Watchlist Concerns

Who can see my watchlist? And who can see which users are watching a given page? Everybody? Or just sysops?

I'm asking this for privacy reasons - I prefer to remain anonymous. For example, I don't want people to work out who I work for (which would be possible if I added my employer's article to my list), and I don't want my employer to work out who I am (by seeing who's watching my employer's article and then looking up the watchers' watchlists).

I know I could just create separate accounts for different partitions of my life, but it's inconvenient and I'd rather not do it if I don't have to.

I've looked at wikipedia:watchlist and the MediaWiki equivalent but can't see anything that answers my questions. --A bit iffy 08:30, 24 September 2005 (UTC)

As far as I know, there is no way for a regular user or an adminsitrator, at least, to see either what someone is watching or who is watching an article; I doubt bureaurcrats have such power either. I'd guess that the software developers would be able to go into the software and examine a watchlist, though I'm not sure. The short answer is that you should be able to watch your employer's article freely and not worry about people finding out—as an administrator, I have no power to see what is on your watchlist. — Knowledge Seeker দ 08:39, 24 September 2005 (UTC)

Only you can see your watchlist using the software; the developers can of course look in the database and see everyone's watchlists, but they will not do it unless they are chasing a bug, because of the obvious privacy implications. --cesarb 13:28, 24 September 2005 (UTC)

As CesarB says, noone can see your watchlist (well, in theory developers can, but they are all sweet little puppies who know ridiculusly much about squid servers). Everyone however can see your contributions, so if you really are very worried that someone should find out who you work for, don't edit your companys article. However, I seriously doubt anyone is going to go to the trouble of finding out. gkhan 04:28, 25 September 2005 (UTC)

No underline

This must have been discussed before but I don't know where. Could someone point me to it? Why have the links suddenly lost their underline? Is there any way to get the underline back? Even if it's just when I log in... --Celestianpower ^hab 20:44, 24 September 2005 (UTC)

Preferences, Misc, "Underline links". --cesarb 21:06, 24 September 2005 (UTC)

Spammy message

Is there any way to get rid of ther spammy message displayed on our watchlists. I read it the first time, and it OTT to get iot every time one goes to one's watchlist, and interferes with the page layout (you have to scroll to lopok beyond the 1st change, SqueakBox 01:39, 25 September 2005 (UTC)

You may have read it, but you didn't heed it. See your talk. Superm401 | Talk 05:03, 25 September 2005 (UTC)

But why is it there? It cannot be more important than say the fight against vandalism which it is getting in the way of by creating a much less user friendly watchlist. People tend to use their watchlist many times, and the message only needs to come up once, SqueakBox 05:10, 25 September 2005 (UTC)

It obviously doesn't, because you didn't follow it the first time you saw it. Also, in some ways it is more important than vandalism, because vandalism can't get Wikipediaia sued, but copyvios can. Also, what's wrong with Special:Log/upload. It misses uploads.Superm401 | Talk 05:15, 25 September 2005 (UTC)

There are (at a guess) comfortably north of ten thousand unsourced images currently on Wikipedia. These need nuked; every one places us at risk of a lawsuit which would do nasty things to the project. If we went ahead and did so people would scream fit to raise the heavens; so there needs to be some way of communicating it to all users, get them to do the work before we have to list threatening messages on their talk: pages. However, unlike things like the fund drive, it's a message only targeted at editors, not at casual readers - so it gets limited to the watchlist. Shimgray 11:04, 25 September 2005 (UTC)

Function "Mouse over"

I just wonder how it works when i pull the cursor over text in the English Wiki and can read synonyms in other languages (Spanish). Is it a sort of Java Script like "Mouse over"? It seems to work in my version of Firefox but not in Explorer.

I admire it and would like to use the same model in a project of my own.

Can you advise me a little? Jan Årmann, Sweden

I believe that's a feature of the Google extension to the Firefox browser. Therefore it's on the client side, rather than the server side (as in, the server makes no instruction). It would be technologically possible to write a combination of scripts to do the same, but it would be difficult and costly in resources. We are not the people to ask about this! Cheers, and I hope this helps. [[Sam Korn]] 12:32, 25 September 2005 (UTC)

Download bug - "Opening index.php"

Can anyone shed any light on this bug?

When I am logged in I can't edit pages using the 'edit this page' tab. Instead a window comes up saying "Opening index.php - You have chosen to open index.php" (in Firefox) or "File Download - Do you want to save this file?" (in IE).

As this suggests, the bug is not browser-specific. It's also not machine-specific: it happens on both my computers (both running Win XP), at different IP addresses.

I can't believe I'm the only one who suffers from this problem, but I can't find any help about it either on Wikipedia or through Professor Google.

There are work-arounds: oddly, I can edit sections, but not whole pages. And the bug doesn't occur when I am not logged-in. But this means (a) that I can't keep a record of all my changes; and (b) I can't create new pages.

Any help gratefully appreciated: it's all rather frustrating.

Oldhamlet 15:02, 25 September 2005 (UTC)

This sounds similar to a prob I had. Check your preferences tab at the top of the page and make sure you don't have it set to "Use external editor by default" under the "Editing" bullet. That was what caused the problem in my case. StuRat 15:40, 25 September 2005 (UTC)