This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.
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 Keane12:45, 14 September 2005 (UTC)
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.
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:
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.
Balearic_Islands entry is goofed up?
Hey -
when I try to look at http://en.wiki.x.io/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)
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.wiki.x.io 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? --Durin17:53, 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! --Tuvic18:55, 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. --cesarb19:13, 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. Martin23:34, 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)
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.) --Brion09:21, 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. --Durin17:24, 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. --Peripatetic19:49, 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 :) Renata303:28, 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 Wis07:26, 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 Wis07:14, 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. --Durin17:33, 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 Palin00:31, 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. Plugwash01:10, 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! --Mb100019:31, 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.wiki.x.io/ and it would send me immediately to the "Wikimedia servers have a problem" page. Does anyone know what's going on, and why? Titoxd02:13, 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.) --Brion09:25, 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 | Talk03:16, 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. --Brion09:25, 19 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 [1][2]
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.
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.
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 | Talk12:28, 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 ‎‏Wikipedia‏‎ 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.
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. --Yodakii18:21, 15 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:
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.
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)
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)
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 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)
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 [3] --Brion11:08, 14 September 2005 (UTC)
I believe this sometimes happens when the servers are under very heavy load. Bovlb 06:06:06, 2005-09-11 (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. Shanes07:12, 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. Shanes21:13, 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 Bjarmason17:37, 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.) --Brion21:33, 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. Bovlb22:46, 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 Bjarmason16:20, 19 September 2005 (UTC)
Thanks. How many repetitions are there? Maybe it would be simpler to collate the most frequent N for each day. Bovlb18:55, 19 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)
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)
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?
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.--81.42.154.11219:36, 13 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. Nandesuka04:19, 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 isBroken00:12, 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.--81.42.154.22513:28, 20 September 2005 (UTC)
Vertical alignment of images within a border
Hi, the subject says it all really - the case in question is my userpage (User:HighHopes), which is still under construction. On the left hand side, you'll see a box with a picture of a ballot box and link to the Politics portal and WikiProject. What I'd like to do is have the image on the left of the box but vertically in the middle. Can this be done? Thanks, --HighHopes(T)⋅(+)⋅(C)⋅(E)⋅(P)20:27, 28 September 2005 (UTC)
Hi. I'm an user of the italian wikipedia. I saw that in your search page, if the query doesn't match, there's a red link that lets create the new corresponding page. So in our wiki it doesn't exist. Is it a new feature of a new version of media wiki? Or we have to fix sth manually? please answer me. it:utente:The_Doc 28/09/2005
ideas for improving history
1. when displaying history/recent changes, show the number of lines/characters added/removed in each change, and the number of lines/chars in each revision. this will make it easy to find vandals and article blankings, and also useful when going back in the history to find when a particular change was made.
2. have an expanded view of history that shows a bit of text from each change. the summary is not reliable because the user could forget to fill it in or be misleading.
Example of why this is necessary is on the Albert Einstein page. On Sept 21, some anon deleted a bunch of text and replaced it with the word "gay". Then another anon removed the word "gay". The text was not replaced until a week later, probably because everyone thought the vandalism had been fixed. If the number of lines in each revision were shown, someone would have noticed the missing text, and it would have been much easier to find when it was removed. Also a few revisions later, some vandal doubled the text of the article, and then someone fixed it, and then someone reverted the fix, thinking someone had blanked the article. This confusion would be less likely with a line count.
Identifying users
So, we have a new users log. We have a contributions log. Is there any way that I can identify if an account has been registered if a) it has no contributions and b) it's too old for new-users? Just idly curious... Shimgray | talk | 15:11, 28 September 2005 (UTC)
i have a small issue with the "Amethyst" theme. I'm not sure if this is intentional or not or if your still editing it but there is no "search box" down either side or at the top and bottom. I quite like the this theme but i'd also like a search box :-P
Page hit counters
hhmmm.... probably not the best place to mention this but anyway. Is it possible to add counters to particular web pages in wikipedia, or just view how often particular sites have been viewed etc. 211.28.76.16517:00, 27 September 2005 (UTC)
Bogus error message when uploading an Image with a . in the first 3 chars of its name
The article Mr. Popo uses a standard template that calls for an image to be used with the name <article_name>_photo.jpg
This calls for Image:Mr._Popo_photo.jpg, but the trouble is, you can't upload such an image filename. You get the entirely bogus error "File names must be at least three letters."
The presence of a . in the filename is not the problem: as a test, I successfully uploaded Image:Testing. 1 2 3.jpg. The problem is the bogus length check, which is obviously searching for the first "." in the filename and not the last one. Should be an extremely trivial fix.
Perhaps someone could be kind enough to report this to Bugzilla, I don't use e-mail and therefore can't use Bugzilla. -- Curps19:52, 26 September 2005 (UTC)
I hit this last night - it seems to be that it assumes you have filename.extension, but doesn't quite realise that you can have file.name.extension. Hmm. Shimgray | talk | 20:02, 26 September 2005 (UTC)
Yeah it only counts up to the first . for filename length, This was meant to avoid counting the extention length but as you say it screws up on the case of multiple dots. There is already a bug report iirc but its not considered that serious by the devs so the chance of a speedy fix is pretty low. Plugwash10:12, 27 September 2005 (UTC)
Yes but the fix is absolutely trivial... just prefix the letter "r" in the right place (or does Wikimedia use a different programming language?) And the "right place" is known by just grepping for the error message given above. Shouldn't be more than 30 seconds, no? And two people independently bumped into this bug within a short time interval. -- Curps10:20, 27 September 2005 (UTC)
Btw i think forcing the image name from the template is a crazy idea in the first place. It encourages very short nondescriptive filenames, copying of images and replacing an image with a very different one whilst keeping the same name. None of theese are things we should be encouraging. Plugwash12:38, 27 September 2005 (UTC)
Perhaps, but you should take that up with the folks who edit Dragonball pages. I just stumbled on this by noticing and investigating an inexplicable Mr. Popo→Mr Popo move in the move log. -- Curps16:54, 27 September 2005 (UTC)
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 Welshταλκ12:57, 22 September 2005 (UTC)
What is your screen resolution? Did you fully open the window?
Firefox 1.0.7 refuses to remember my Wikipedia password: the Wikipedia login screen does not activate the password manager. It works with Mozilla 1.7.12. How can I get Firefox to do it, too? (On Win XP.) Lupo15:12, 26 September 2005 (UTC)
I'm not sure if this was mentioned before; I'm having some difficulty saving edits. I'll hit save but I keep getting a preview version; sometimes it takes me six or seven attempts to get the page to save. I've never had this problem before. (for this comment: at least ten undesired previews, tired of counting now) — Knowledge Seekerদ04:58, 26 September 2005 (UTC)
Hmmm, after ten failed attempts to save the page, I got frustrated and clicked "Save page" rapidly probably twenty times in a row, before it had a chance to do anything, and the edit took. — Knowledge Seekerদ05:00, 26 September 2005 (UTC)
Yea, I tend to do many edits instead of one big one because if I don't:
There is a higher probability my work will all be lost due to a Wiki error.
There is a higher probability I will get an edit conflict.
Should moving User: and User talk: pages be disabled now?
User and User_talk pages only seem to be moved under two circumstances:
Page move vandals
Clueless newbies who want to change their username but don't know that they're supposed to contact a bureaucrat.
Is there any legitimate circumstance to move a userpage anywhere, other than performed by a bureaucrat as a rename? If not, why not just disable it for anyone but bureaucrats (or maybe admins too, though offhand I can't think of a reason). On the other hand, the reverse should not be disabled (userfication of vanity pages in article space, etc). -- Curps00:51, 26 September 2005 (UTC)
Some people choose to draft new pages as user subpages and then move them into article space. Also some people archive talk pages by moving the entire thing to an archive name and starting over fresh. Dragons flight00:56, 26 September 2005 (UTC)
Yup, as mentioned above, I'm one of those people who moves off his talk page for archiving. Is there a major problem with people moving their user page? If not, I'm not sure we should be limiting any features. --fvw*01:00, 26 September 2005 (UTC)
OK, but how about the actual user page itself, as opposed to subpages of it? That probably shouldn't ever move. -- Curps01:02, 26 September 2005 (UTC)
The problems, as mentioned above, are vandals and clueless newbies. As far as I can think of, those are the only reason that user pages (not user subpages) get moved, other than a bureaucrat doing it the right way. If it's disabled, it's one less thing that can be messed up. -- Curps01:05, 26 September 2005 (UTC)
Actually I occasionally move a user page to the user talk page, where the welcome message has been posted to a new user's user page in error. Also I frequently move to user space, where a newbie has accidentally posted their user page in main namepsace (without the User: bit first) -- Francs2000File:Uk flag large.png01:07, 26 September 2005 (UTC)
The problem is, I'm pretty sure, that the <math> tag messes up the alt field. And since, according to Wikipedia:Alternative text for images, there is no way to separate caption and alt-text, you SOL. My suggestion is either to change the text into english instead of mathematics, ie. "This is a plot of the cube-root function", or simply use tables, like this. Now would be a good time for somebody who actually knows how to make neat little tables to come and show us how it's done :P gkhan22:38, 25 September 2005 (UTC)
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.
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. StuRat15:40, 25 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, SqueakBox01:39, 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, SqueakBox05:10, 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 | Talk05:15, 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. Shimgray11:04, 25 September 2005 (UTC)
How long do we havwe to endure this spammy message for? I must have uploaded it a couple of hundred times because I cannot access my watchlist without it. It shows a real lack of respect towards regular editors. Is there somewhere else I can discuss it, as presumably the decision to put it on the watchlist was taken somewhere by somebody. And can it go down again asap, SqueakBox20:21, 25 September 2005 (UTC)
Approximately a week, would be my guess. Since we are in the process of deleting 15000+ images that are not properly licensed and sourced, there needs to be some warning to general users to take care of their business, and it needs to remain long enough that regular but infrequent editors will have an opportunity to see it. A large fraction of these images are actually in use in articles, and if they can be saved by the addition of proper source and licensing information, we would like to see that happen. Keep in mind that there is no way, save reuploading, to restore an image that is deleted, so image deletion is a fairly big deal. Other discussion on the topic: User talk:Jimbo Wales (the message at top and 3 or 4 seperate threads), Mediawiki talk:Watchdetails (to discuss the text of the watchlist message), Wikipedia:Administrators' noticeboard#New speedy deletion criterion.
I eas able to fix it in my monobook.css using
#watchlist_notice { display: none; }, which entirely solves the problem. Great! SqueakBox23:26, 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... --Celestianpowerhab20:44, 24 September 2005 (UTC)
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, -- pmam21talkarticles02:31, 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. --Brion08:15, 24 September 2005 (UTC)
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_wnrEsc22:42, 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|popups02:18, 24 September 2005 (UTC)
This is probably because you guys have "</script>" appear in your scripts. I'll see about a workaround. --Brion02:00, 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)
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).
Clarification: I know about the(top)thing; basically, I'm wondering if there's a way to only show the entries with (top). --Ashenai14:55, 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. --Brion07:52, 23 September 2005 (UTC)
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?
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+ new23:54, 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. --cesarb00:13, 23 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.Gator121:03, 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 | Talk21:08, 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 | Talk12:38, 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. --cesarb21:11, 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 (�). 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. --DanielDarabos09:37, 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.
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. --DanielDarabos12:01, 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! --DanielDarabos13:24, 22 September 2005 (UTC)
The Unicode character in question looks like (金+波) and the code point is U+28A0F = 𨨏 — 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:𨨏]] (not a problem for the parser).
Between April 30 and August 30 2005, FlaBot's edit changed this to a bogus interwiki link [[zh:��]] 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 𨨏 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. -- Curps09:57, 24 September 2005 (UTC)
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? --Ixfd6408:56, 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. --Ixfd6409:14, 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. --Brion09:00, 22 September 2005 (UTC)
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 | talk20:58, 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. --cesarb02:29, 22 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? android7919:09, 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. --Golbez19:34, 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);
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:
/* 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
Wow, thanks, guys! BTW, Lupin, your admin-tools script works wonderfully as well. (Note to self: learn JavaScript at some point.) android7915:12, 23 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 Starling06:19, 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? Bobblewik13:55, 21 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 klenk01:23, 21 September 2005 (UTC)
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.
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)
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 klenk06:28, 20 September 2005 (UTC)Secchi
Click 'edit this page'
Make change
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. :) --Brion09:56, 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)
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 | Talk05:04, 20 September 2005 (UTC)
Yes, we've switched to a faster, but somewhat buggier, version. Hopefully the bugs will get worked out. --Brion05:32, 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. [4], [5]), and it has improved at en: but not completely (e.g. [6]). 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)
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? Kaldari04:31, 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. Kaldari04:58, 20 September 2005 (UTC)
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? 213.94.144.10900:58, 20 September 2005 (UTC) 01:58 BST 20 September 2005
@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:
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).
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. Shinobu02:20, 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*Shinobu03:03, 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.
Solution Found: For others who have this problem, I have solved it by going into Tools, Internet Options. Near the bottom-right you will see "Accessibility". Click this and then check the "Ignore Font Styles specified by web page" (under the formatting heading).
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)
I set the size to the smallest which shows the level of detail needed to see what I'm trying to show, so really don't want it to be overridden by user default settings. StuRat20:43, 25 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? Shinobu14:48, 16 September 2005 (UTC)
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? Boojum14:49, 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. --Ixfd6423:22, 16 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)
Special:Export Full History alternative?
I am researching the use of wiki's as a tool for emergency services coordination. As a part of this research I would like to conduct some analysis of the Hurricane Katrina article. I am trying to get a full history dump of that article, but unfortunately this capability has been disabled on the Special:Export page. Unfortunately, I don't have access to a machine that has enough capacity to deal with the full database dump. Does anyone have any suggestions for getting a full history dump of just the Hurricane Katrina article?
Thanks a lot. You guys rock! Unfortunately you've also deprived me of my best excuse for procrastinating on this thing. Now to figure out how to tease some interesting data out of all this text...
Save page does not work, brings up Preview instead
had to work for a while to get to this page, but Is anyone else getting this message on any random page?
Fatal error: Call to undefined function: linkprefix() in /usr/local/apache/common-local/php-1.5/includes/Parser.php on line 1232
Wikipedia's slow, and there must be some error...-- pmam21talkarticles00:43, 10 October 2005 (UTC)
version shows differently in a diff than on direct viewing?
I've found something rather unusual; I don't know if it's a bug or something I'm not understanding, but it's a case where the diff between two versions shows one of them with different content than that version seems to have when examined directly.
The diff is [7]; it shows that the previous of the two versions has only the "Frey" in "Allan H. Frey" wikilinked. However, looking at that version directly [8] shows, in every way I can think of to test, the whole of "Allan H. Frey" wikilinked. Any logical explanation for this? -- Antaeus Feldspar22:22, 9 October 2005 (UTC)
Never mind; I figured it out. The first link of "Allan H. Frey" didn't change; it was an unnecessary link to "Frey" later in the article that became a redundant link to "Allan H. Frey". -- Antaeus Feldspar22:28, 9 October 2005 (UTC)
Using defined templates to build a list
Once people start defining entries using templates, is it possible to use those templates to utilize the defined data to extract lists from all the articles?
For example, I developed a list of Soviet/Russian air bases on my site ([9]) and am mulling over how/if to put them on Wikipedia. I put a prototype page here: Tiksi (aerodrome), which uses my MilitaryBases template. Now is it possible to construct a page that scavenges all of the defined airbase data to show a master list like:
NAME LATITUDE LONGITUDE ELEV OPERATOR
Moscow/Air Base 1 57 04'N 37'05E 439 436 IAP (MiG-25)
Vladivostok/Some Air Base 52 44'N 144'02E 124 113 TBAP (Tu-22)
Obviously I am worried that the usefulness of the data will diminish if it is placed into a format where the data cannot easily be extracted; i.e. where someone has to visit each and every article to get all the data for a region. Any thoughts?
— Preceding unsigned comment added by Timvasquez (talk • contribs) 11:55, 9 October 2005
Your template is identifying each piece of data. Wikipedia itself does not have a tool for extracting fields from article data, but someone could make software (a "bot") which manipulates the data. At present, obviously each article would have to be accessed to extract the data. One of the advantages of templates is that the information has been identified and future software can further manipulate it. (SEWilco21:51, 9 October 2005 (UTC))
The way some templates are being used suggests that, longer-term, something like XML might be useful to think about. (Not that I know too much about it.) Rd23223:05, 9 October 2005 (UTC)
There have been several proposals for storing data items individually, so such data as country capital cities and populations can be stored in a single location. Something like XML would be useful if the data is stored within an article, rather than stored in a data space. (SEWilco19:35, 10 October 2005 (UTC))
Printable page
When using the 'Printable page' function, all links are rendered fully, so that no (useful) information is lost. Unfortunately, no distinction is made between links in the article itself, and links added to the article by a template. This causes links such as edit-links to mess up the layout of the page. See Liberalism (printable) for an example.
Shouldn't links in templates be left as mere text? Or should an option, similar to 'includeonly', be available to prevent the 'printable page' displaying useless links? -- Ec561812:06, 9 October 2005 (UTC)
Note that the URLs (internal and external) don't appear in the classic skin, but (as you say) do appear in monobook. Yes, printonly and screenonly (and maybe voiceonly) sound like good ideas. In the meantime, we could suppress that "edit this box" using CSS. -- Finlay McWalter | Talk12:45, 9 October 2005 (UTC)
The printable version is generated merely by a different stylesheet no changes in the html itself. There is a class called noprint for stuff you wan't to hide when using the print stylesheet which i have added to the sup tag surrounding that edit link. Plugwash21:44, 9 October 2005 (UTC)
Web TV
I just want to point out that Wikipedia on Web TV looks really really bad. It's all I have available right now, and it's very painful not to have a text-only option. I expect the same applies when trying to use it on a mobile device. The toolbar is somewhere off the end of the screen. The logo is not transparent. The background is an awful grey colour (oops - I'm in America- I meant gray color). The worst part is the tabs at the top. I only see the ones up to edit. Everything after that is missing off the end of the screen and not accessible (and access keys don't work, so I can't get to them at all). Is this something fixable in the css, or is a new skin needed (I think there was a new skin for mobile devices, but unless that is given automatically to people using those browsers, it won't be very useful to non-editors)? Angela 23:47, October 8, 2005 (UTC)
try the "chick" skin it does do away with all the shit round the side of the page and gives you simple text navigation links below the page. You still get any images in the wikitext itself ofc but i can't see why theese would be too much of a problem. Plugwash04:01, 9 October 2005 (UTC)
Hmm, that's a good point. We could tweak mediawiki to select a specific skin (for anon users) based on the HTTP GET's user-agent field. I've no idea, however, how we'd let the squids know about the difference. We'd also need a list of mobile user agents, and even then we'd get issues (as some will pretent to be IE for the sake of stupid websites that complain when you use anything else). -- Finlay McWalter | Talk12:54, 9 October 2005 (UTC)
Yes, the chick skin might work, but going to an URL with ?useskin=chick after it works only for that page and is forgotten when you click the next link (not useful on an unsecure connection where you don't want to log in). I tried going to a "detect my browser" site and it did claim I was using IE 5. Angela.23:10, 10 October 2005 (UTC)
According to older docs WebTV classic is supposed to identify itself as webtv, and MS's rather outdated docs for MSNtv (http://developer.msntv.com/Develop/browser_sniffing.asp) say the same. The old WebTV classic browser was based on Spyglass, but I guess you probably are using IE5 on a new box. If they're now pretending to be a regular PC, there's not much we can (practically) do to detect them. -- Finlay McWalter | Talk23:45, 10 October 2005 (UTC)
vortexing
First of all I'd like to thank you for setting up and maintaining such an informative site.
My thesis work has led me to the problem of trying to define the reaction kinetics of a 10 um particle with small molecules in solution during vortexing. Have you characterized particle behavior in solution for your vortexers? I'm trying to validate the benefits of vortexing but, I can't come up with data, or a mathmatical model, to make this claim.
I guess vortexing would cause the solution to become homogenous. With all particles easily accessible the reaction rate should improve. However, that would depend on the concentration of particles you have. N was right, please visit the reference desk for further questions. - Mgm|(talk)21:17, 8 October 2005 (UTC)
I copied some interesting code into my custom monobook.css file. I've changed to a serif font and made rounded corners, and this makes the "edit this page" tab look a bit odd, a problem which could be easily solved if I could manage to remove the bolding from words "edit this page". What can I add to monobook.css that will make it unbolded? — Ливай | ☺20:11, 7 October 2005 (UTC)
Yes, I know how to use :Cat It's not in the code on the page, according to the export function. It's protected or special, I couldn't edit it even if I did find it. SchmuckyTheCat21:41, 7 October 2005 (UTC)
Suggestion for cutting down on vandalism
A tool which I hope is implemented in the future would be to require users to view the differences made by the last editor of an article when submitting a new edit. Thus, everyone becomes a member of RC patrol, and much more vandalism will be caught. This could be as simple as sticking in the content that normally appears when you click "diff", above/below the editing box, so that the bright red text can easily be noticed. — BRIAN0918 • 2005-10-7 11:48
Not a bad plan, but make it a preference that can be turned on or off... I can see it confusing some of the truly new, especially if you combine it with the 'Automatic preview, Preview above edit box' options... nae'blis(talk)19:29, 7 October 2005 (UTC)
Good suggestion. I agree it should be optional. This would be much better than the current preview notice that asks everyone to check what they're about to save for previous vandalism, without showing the diff -- this might make newcomers feel as if they're responsible for everything they save, not just their changes, and could thus deter editing. Is/was there a discussion about this somewhere? How/where can I propose to change that? Joriki04:25, 8 October 2005 (UTC)
Yup, this new version of the preview notice just takes a simple warning note and turns it into a long notice that most editors won't even know how to respond to. Where do we change it back? —MichaelZ. 2005-10-10 17:27 Z
Excellent! Thank you!! Recently I caught a vandal making junk pages outside of the main article space; this will help find them. Antandrus (talk)15:44, 8 October 2005 (UTC)
I have three different Indian rivers with exactly the same name. Is it necessary that they be distinguished from each other in name? In ditionaries/encyclopaedias such entries can appear successively without any problem. Is that possible in Wiki?
As to what to call the pages (whether to use commas or parentheses) most countries have specific rules for this. Not sure whether India has, but Wikipedia:Naming conventions would be a good place to look before naming the pages in case it does. Some countries (like the UK and New Zealand) use the comma for rivers, others (like the US) only use the comma for towns, and use parentheses for rivers. Grutness...wha?23:35, 6 October 2005 (UTC)
I've been working on a List of Formula One drivers (here), which has a lot more information than the current version (here). I'm only up to C (179 of 783 drivers, approx 20%), and I've reached the 32kb limit mentioned on Wikipedia:Article size. Is there a policy/standard procedure for splitting up lists into different pages? AlbinoMonkey(Talk)11:53, 6 October 2005 (UTC)
Don't worry, the largest page I ever made was 200kb, but if it does get crazy then just split it either into 26 different pages for each letter or into a few pages (i.e. A-C, D-F etc.) Martin12:11, 6 October 2005 (UTC)
The 32kb size is mainly a recommendation rather than a hard-and-fast rule, and is in place because a lot of older web siftware can only handle 32kb in its editing windows. So the main thing is to make sure that no section of your file is over 32kb. Since you've broken this table into sections alphabetically, there should be no problem with that, as each section is editable separately (you should make sure, of course, that no section is over 32kb, although I doubt any will be). It's probably worth noting that the page you're looking at now is - at the time I'm typing this - 51 kb long. Some pages can grow well past 32kb - the WikiProject Stub sorting pages i'm most imvolved with often get close to 200kb, for instance, and there are quite a few article pages that are beyond 32 kb (List of interesting or unusual place names, for instance, and List of volcanoes). The largest lsit I've seen - one that thankfully has now been shrunk to a shadow of its former size - was 120 alarmingly confused and virtally incomprehensible kilobytes of Places in Afghanistan. It took an army of wiki editors several weeks to get that into shape, a job that would have made Hercules run away. If you want to see something frightening, have a look at some of the early versions of that file from late December last year! Grutness...wha?12:44, 6 October 2005 (UTC)
Monobook.js changes?
I don't see any recent changes in MediaWiki:Monobook.js or in my Monbook.js, but I now have Edit this page, baby! on categories, and the Save Page link is now bold. Has anyone else noticed this? If so, where were the changes made? ∞Who?¿?08:36, 6 October 2005 (UTC)
Actually I did check there, I was just looking for the edit tab changes and didn't even notice the edit summary. D'oh!. Thanks. ∞Who?¿?11:29, 6 October 2005 (UTC)
The language used
You have two ways of translating opinions in serbian language:Serbian and Serbocroatian.
Since Croats have had the gift of having their first grammar in the year 1604 by Barol Kašić,wouldn't be fair that this great Wikipedia has translationes in CROATIAN language?
Who decides about this anyway?
Lately my third-floor apartment has been infiltrated by roughly six or so mice, which are a wily bunch to capture and banish.
The mice surpirisingly appeared in bulk on the day that a departing roommate vacated. He had a strange preference to store his food in his closet, from which by chance the mice emanated.
I've bought mousetraps (after much searching) and they have gone un-moused for several days.
My real issue does not include the mice, but what they bring along with them. The spiders, flies and crawling bugs I can deal with by conventional methods, which I am well-versed. The big problem, and the most irritating, is the 'invisible' silk-like strands of whatever which are attracted to the body, especially to the hair. From the head's hair the 'whatevers' are drawn to any skin exposed, and bug you!
-- What the hell are they?
-- How the hell do you get rid of them? I have tried disinfecting spray, Febreze, leaning over a lit gas stove, smoke . . . and they're still here.
What do I do before I commit myself?
You'll probably get a better response to your question at the Reference Desk pages. (That said, I'd suggest calling your local public health/pest control department or the like - whilst they probably won't get rid of them for you, they might well know what they are) Shimgray | talk | 00:47, 5 October 2005 (UTC)
They sound like spider webs to me. Or, more specifically, the same type of threads which spiders use to make webs. They also use them as support lines for webs and to lower themselves from the ceiling safely. They are quite annoying when they get in your face and are invisible unless they catch bright light just right. They usually last longer in areas with less traffic, like corners. Using a dust brush in the corners will normally remove them. To stop new ones, you need to kill off all the spiders. Not all spiders leave such threads, but I don't see any reason not to kill them all and ask questions later, LOL. StuRat11:04, 5 October 2005 (UTC)
But the spiders eat the flies, so it's best to lay'em alone. Upset the delicate ecosystem...
We all know that some anonymous IP users cause the vast majority of Wikipedia vandalism, fortunately most can be blocked quite easily, some however cannot. As far as I understand dynamic IP userss (AOL etc.) change their ip very frequently and thus cannot be blocked.
Would it be technically possible to make it such that users who edit from this type of IP address are only allowed to edit if they register? Meaning that we could block them properly, while allowing others to edit without the fear of being blocked by accident. Martin08:51, 4 October 2005 (UTC)
Please give me the idea behind the logic of designing conversion formula. For eg:
There were 2 systems A n B.14 degrees in A is equivalent to 36 in system B and 133 in A is equivalent to 87 in B. now what is the temperature where they both r equal?
Ans:51.25 (conversion A=(7/3)B-70)
I'd like to know how to design such conversion formula.
Three times in the last couple of weeks I have made fairly extensive edits to articles on my watchlist, only to have the edits disappear entirely -- they're not in the page history or my contributions list. One of them was 90 minutes worth of work, so I was more than a little perturbed.
After the first time it happened (to William Randolph), I've been paranoid about making sure articles have actually saved and not just previewed (I do not have automatic preview on save set in my preferences). The last time (after working on Duran Duran), I checked the edit history after saving, and there was my edit at the top of the list. But when I looked at it again a few days later -- the next time the article popped up on my watchlist -- all traces of my changes were gone from both the article and the history.
I thought it might be a lag issue with caches and such, but the edits have not miraculously come back yet.
The history shows you making several edits to Duran Duran, last on September 18 (summary "/* Singles */ - + club play"). Can you confirm that the lost edit was after that, and if so can you perhaps narrow it down to some time range, contents, summary, anything you can remember that might help in tracking down traces in the database? Same of any other lost edits that you're sure about having saved.
A couple times people have had the history page get stuck in their browser cache somehow, so it wasn't showing later edits for them (including their own recent edit), but if you stick some extra parameters on it for instance this should then show the correct current history. Can you confirm if this does or doesn't help here? --Brion05:20, 4 October 2005 (UTC)
Duran Duran was definitely after the 18th -- looks like either the 23rd or 24th, after this edit -- I fixed all the non-MOS headers, and made about half a dozen other changes.
It looks like I was working on William Randolph on the 12th, because I made related minor edits to William Berkeley, Tuckahoe, and Isham Randolph that day. That's the one where I lost a lot of research and external links; 'twas annoying because I was responding to a request on my talk page. Anyway, wanted to see if it had happened to anyone else before I bothered a developer with it (since despite my paranoia it could still have been my fault in some way) -- I thank you for your attention, Brion. — Catherine\talk03:20, 5 October 2005 (UTC)
Some general suggestions to avoid losing work:
Edit in smaller pieces. A 90 minute edit means you could lose it all if your computer goes down, you lose your internet connection, or there is a Wiki problem. A series of 5 minute edits reduces your risk.
Cut-and-paste revised text to your own computer as a backup, before attempting to save it to Wiki, as this is when it is likely to be lost.
Can anyone please give me an estimate of how many updates are made to wikipedia per day or month? Also, can anyone provide any information about the average bandwidth usage of wikipedia? Thanks for any help.
Rob
I'm not sure where to report this, so I'm trying here. In the last two or three days, three e-mails that I've sent to editors via Wikipedia have not arrived. Has anyone else noticed a problem recently with this? SlimVirgin(talk)07:27, 3 October 2005 (UTC)
I haven't but Ludraman wouldn't complain about not receiving the message I sent him if he doesn't know about it. Are you sure the receipient's spam filter isn't throwing it into the trash? By the way, Raul managed to sent me an email with standard Wikipedia title a few days ago, so it seems to work. - Mgm|(talk)12:45, 3 October 2005 (UTC)
I was recently posting to User:Encephalon's talk page, and had in my post the tags {{tl|PD}} and {{tl|PD-USGov}}. When I went to post, it completely destroyed the page's formatting and actually removed previous posts (see these two diffs, [11] and [12] (I tried it twice)). I eventually figured out what the problem was, and that I could bypass it by changing {{tl|PD}} to [[Template:PD]] (see this diff, here). Now, my question is, why is this happening? I've used this tag many times before and with no problems, and would like to still be able to do so. Thanks in advance for the help, --Blackcap | talk20:36, 2 October 2005 (UTC)
Recently, a reporter spent time reading the Encyclopedia Britannica from A - Z, and then reported on it. I was wondering, is there a mechanism for reading Wikipedia from A - Zzyzx, California (the alphabetically closest to the end I could find) in the same manner? Does Wikipedia have a feature which categorizes articles alphabetically? Yes, I realize that in the months/years reading this encyclopedia from A - Z, many articles will change dramatically, but I am more interested in the process than in the actuality of completion.
It's a pretty Sisyphean task, but have a look at Wikipedia:Quick index, which dynamically generates alphabetical subpages. However, this includes redirects (alternate names for articles), which means it won't in practice be alphabetical. Shimgray | talk | 00:53, 2 October 2005 (UTC)
Cool! Thank you both. I (am thinking about trying/wish to attempt/am in awe of anyone who could complete) this feast, dispite my mention to the contrary, yet, I'll bet I don't get half-way the letter "A"
I'd be suprised if you got as far as the letter A! - Our 753,857 articles currently start at ! and end at ﹀, if you restict yourself to articles that contain only un-accented Latin-alphabetic characters, you'll still have to read from A to Zzzax. Allowing all alphanumeric charcaters that are not redirects the range is 0 to 鬼葬. Thryduulf12:03, 2 October 2005 (UTC)
Of course, the real problem is that once you've finished you'll have to start over again because so much content will have changed! Dunc|☺21:33, 2 October 2005 (UTC)
Indeed, in the approximately 10 hours since my previous post, 865 new articles/redirects have been created. Thryduulf21:55, 2 October 2005 (UTC)
If you read at 1 page per 30s for 20 hours a day then it will take you round about a year to complete.--Clawed22:14, 2 October 2005 (UTC)
Go to that article and click edit. Most people seem to get the "WARNING: You are editing an out-of-date revision of this page. If you save it, any changes made since this revision will be removed." Someone has had to reconstruct the page several times, I don't know if it's grabbing random bits from history or what. Check history for the corruption fixes. --Golbez20:09, 1 October 2005 (UTC)
I've fixed it in this case and I'm running a search for any more such problems. Tell me if you see it again, it's quite serious. -- Tim Starling00:12, 2 October 2005 (UTC)
Strange article problem
...as in a strange problem, not a strange article. University of Otago opens fine in Safari and Mozilla, but crashes IE every time I try to open it with that. This isn't a new problem - it occurred a few months back as well, but that was during the upgrade to a newer version of MW, so I thought it probably had something to do with that. Quite clearly, though, it doesn't. IE works fine with every other article I've tried, but it crashed three times in a row on this one. Any thoughts? Grutness...wha?10:41, 1 October 2005 (UTC)
Not really... I'm using 5.2.1, with a Mac (not a particularly viable combination, which may be the problem). It's odd that it's just that one article though. Grutness...wha?11:07, 1 October 2005 (UTC)
Also fine for me (IE6.0.2900/XP-Prof SP2) and mine chokes on just about anything even slightly funky. Have you checked your PNG renderer? —Phil | Talk13:42, 4 October 2005 (UTC)
}
</css>
I've installed an extension on the Wikimedia cluster which allows per-page CSS using <css>your css here</css>, my rationale for doing this is that it would allow a much cleaner implementation of pages like the Main Page which currently use a lot of inline styling, additionally you can specify @media print, projection, screen... inside <css></css> so you can have different styles for normal views and printouts amongst other things, finally this should hopefully reduce the amount of things that in MediaWiki:Common.css and MediaWiki:Monobook.css that aren't used globally but are just there because of the limitations with inline styles stated above. —Ævar Arnfjörð Bjarmason16:24, 30 September 2005 (UTC)
It escapes the data sent to it so that for example doing something like <css></style><script type="text/javascript">malicious code here</script></css> isn't possible, however putting invalid CSS which wouldn't validate as such is possible, however that's pretty much a non-issue since you can invalidate the XHTML output with various evil tricks anyway. –Ævar Arnfjörð Bjarmason16:50, 30 September 2005 (UTC)
Ævar, could it be that it's only installed on some servers? I've seen your example above work once, but not now (even in a different browser, when logged out, or logged in under a fresh test account). Bypassing the cache didn't help. -- Finlay McWalter | Talk18:02, 30 September 2005 (UTC)
Ævar please don't delete discussions, if you've added a new feature and then disabled it again for security reasons please give details don't just try to cover up that it ever existed. Plugwash19:39, 30 September 2005 (UTC)
I removed the discussion because I didn't think an announcement of something being installed which isn't installed anymore was very useful, I wasn't trying to cover anything up, you'll notice that it's still in the server admin log as well as still being in CVS, I would probably have removed it from both if my intention was to "cover up that it ever existed.", don't be so paranoid. —Ævar Arnfjörð Bjarmason21:02, 30 September 2005 (UTC)
Web Services
Does Wikipedia offer Web Services? If so, how can I invoke them via .NET?
Using and attributing non-image media. How is it done?
Let's say I've got some sounds to upload, how do we store copyright info, etc, about them? It doesn't seem to have an image description page.
How can the media file be imbedded in the page in a way that displays this information? Apart from directly linking to Media:Example.ogg? is there a standard way of doing it with a thumbnail type thingy? Should I enter a bug report to request this? Dunc|☺11:10, 30 September 2005 (UTC)
I just looked at the few images that I've contributed, and I find that now there is metadata there, and you can tell how cheap my camera is by how ugly the metadata. Gak! Is there any way to improve it? It's really horrible. The images in question are:
Photoshop has limited metadata editing capabilities, and when you use the save for web function it removes all metadata from the file. Though this probably isn't what you want to hear if you don't already own it as it is very expensive. Martyman03:50, 30 September 2005 (UTC)
GraphicConverter can do some of that too, and it's much cheaper (see [www.lemkesoft.de] — Mac only (X and earlier). But I'm wondering how I can get to the metadata on the images I've already uploaded to clean it up. (And on my Mac with GC the metadata looks OK. The "M+M+M+M+..." business is in the comment section, BTW.) ;Bear15:42, 30 September 2005 (UTC)
I fooled around with jhead which allows one to remove the comment and/or EXIF sections from JPEG files. So that's the brute-force way of fixing your problem, I suppose. -- Finlay McWalter | Talk17:27, 30 September 2005 (UTC)
Currently it's set on 180px, you can change your preferred size in user preferences, but there are two problems with this.
People don't realise that you can set the size in their preferences and so force them to be larger, which defeats the point of having a small thumb size, and the ability to change it in user preferences.
Most users are going to be using large screens. It is worth having to bother with some slightly ugly looking page formatting issues just for them.
I think an increase of only 20px to 200px brings quite good effect. 250px might be a little large, though it's what I've got mine set on. What do other people think? Who is using a diddy browser screen? What does it look like to you? Dunc|☺15:51, 29 September 2005 (UTC)
On a related note it would be nice to be able to specify sizing relative to a users preffered size. Some images need to be bigger than others for various reasons. Plugwash17:35, 29 September 2005 (UTC)
We're seriously considering removing the user preference option to change the thumbnail size, it impacts on our bottom line fairly heavily. Considering that, changing the default to something that suits everyone is probably a good idea. -- Tim Starling17:40, 29 September 2005 (UTC)
Why is this a big deal? It seems like when something is uploaded you should create and cache a selection of default sizes and then just ensure the right one is delivered upon request. Aside from the initial step it wouldn't be computationally intensive and I would be surprised if such thumbs are creating a serious hard disk problem? Is it that you are presently creating thumbs on demand? Dragons flight18:09, 29 September 2005 (UTC)
Well, it multiplies the resource requirements of everything. If we have three settings, and people use them, then that triples the amount of CPU time, disk space, and cache space required. Options that affect cachability of pages affect all pages, not just those which actually contain images, so having say three popular thumb sizes cuts the amount of stuff we can keep cached by two thirds. --06:10, 30 September 2005 (UTC)
Does anyone know what I think is called the hexadecimal code (i.e. #FFFFFF is white) for the standard wikilink colour, i.e. this: --HighHopes(T)⋅(+)⋅(C)⋅(E)⋅(P)10:17, 29 September 2005 (UTC) Your help is much appreciated. Thanks!
The articles already mis-sorted won't be fixed in the category listing until an edit (even a null edit) is made to them. A null edit is "edit, then save" (with no summary and no changes). -- Rick Block (talk) 01:12, 29 September 2005 (UTC)
Optional template parameters
I'm sure someone mentioned adding theese but i can't find any reference to how they are used can anyone gimme a hint? Plugwash12:38, 27 September 2005 (UTC)
And, if you're actually interested in optional parameters, they aren't supported by the existing template mechanism. Although a variety of folks have figured out extremely clever workarounds, I'd advise waiting until there's real support for this in the template implementation. This is an existing enhancement request, see bugzilla:364. -- Rick Block (talk) 01:26, 29 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)
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.
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 Seekerদ08:39, 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. --cesarb13:28, 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. gkhan04:28, 25 September 2005 (UTC)
Also remember if you are actually logging in from work your employer can of course packet sniff you and get your wiki password. Plugwash18:56, 1 October 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:
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? Martin19:51, 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)
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. (SEWilco20:08, 20 September 2005 (UTC))
It broke half of the infobox templates at Polish Wikipedia... Ausir
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}}}}} --Brion20:59, 20 September 2005 (UTC)
Any idea if/when we will get this back it would save a huge amount of pain when you wan't to change something in a template for one case without having to go round editing all the pages using it or substing the template. Plugwash18:51, 1 October 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 - Martin20:47, 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 - Xed20:53, 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.--81.42.154.22513:43, 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\talk02:14, 22 September 2005 (UTC)
It gets a bit more complicated when you start looking at redirects, categories, etc. If you include everything, that's fine, but otherwise you'd have to have some way to parse it all. --Blackcap | talk20:39, 2 October 2005 (UTC)
Images and edit section fields
At Alliance 90/The Greens, I put some images into a row in the right side of the article. This results in all the edit fields of sections following the images (and displayed left to theme) now are stacked left to the last image in the middle of the text. Known bug? Is there some workaround? -- till we ☼☽ | Talk22:40, 23 October 2005 (UTC)
Back button not working
Has anyone else noticed this. Usually happens when if i'm trying to get back to the Watchlist. Just started in the last day or so. I'm using Firefox on Linux. josh19:03, 23 October 2005 (UTC)
Sometimes a raw category page doesn't give enough information about the articles in a category to know if they're worth clicking on or how they are related. For instance -- you wouldn't have any idea what "Key System" means on Category:Streetcars_in_North_America Is there any way to add text to a category page for a particular item in the category? -Archier18:00, 23 October 2005 (UTC)
Forgive me if this has been raised already, but I'm increasingly getting 404s on many different machines whenever I try to create a new page. What could be the matter? JFW | T@lk11:12, 23 October 2005 (UTC)
Redirects to special pages and other wikis have been disabled since they're much harder to revert than normal redirects and commonly used for vandalism. Angela.00:35, 23 October 2005 (UTC)
Hello again. I asked a question above, and am grateful for the answer I got there-- to follow the thread:
When publishing an article in a book, or other printed publication, is it necessary to list all contributors? In many articles, there are 2-3 people that have made 90% of all edits. Would it be sufficient to just mention those? And can users only making minor edits be disregarded? I am referring to the legal issues here, not just recommendations/suggestions. // Thanks, Fred-Chess11:32, 22 October 2005 (UTC)
Section 4b reads:
List on the Title Page, as authors, one or more persons or entities responsible for authorship of the modifications in the Modified Version, together with at least five of the principal authors of the Document (all of its principal authors, if it has fewer than five), unless they release you from this requirement.
So I personally take it to mean you need to credit at least five people. Ink and paper being cheap, I don't see why you wouldn't credit everyone. That's my recommendation/suggestion - but if you want a definiative answer, you'll need to hire a lawyer. The technical matter of figuring out who those five authors are, and rejecting any contributors whose edits fall below the threshhold of copyrightablility (reverted vandalism, fixing a simple typo, etc.) is not a problem that the mediawiki/wikimedia community has chosen to solve for you. -- Finlay McWalter | Talk11:41, 22 October 2005 (UTC)
It appears that if you don't want the link to activate, you format it [[:ko:Municipality of Strathfield|Korean version of article]]. You had [[ko:Municipality of Strathfield|Korean version of article]] Superm401 | Talk01:37, 22 October 2005 (UTC)
For all those cases when you only want the link to appear under "In other languages" and not explisit in the text, maybe? Shanes23:50, 22 October 2005 (UTC)
You're confusing interwiki links and interlanguage links. Interwiki links link to other wikis as part of the text. Interlanguage links link to the same article on another language edition of the same wiki and appear in the sidebar (on the default skin). If you want an interwiki link to the Korean Wikipedia, use :ko:article or w:ko:article (or [[wikipedia:ko:article]] from most non-Wikipedia wikis), not just [[ko:article]]. Angela.00:18, 23 October 2005 (UTC)
Note the discrepancy between the first field and the second field. It seems $4 is interpreted as a parameter and substituted with the empty string.
I think this is just a display issue, the $4 appears to be correctly stored in the database, nevertheless, it displays incorrectly in Special:Log/block: [14] and in Special:Contributions (here) (not that there are any contributions for this user, or ever will be). Note mouseover shows that it was indeed the correct $4 version that was blocked.
I haven't checked if this applies to articles with a $+digit, or other places where the display discrepancy occurs other than block log and contributions history.
Wow. You're right. That's a disturbing bug. What you described is bad enough. However, when there's a username with just $, then a number, like "$4", it doesn't show up at all in many places.
Block log fixed (and in general, most such that didn't save bogus version of text to db should be fixed now). New user log recorded the bogus text to the database at the time, but any new ones should be correct (eg [15]). --Brion23:10, 22 October 2005 (UTC)
Furthermore, this does apply to page names as well. I created a sandbox page called User:Superm401/Sandbox/$5Test. When you edit it, for example, the $5 doesn't show up in the Editing ... header. I'm sure the article name version of the problem extends to many other places as well. The best solution is probably to ban it in article and usernames. Several other characters are already banned and people make do. See Wikipedia:technical limitations. Superm401 | Talk21:25, 22 October 2005 (UTC)
As you found out the hard way when you created User:$4, any accounts with $+digit are now getting automatically blocked on creation. The dollar sign followed by a non-digit should still be OK. -- Curps21:32, 22 October 2005 (UTC)
Works fine; presumably the fix above fixed that as well. Please don't run around declaring entire categories of things to be banned for technical reasons just because nobody bothered to tell us there was a bug so we could fix it. --Brion23:27, 22 October 2005 (UTC)
This is identical to a bug with a dollar sign in the title of an article. This comes up quite often in WikiNews. I was told they intentionally strip out the dollar sign due to some security concern. However, the inconsistent way it is stripped out, taking the following character with it, is certainly not intentional. StuRat21:56, 22 October 2005 (UTC)
Is there a way I can make an RSS feed for my watchlist? Sorry if this has been asked before, I couldn't find anything. It would be a really nice feature. Pojo19:12, 21 October 2005 (UTC)
I've got some template coding issues that I hope someone can help me with, or direct me to where I should ask if this is not the right place. First question; are recursive template calls possible – e.g. can a template call itself with different arguments based on which link inside the template you click? How would you code it? Typically the template would be a front page menu, and you want to stay on the same page even after selecting some menu options. Say, by default the template {{menu}} is called without arguments and it displays some text and some links, and two of the links are something like [[{{menu|1}}]] and [[{{menu|2}}]] (I know the syntax is wrong, but bear with me), which conceivably leads back to the same URL but with different behaviours, i.e. the text displayed is different depending on the arguments 1 and 2. Second question; are there other, simpler, ways to solve such a front page menu task? Thanks a lot. --Eddi(Talk)09:43, 21 October 2005 (UTC)
To make a template do something different, you'd have to pass it something which is different between two "same URL"s. (SEWilco12:29, 21 October 2005 (UTC))
As far as I read the guidelines they recommend against templates calling other templates in cases where the calling page and/or the called page are frequently edited and/or they are linked by many other pages. This may or may not be the case, depending on what is the "front page" as I call it. I'm basically looking for the right syntax to achieve the task but still not be "too recursive". – Or a completely different approach altogether providing the described menu behaviour. Hope I'm not too greedy... --Eddi(Talk)21:33, 21 October 2005 (UTC)
Crediting the author of articles
Hi.
According to recommendations one should mention five major editors of an article when one is to print it or reproduce the article elsewhere. My question is how to value major contributors since this is hardly practically possible in most cases?
I am requesting the log pages such as Block Log, New Users and Recent Changes be made available in XML format. This not only reduces server load but also is much easier for programs to interpret than standard HTML. 67.60.52.15514:53, 20 October 2005 (UTC)
You haven't blocked images from this server have you? preferences|web features|images|exceptions. It's easy to do by accident Ojw21:14, 21 October 2005 (UTC)
Template Help Needed
I run a wiki at www.youthrights.net, one aspect of that project is a law library for laws that affect youth. It is organized by state and by issue/law. I'm trying to use a template to better organize the index/table of contents page in the law library.
The issue is that I'd like to use that same template for all the various issues/laws the law library will cover, and use variables to customize it for each. So I'd like the template to have the state name (i.e. Missouri) as standard in the template, and the issue be a variable (i.e. boating age) that can be customized per page. Its tricky in that each of these is also an internal link.
I'm probably not being clear, but that's just how clueless I am on this, I don't even know the proper language to describe it.
This is basically what I wanted to do, but it doesn't work:
[[Missouri {{variable like Boating Age, or Curfew}}]]
Anyone have a clue what I'm talking about and how to do this?
See any of our city infoboxes, but the basic idea is: you call the template like {{Missouri|age=18|curfew=none}}, and on the template you use {{age}} and {{curfew}} to get the value of the parameters. --cesarb00:49, 20 October 2005 (UTC)
(edit conflict, what I said is mostly duplicated) Yes. What you need to do is make the parts of the template you're going to customise be filled only with {{{<name>}}}, where <name> is a tag you can call.
For example if my template "Ingredients" contains {{{chocolate}}} and {{{coffee}}} I can then call them like this...:
Please check out those links, cause you can see the level I'm at I guess and what I want to plug these things into. I know how to call a variable from the template, but I want to drop that variable in a link. For example I want [[Missouri {{variable like Boating Age, or Curfew}}]] as an item within the called template, to function just like a regular link such as [[Missouri Boating Age]]. Does that make any sense? - KPalicz01:26, 20 October 2005 (UTC)
Ah, kick ass, I figured it out. Go here to see the template, and here to see it used. The error (seemingly) on the template page is what threw me off. I thought it wasn't working the whole time, when it really was. Damn I'm dumb. KPalicz01:45, 20 October 2005 (UTC)
It may be a factor that templates use two braces, {{ }}, while template variables use three braces, {{{ }}}. (SEWilco03:35, 20 October 2005 (UTC))
Blocks
I have no idea if I have put this in the right place, but I was wondering what the point is in not being able to edit articles when logged in, if the IP address you are using is blocked. Is it because of technical limitations? FireFox18:07, 19 October 2005 (UTC)
Template behavior changes: Default and arg transclusion
Someone named 'avar' in #mediawiki says that we now support syntax like {{{arg|default}}} and {{template|arg = {{template2}} }} (SEWilco16:00, 19 October 2005 (UTC))
I haven't seen any details. For example, although {{{URL|{{{url}}}}}} allows both spellings for an arg I have not tested {{{URL|{{{url|{{{Url}}}}}}}}}} nor do I know how deeply it can be nested. (SEWilco19:09, 19 October 2005 (UTC))
IIRC the breakage was a bug; I'd suppose he has tested against the cases which broke last time before going live with it again. --cesarb00:54, 20 October 2005 (UTC)
Okey, what it does support now is {{{param|default if param is not specified}}} and {{template|paramater = {{woo, a template inside a paramater}}}}, and for the record, this code is not mine (unlike the previous one), which is why it's working;) —Ævar Arnfjörð Bjarmason17:47, 20 October 2005 (UTC)
For those interested this means that we can now have optional sections as well as optional values in templates using some hacking, for instance say you wanted to create a mointain template as I've done on User:Ævar Arnfjörð Bjarmason/Mountain and always wanted to specify the height and name of the mountain but not always the mointan range (as it might not be in one) then you might call it as:
{{User:Ævar Arnfjörð Bjarmason/Mountain
| name = Mount Everest
| height = 8844
| rangeoff =
}}
This was already possible using {{If defined call1}} (see Template talk:Peru region table for instance). However, it becomes now possible to have optional sections which won't appear if the parameter is undefined. This means you can introduce optional sections without having to modify the pages calling it. Demonstration :
{{User:Rorro/Mountain
| name = Mount Everest
| height = 8844
}}
yields:
Mount Everest
Height
8844
If you want to specify the range,
{{User:Rorro/Mountain
| name = Mount Everest
| height = 8844
| range = Himalaya
}}
Previously, a parameter value could contain (text)]](text)[[(text). Now, this prevents the template from being recognised in the code and the raw code is output. Is this fixable? _R_13:38, 21 October 2005 (UTC)
Nothing in what has been mentioned indicates such a change in behavior. Whatever changed should be documented. However, I don't know whether this change should be called a bug or a feature. (SEWilco19:51, 21 October 2005 (UTC))
I don't know either, but some templates on fr: relied on the old behaviour (and I bet some here do, too). That's the reason why I'm asking. _R_22:25, 21 October 2005 (UTC)
As we're discussing changes and announcements of changes, the changes to that page are in [16]. But I did not notice mention of changes affecting behavior of argument content. (SEWilco06:30, 22 October 2005 (UTC))
Alt language
This has happened to me in the past, and it is happening to me again now. When I hold my cursor over a link (either the ones on the left, or the tabs at the top of an article, or the ones at the top right corner) the alt descriptions appear in another language. I always have my language set to English, and never change it. It looks to me like Norwegian. Take a look. Does anyone else experience this? Is it a known issue? --Daniel Lawrence15:33, 19 October 2005 (UTC)
Is this only happening to me, or is it happening to other people? Every now and then I'm just being logged out automatically, and its starting to happen very frequently now. Is this some kind of bug? 213.122.126.15613:07, 19 October 2005 (UTC)
Probably not related, but no matter what permissions I allow for wikipedia.org in Zone Alarm, I can only stay logged in if I shut down Zone Alarm. Any recommendations from Zone Alarm users who don't have this problem would be welcomed. --Craig12:56, 27 October 2005 (UTC)
Vanishing rvv
I'm pretty sure I've had at least four rvv vanish today; I'll pay more attention to future ones. The updated page was redisplayed after Save, but the vandalized version still showed in "My watchlist" and "History". (SEWilco05:38, 19 October 2005 (UTC))
god mode lite
Not sure where to post this, but just to warn anyone who uses this script to be careful. I've been having a few problems with it exchanging characters on reverting (e.g. [17], [18], [19].) I'm not sure but it could be since I started using Firefox 1.5 Beta 2, so keep an eye out. I've notified the author Sam Hocevar. the wub "?!"23:28, 18 October 2005 (UTC)
Actually I think this is more the fault of MediaWiki 1.6 Alpha (or whatever it is we're running at the moment). All the browser does is send a username and a special ID and tells the server to replace that user's contributions string. I don't think it has any direct control over the content. GarrettTalk04:46, 19 October 2005 (UTC)
incredibly wide diff pages under firefox
I recently moved to firefox and found that the diff pages' width didn't adjust correctly: I have to scroll horizontally to see both the removals and the additions :( Anyone knows how to fix that? (Btw, that also sometimes happens with my personalized google homepage when it reloads automatically, if that's any help...). Thanks, Jules LT21:25, 18 October 2005 (UTC)
There is a fix for that by adding horizontal scrollbars to oversized sections. I don't have a link handy but it's in my user JS. You only need the "cesarb_fixDiffOverflowLoadListener" section. It works wonderfully, even enormous URLs won't make you scroll the whole page horizontally. GarrettTalk04:52, 19 October 2005 (UTC)
Except when it doesn't work, which causes one of the columns to get incredibly tiny. I am the one who created that function, and gave up using it because of it. If you find somewhere that bug happens, and manage to fix it, please tell me (some of the problematic diffs can be found at User talk:CesarB/monobook.js). Sometimes I wonder if I'm the only one to whom that function doesn't work well :( --cesarb17:09, 19 October 2005 (UTC)
Yes, I'd be glad to know too... Thanks anyway, Garrett; reloading hardly ever fixes it, but sometimes at least... Jules LT19:21, 19 October 2005 (UTC)
four tildes don't work
When I type the four tildes, it makes my name come out as SoothinR, instead of SoothingR and doesn't redirect to my profile page anymore. It worked a week ago, however suddenly stopped working today. Does anyone have an idea how this can be fixed? --SoothinR 20:49, 17 October 2005 (UTC)
Ive read the download wikipedia article in the help section and I don't understand exactly how you set it up. I wish to download the whole of wikipedia minus the images if possible so I can look through if offline. Could you possibly provide a step by step guide to downloading and installing wikipedia? I read about file 'dumps' and xml but am confused as to how it all fits together as the help section article seems to be oriented to people already familiar with that type of software and files.
Wikipedia is database, not set a regular web-pages, so just simple downloading (with offline browsers httrack, TeleportPro or similar) I suppose will not work. In short You have to:
where do i get WAMP/LAMP from? so i install that, then sql, then I download and install mediawiki. what im confused with is the next part. when i navigate to index of wikipedia it has a list of 30+ different files ranging in size from a few bytes to 30gigabytes. If i just want the whole list of wikipedia articles in english with no images is that where I download it? if so, which one is it?
You should read the section about this in Wikipedia. I am not going to seach for it, but there is a *complete manual* for it (I have seen it). It helps you more then we can help you (and wastes less of our time answering something that has already been answered). On top of that, I would not advise you trying to do it, since it is A. very technical B. very resouce intensive (not only gigabytes and megahertzes but also human time).--Msoos12:51, 21 October 2005 (UTC)
working toward watchlist rss feeds ( bugzilla #471 )
Have been attempting to dig a bit toward this as it seems long coming and increasingly important.
It's less restrictive than GFDL or cc-by-sa. cc-by allows someone downloading it to distribute altered versions without distributing them on the same terms as the original. If my understanding is correct, I could upload a cc-by image (it's one of the licences in the drop down box on the upload page). Someone else could then alter that, and they wouldn't be required to release it under the same licence. That doesn't appeal to me, but if planiglobe doesn't have a problem with it then I suppose we don't either. -- Finlay McWalter | Talk20:53, 16 October 2005 (UTC)
I plan to upload a bunch of images soon since I'm working on articles on each Metro station in Mexico City. Each station has its own logo and we are still missing many of them (~50) all of them fair use. I was wondering if there is a bulk upload tool for EN similar to the one someone created for Commons. Thanks! -- Rune Welsh | ταλκ | Esperanza19:52, 16 October 2005 (UTC)
I thought so, but still wanted to ask anyway. The thing is, having the logo of each station in an article is just as important as having the article itself (read Mexico City Metro to see why). As far as I know the designs belong to the city and there may be no easy way to get free versions of them. -- Rune Welsh | ταλκ | Esperanza00:15, 17 October 2005 (UTC)
Have you taken look at m:Pywikipedia. It's a bot, and it's upload.py program (if that's what you call it) could be helpful. The description for upload.py: "Script to upload images to wikipedia...". I'm pretty sure it will work for en, but not certain. Oh, and you don't need to be Boris Grishenko to get it going either.--Commander Keane07:12, 17 October 2005 (UTC)
There is a list of allowed bots in Wikipedia somewhere. I have seen it. There is one specifically for uploading a lot of images. The author(and user) of the bot said on his user page that he is willing to upload a lot of images if he is asked. Go, find it, it should be pretty fast (there are no more than about 20 bots allowed in wikipedia), and ask the author. He seemed pretty freindly on his talk page. --Msoos12:45, 21 October 2005 (UTC)
Well, just to support that I don't like those changes. It bothers me quite a lot, it kind of makes wikipedia outside of the "internet link font" Janarius13:25, 21 October 2005 (UTC)
Newbie error or vandal?
Someone from IP 129.234.4.76 (talk·contribs) (not me) just tried logging into Wikipedia using my username: I got bounced from the system (logged out) and the system has changed my password, and sent me an email informing me of this. What is going on? -- Francs200015:07, 16 October 2005 (UTC)
This happens all the time. Someone thinks they had a particular account but fail to log in so of course request a password. The new password is sent to the owner which isn't always them. Both passwords, old and new, should now work when you go to log in, unlike some forums where the old password is deleted upon request of a new one. So don't worry, your account is still safe. Whew. :) GarrettTalk16:25, 16 October 2005 (UTC)
I agree. No amount of tinkering in MediaWiki:Subject can fix it. It's specifically centered around something strange with the CSS labels of (span id) "wpSummaryLabel" and (label for) "wpSummary" and maybe the (div class) "editOptions". --AllyUnion(talk)10:06, 16 October 2005 (UTC)
Redirecting a page is fundamentally a different concept than editing it... actually, it's exactly like a page move, except the "page move" target already exists. It should be logged in a different way than ordinary edits, more like the way Pagemoves are logged (in Special:Log/move and in Special:Recentchanges).
The question is, is this technically feasible? When does Mediawiki recognize that a page is a redirect... when it is saved after an edit, or only later when it is clicked on?
Redirecting a page is exactly like editing it; it *is* editing it. Previous versions are still in the history at the original title, etc. --Brion04:26, 16 October 2005 (UTC)
Apparently (some?) users have access to other users' edit histories. Is this privileged, or may any-one see other users' edit histories?
All edit histories are publicly visible. For the edit history of an article, click on the history tab at the top of the page. For a user's contributions, go to their user page and click user contributions in the toolbox at the bottom-left of the screen. --fvw*06:29, 15 October 2005 (UTC)
Images/media and red links.
Is there a reason why text links (the prefixed with a colon kind) to the Image: namespace don't produce red links? Sure you can use Media: links but why the distinction? Image:SomeRandomFilename.jpeg looks just fine while Media:SomeRandomFilename.jpeg does result in a red link? The main reason it's bugging me is that the watchlist use the Image: kind of link so it's hard to weed out deleted images when editing the watchlist, and also image links on various maintainance lists and talk pages also use the Image rather tha Media "namespace" causing you do end up on non-exsisting pages quite often (closed IFD pages also look "wierd" without any red links). --Sherool02:30, 15 October 2005 (UTC)