Wikipedia:Village pump (technical)/Archive 19

Getting rid of unwanted watchlist item.

(Copied from Wikipedia:Help desk)

The last entry on my watchlist is:

Wikipedia:SOURCE

It is marked as a dead link, but it isn't, of course. When I look at the redirect page, it does not have an "unwatch" tab. When I check it and click on "remove titles" it still remains on my list.

When I "edit raw watchlist" and delete it, it is still on my list.

I tried "watch" then "unwatch". It is still my list.

How can I get rid of it? --Softtest123 (talk) 13:47, 2 February 2008 (UTC)

That sounds very strange. It's only possible to make general suggestions, not knowing what's causing it, but try "edit raw watchlist" again, making sure to hit the "Update Watchlist" button, and try clearing your browser cache. If that does not fix it, post at WP:Village pump (technical) - in case there is some strange database issue the devs should know about. Hopefully this will help. • Anakin (talk) 13:56, 2 February 2008 (UTC)
It also sounds strange to me. The general treatment of WP: has changed. I don't know whether this can cause watchlist problems if you were watching a WP: page like WP:SOURCE. Does http://en.wikipedia.org/w/index.php?title=WP:SOURCE&action=unwatch or http://en.wikipedia.org/w/index.php?title=Wikipedia:SOURCE&action=unwatch work? PrimeHunter (talk) 17:55, 2 February 2008 (UTC)
This may be some side effect of the WP: to Wikipedia: "namespace alias" feature change in the MediaWiki software. Some discussion about this change has appeared on the Help desk: Search Help desk for: WP: Wikipedia: namespace alias. Speaking for myself only, I have not been too fond of this change where I have noticed it. It seems likely to break some things that subtly depend on the previous way things worked. Wikipedia is so intricate that it's hard to change any basic feature without breaking something. --Teratornis (talk) 19:18, 2 February 2008 (UTC)
Thanks. I hadn't tried clearing my cache, so I did, though it seems that when I have had cache problems with Firefox before all I needed to do was simply reload the page. The problem persists.
I get:
Your watchlist has been updated.

1 title was removed:

    * Wikipedia:SOURCE (Talk)

Your watchlist contains 25 titles, excluding talk pages. 

and

...
Voluntary Voting System Guidelines
Voting machine
Wikipedia:SOURCE
User:WimdeValk
Wikipedia:Template messages/Sources of articles
Wikipedia:WikiProject Citation cleanup
...
every time I delete it
I'll submit this section to WP:Village pump (technical).--Softtest123 (talk) 22:51, 2 February 2008 (UTC)

Two issues at Wikipedia:Translation

The first problem can be seen at Wikipedia:Translation/*/In Progress and Wikipedia:Translation/*/Proofreaders Needed. In the template box in the upper left corner of the page, only as late as November 2007 is shown. I put the months up to February 2008 into the template (the template is Template:Translation/Months, which shows the same problem). Interestingly, all of the months do show up in the problem pages' tables of contents.

The second problem can be seen at Wikipedia:Translation/*/Translation Requests and Wikipedia:Translation/*/Completed Translations. Neither of these pages even show the table of contents. The template problem applies to these pages, too.

I'm not familiar with the code {{#if: {{{include_all|}}} so I won't attempt to work it out on my own. But I think that piece of code, found at Template:Translation/Months, might be the source of both problems. − Twas Now ( talkcontribse-mail ) 01:08, 3 February 2008 (UTC)

Issue one is fixed. Looking at #2 now. --TheDJ (talkcontribs) 01:16, 3 February 2008 (UTC)
Got it. Issue two was an overflow of the post expand limitation that the software puts on templates. I suggest simply deprecating this method that is used now. With this I mean that on the pages that include everything (Wikipedia:Translation/*/Translation Requests and Wikipedia:Translation/*/Completed Translations), you should no longer use the {{Translation/Months}} template, but simply hardcode each separate month inclusion. It might be a tad more work, but you simply cannot put as much text in a single template. If you need me to set this up, than leave a note, and I will fix it tomorrow, if no one has responded before me. --TheDJ (talkcontribs) 01:43, 3 February 2008 (UTC)
  • Bugzilla link: 709

While it's not currently possible to move/rename images, Betacommand has a bot that can perform the function for us the hard way.

Requests for image movement should be placed at the bottom of Wikipedia talk:Image renaming. Admins are automatically authorized for the use of this tool, and non-admins may be added by having an admin list you at the bottom of Wikipedia:Image renaming, which also includes instructions on using the tool.

There is no "Requests for" process involved, you just need to have a reasonably good edit history.

Related pages can be found in Category:Image renaming ~Kylu (u|t) 06:17, 3 February 2008 (UTC)

database?

Did the database just crash or something? Wikipedia was down for several minutes for me.--172.168.50.11 (talk) 23:47, 27 January 2008 (UTC)

There were problems for a few minutes; they seem to have subsided. --MZMcBride (talk) 02:03, 28 January 2008 (UTC)
Few minutse?! I thought it was at least an hour. --Coppertwig (talk) 12:56, 28 January 2008 (UTC)
If I had a pound for every time the server went down, I'd be a very rich person. There's no point in worrying about crashes, let alone starting a thread to discuss them: they're always going to happen. Anthøny 23:18, 3 February 2008 (UTC)

Viewing new articles created by people

Is it possible to look at someone's (e.g. your own) contributions only in terms of new articles (preferable even just non-redirects)? It seems impossible to find the new pages you've made among the jungle of edits.

A second question: is Wikipedia's number of articles (~2.2 million) exclusive of disambiguation pages? Richard001 (talk) 01:56, 28 January 2008 (UTC)

  1. I don't believe so.
  2. See Special:Statistics for what it includes, it probably does include disambig pages.
Prodego talk 02:06, 28 January 2008 (UTC)
I have a tool that looks through user contribs to report on link add/removes. It currently spots new articles as "warnings" (and only runs on Windows and takes a while). Tweaking it to scan for new articles only and report them properly is on my list. If you don't get satisfaction elsewhere, watchlist my sub-page, I will announce there when I get that task done. Franamax (talk) 04:55, 28 January 2008 (UTC)
I can't get the tools.de site to work, though I'll try again later. As for the number of articles, I would have asked at the page you linked but it isn't the sort that has a talk page and the related pages are very inactive. Should we actually be including disambig pages in our article count, assuming we actually are? They're not really 'content' pages are they? Richard001 (talk) 07:59, 31 January 2008 (UTC)
Nope, still not working. When we have something running we should add a link at the Wikipedia:User contributions page so others can find this info easier. Richard001 (talk) 23:43, 3 February 2008 (UTC)

Improving appearance of Wikipedia articles...

Dear Friends,

On most Wikipedia pages that have footnotes the gap in the line above the note is increased, and gives the whole article a ragged effect. A piece of css code can fix that easily.

Just add the following to your main .css: sup {vertical-align:text-top;} and see the difference!

Best wishes,

Ananda

Sunday, 2nd February 2008 —Preceding unsigned comment added by 60.54.67.33 (talk) 23:51, 2 February 2008 (UTC)

That would push the <sup> text down to the same level as the text itself, so I don't thinkit's a good idea. EdokterTalk 00:27, 3 February 2008 (UTC)
User:Mzajac/monobook.css/Superscript fix this might be something to look at as well if want to fix it. I believe that so far there has not been a solution that worked properly across all browsers... --TheDJ (talkcontribs) 01:07, 3 February 2008 (UTC)
It would not push the text down to the same level as the text, but to the top of the font height. I can't demonstrate it here. But on my webpage the effect can be seen, and is quite acceptable. Have a look at http://www.ancient-buddhist-texts.net/Buddhist-Texts/K3-Udana-BJT/1-Bodhi-10.htm for an example. Ananda-53 (talk) 02:15, 3 February 2008 (UTC)
Seethis vs. seethis. x42bn6 Talk Mess 05:03, 3 February 2008 (UTC)
That's using Ananda's code suggestion.
Using Mzajac's you get: Seethis vs. seethis. Which looks the same to me (in current firefox, under ubuntu linux, using the Freesans font).
And I strongly agree with somehow fixing this, if it's possible. The articles do indeed look terrible in Opera and Safari (and any other affected browsers). -- Quiddity (talk) 08:24, 3 February 2008 (UTC)
It seems to be dependent on the font used. Certain fonts produce especially large vertical gaps (possibly varying still further across platforms/browsers?), OpenSymbol, Chandas, Jarul, and "serif", are very bad on my system. -- Quiddity (talk) 08:48, 3 February 2008 (UTC)
They look bad in my IE browser too, but I have a fix. I just take all the footnotes out of the article and it looks fine. ;) Franamax (talk) 10:14, 3 February 2008 (UTC)
The solution of Mzajac works very well for me in Safari, Opera and Firefox (all latest versions). The problem is mostly with IE i think (which i cannot test), and older versions of browsers. We might be able to construct some tests that limit the Mzajac solution to just a few browsers perhaps. I too am really interested in it, so i'll look further into it on Monday perhaps (won't get around to it today i think). --TheDJ (talkcontribs) 11:07, 3 February 2008 (UTC)

Unicode is currently planning on adding a capital version of ß (sharp s) in Version 5.1 (expected March 2008). That presents a potential technical problem in that currently we map initial lowercase letters to uppercase. So what will be done for ß when support for Unicode 5.1 is added to the MediaWiki software? Special case it as a lowercase letter for its one article (and 10 redirects added for articles about subjects that begin with the Greek letter β), move it to the uppercase, or use it as an excuse for ending automatic case folding? Caerwine Caer’s whines 00:36, 2 February 2008 (UTC)

We could look at how Greek letters such as π are covered. If the capital is truly the same letter, we could also used the {{lowercase}} template. — Carl (CBM · talk) 00:40, 2 February 2008 (UTC)
Granted that Π is a redirect to Pi, but you enter п, you get told that you have been redirected to Pi from Π, not π because of the case folding. However, when Mediawiki eventually does go to 5.1, if we don't make plans for it, we'll have 11 entries in the article database that actually begin with something no entry now does, a lowercase letter, and that could well cause problems, especially since there are quite a few links to ß within Wikipedia, not to mention those that are without. Caerwine Caer’s whines 00:58, 2 February 2008 (UTC)
I see what you mean now. If we do switch to a new Unicode version, the devs can easily write and execute a script to change the names of the articles in the database to use the correct case. — Carl (CBM · talk) 01:19, 2 February 2008 (UTC)
MediaWiki is already fully case-sensitive. The only reason first-letter capitalization does not matter is that the $wgCapitalLinks switch is set to case-insensitive. Titoxd(?!? - cool stuff) 08:17, 4 February 2008 (UTC)

Color picker

 

Good evening,

On #wikipedia-fr (the French Wikipédia IRC channel), an user has requested help about adding colored text in articles (e.g. for a legend "in red").

As a model is not really a good idea in these case (how to choose the hexadecimal color code?), I've coded http://www.dereckson.be/tools/colorpicker/ to help him.

The left area is to pick the right color, with a Photoshop-like interface.

The right area is divided in 3 tabs:

  • at left, you write the text to hue
  • switch to the middle tab to preview hued text (update in real time)
  • the last tab gives you the code to copy/paste in the wiki.

Enjoy ;-) --Dereckson (talk) 03:58, 3 February 2008 (UTC)

That's some gorgeous JavaScript right there. x42bn6 Talk Mess 05:04, 3 February 2008 (UTC)
We used to have a list of online color pickers at Color tool (which was useful at one point, but was starting to get too big, and is now sadly all gone). This is a particularly good example though. Nicely done. -- Quiddity (talk) 04:56, 5 February 2008 (UTC)

Spare bot in town?

I often need to change hatnotes generated manually by users into {{dablink}} templated messages. The benefits are obvious - they become non-printable and can be programmed by users in their stylesheet.

Typically, these hatnotes take the form of :'' (indent and then italics). This makes them ideal for a bot-assisted search and replace operation. Anyone have a spare bot that could be used for this purpose? Unfortunately, Wikipedia search does not allow search for formatting. API access is therefore a prerequisite. JFW | T@lk 08:53, 3 February 2008 (UTC)

Best place to ask is at Wikipedia:Bot requests. -- John Broughton (♫♫) 21:54, 4 February 2008 (UTC)

Text and links that appear when creating new articles

Recently, the default text and links that appear when creating a new article seem to have changed. Right now it only says

Information for editors

This is fine, but I seem to recall that previously there was a link to look for links to the article and to search for the term. Am I imagining that these links used to be there? If I'm not mistaken, I thought those links were very convenient to help avoid accidentally creating duplicate articles. olderwiser 17:02, 3 February 2008 (UTC)

I concur thoroughly. While you can use the "what links here" link in the left-hand column to check incoming links, there is now no link to search for the term, which can be handy especially for alternate capitalisations and hyphenations. I'd very much welcome its return. Qwghlm (talk) 18:44, 3 February 2008 (UTC)
Came here for the same reason. This needs to be undone, or at least merged with the previous setup. Richard001 (talk) 23:49, 3 February 2008 (UTC)
Ditto. I support the simplified box because noone seemed to read the old one, but now I realise we need the search link back, prominently, or people will create articles just because they can't spell the title. • Anakin (talk) 17:56, 4 February 2008 (UTC)
Note that the text (which depends on the namespace of the created page) is from MediaWiki:Newarticletext, discussed on MediaWiki talk:Newarticletext. PrimeHunter (talk) 01:35, 5 February 2008 (UTC)

CommonSettings.php

I can get commonsettings.php to work for my new MediaWiki 1.12 copy I downloaded from SVN, however I don't quite get InitialiseSettings.php as per http://noc.wikimedia.org/conf/InitialiseSettings.php.html

why does the CommonSettings.php page (http://noc.wikimedia.org/conf/CommonSettings.php.html) not include the following line:

$wgGroupPermissions['steward' ]['userrights-interwiki'] = true;

According to mediawiki.org it says that line has to be included in CommonSettings.php or LocalSettings.php for stewards to set rights on other wikis - which I did using the above text.

Anyone able to help me with this one?? Also, what would I do to get InitialiseSettings.php working?? Thanks, --Solumeiras (talk) 19:15, 3 February 2008 (UTC)

(P.S. Anyone who can help me with this gets a barnstar!)

The files at noc.wikimedia are horribly outdated. Also, the default install of MediaWiki doesn't use CommonSettings, it uses LocalSettings. It's a personal preference of the Wikimedia sysadmins to use the name CommonSettings. --MZMcBride (talk) 01:42, 4 February 2008 (UTC)
  • I was trying to make it so that when I set up my MediaWiki installs on my localhost XAMPP server, they all have the same settings as Wikimedia does. Thanks, --Solumeiras (talk) 18:10, 4 February 2008 (UTC)

query.php interface will be removed soon

Read bugzilla:12881 and http://lists.wikimedia.org/pipermail/mediawiki-api/2008-February/000345.html for more details — VasilievVV (talk) 04:34, 4 February 2008 (UTC)

It should not be removed just because there's another api available, since it is already being used by several user scripts and perhaps also by other applications outside of Wikipedia. Tra (Talk) 17:00, 4 February 2008 (UTC)
Those applications should use the API instead. GracenotesT § 17:36, 4 February 2008 (UTC)
I think Tra is quite correct that removing what was advertised as a stable interface is a poor choice. On the other hand, it seems to be a fait accompli, so the scripts do need to be changed to use api.php. — Carl (CBM · talk) 17:49, 4 February 2008 (UTC)
I hope the idea for a backwards-compatibility wrapper to be left in place is acted upon. Some tools that use query.php aren't in particularly active development anymore, and those would break were query.php just simply removed. Tuvok[T@lk/Improve] 19:55, 4 February 2008 (UTC)

Minimal ToC?

Other than the alphabetical compact Table of Contents templates, is there a means to either

  • Disable the numbering on the standard TOC, or
  • Limit it to a certain level, e.g. Level 2 and 3 headings only?

I ask because the TOC at Wikipedia:New_articles_(New_Zealand) is very ugly - the item numbers run on into the actual dates which are the headings, and it is excessively long. Just the months might suffice. dramatic (talk) 04:53, 4 February 2008 (UTC)

I don't know if you can limit numbers, but you can do level heading limits with {{TOClimit}}. --MASEM 05:04, 4 February 2008 (UTC)
You can remove TOC numbers with <div class="nonumtoc">__TOC__</div>. --Splarka (rant) 08:27, 4 February 2008 (UTC)
Many thanks, that looks much better. dramatic (talk) 22:12, 4 February 2008 (UTC)

How many pages use a template

Wondering if there is a way to find out how many pages use a particular template such as Template:Infobox Rugby Union biography.Londo06 13:12, 4 February 2008 (UTC)

Go to "what links here" in the toolbox, and select the article namespace. There is no reason for an article to directly link to a template so we can safely assume all links are transclusions (i.e. the template is being used). In thiscase it is used 260 times. Fiddle with the numbered links to figure that out. Graham87 13:30, 4 February 2008 (UTC)
Also see Special:Mostlinkedtemplates, as the top templates will be there. • Anakin (talk) 17:58, 4 February 2008 (UTC)

I believe there are some tools on toolserv that can do the counting as well, and not count page links. Not sure where it is, though. -- Ned Scott 03:32, 5 February 2008 (UTC)

You can also use the API, e.g. http://en.wikipedia.org/w/api.php?action=query&list=embeddedin&eititle=Template:Stub&eilimit=500 . MER-C 04:48, 5 February 2008 (UTC)

Delete reason dropdown script

Wikipedia:Administrators' noticeboard#Delete reason dropdown scriptRandom832 17:26, 4 February 2008 (UTC)

Smart category searching

Is there any webpage out there, or is there any potential to upgrade Wikipedia, so I could cross reference categories? E.g, if I wanted to find Dutch actors born in 1970, I could cross reference our Dutch actors and 1970 births categories and come up with a list of articles that are in both category sets. I've poked around and can't find a way to do this, but it would be a really helpful tool for a lot of people. Any thoughts? Neıl 15:18, 30 January 2008 (UTC)

AutoWikiBrowser has a "list comparer" function that does just that. jwillbur 18:19, 30 January 2008 (UTC)
Also, take a look at Wikipedia:Semantic Wikipedia. Especially, Wikipedia:Persondata and the correspondig wikiproject might be interesting. Waldir talk 18:23, 30 January 2008 (UTC)
And catscan of course, if the toolserver ever finishes importing the database after it crashed last week. --TheDJ (talkcontribs) 23:56, 30 January 2008 (UTC)
Give me 48 hours and Ill write a tool that is not dependant on the SQL Database. βcommand 00:03, 31 January 2008 (UTC)
Wicked !, though I'd prefer a commonshelper tool in that case :D --TheDJ (talkcontribs) 00:06, 31 January 2008 (UTC)
this tool will only be a temp tool. As for the commons helper, Ive got no clue how that operates. If There are any future tool request feel free to ask. βcommand 00:18, 31 January 2008 (UTC)
Can I get a copy of this tool? Might be helpful for some image stuff I currently ask East718 for. MBisanz talk 00:43, 31 January 2008 (UTC)
MBisanz, this will still be hosted by the toolserver. βcommand 01:34, 31 January 2008 (UTC)
Has βcommand's planned new tool been made available for use yet? I couldn't find it on ts. No hurry :) - Neparis (talk) 01:49, 6 February 2008 (UTC)
If this thread hasn't already quietly died in a corner, any news or views? - Neparis (talk) 04:25, 6 February 2008 (UTC)
There is a fault tolerant search system that has been implemented on top of wikipedia data here : http://knecht.cis.uni-muenchen.de/exorbyte/wiki maybe that would help. the suggest layer below the query box is searching a full wikipedia index of article titles in realtime and with dynamicc spell corrections. pretty cool no? - Danicc (Danicc) 12:43, 9 February 2008 (PST)

Advanced template question

Is there any way to force the arguments of a template to resolve/execute first prior before actaully calling a template? That is, is there anyway to have: {{t1|a{{#if:{{{b|}}}|{{!}}b|}}|c}} as the content of template t2 such that {{t2|a=1|c=3}} executes {{t1|1|3}} while {{t2|a=1|b=2|c=3}} executes {{t1|1|2|3}}

Basically, this is try to provide a alternating color infobox where entries are optional; I thought using one base template to provided the alternating color functionality using unnamed parameters, and then having the infobox template format a call to this would work (as done on the Navbox template), but the execution order fails more. Of course, if anyone has a better suggestion/existing solution, I'd love to hear it. --MASEM 01:07, 4 February 2008 (UTC)

It cannot be done in the new parser as you described, but the effect can be done other ways. AzaToth 01:29, 4 February 2008 (UTC)
Such as? :-) --MASEM 01:31, 4 February 2008 (UTC)
Parameter names can be conditional, e.g. {{t1|{{{a}}}|{{#if:{{{b|}}}|2}}={{{b}}}|{{#if:{{{b|}}}|3|2}}={{{c}}}}}.--Patrick (talk) 10:38, 4 February 2008 (UTC)
Not in the new parser. AzaToth 10:44, 4 February 2008 (UTC)
It looks like it still works, I just tried it [1]. The equals sign cannot be conditional.--Patrick (talk) 11:14, 4 February 2008 (UTC)
While that's good to know, I don't think that will work for what I'm trying to do; if the equal sign is required, then so must be the vertical bar to separate parameters. I almost need to be able to construct the template call as a string, then feed that string to the parser to work through.
I'm still curious as to what AzaToth mentioned as other ways to achieve the effect of an alternating color table. As I've noted, I've got a template that has some named parameters for appearance control, but otherwise the other 20 parameters are unnamed, so that the first argument is shown in the odd-row style, the second in even-row style, and so forth. With the above, since I cannot seem to leave out the vertical bar, an empty argument (resulting from an optional parameter in the main infobox template) will create an empty argument to the alternating color template, and will result in either a blank row being shown or two rows of the same color next to each other, depending on if I seek empty vs undef, neither which is a solution. (if only we have in-template variable support! :-) --MASEM 17:28, 4 February 2008 (UTC)
For every entry you can count the number of previous entries with code like {{#if:{{{a|}}}|1|0}} + {{#if:{{{b|}}}|1|0}} + .., and check whether it is odd or even. It is rather cumbersome because the total number of such terms is ca. 1/2 of the square of the potential number of items (20 -> 200). If the conditions are complicated, so preferably not repeated, compute an integer of which the binary representation forms the Boolean array indicating which items are selected, and pass that on to another template, like in m:Template:Short DOW ipv (backlinks edit) discussed briefly in m:Help:Array#Producing_multiple_array_elements_in_standard_order.2C_without_duplicates.--Patrick (talk) 02:23, 5 February 2008 (UTC)
Page title
1 1
3 3
4 4
6 6
7 7
9 9
13 13

(←)Ok, I've figured out a way to do this, basically adding up the number of defined unnamed parameters in an #ifexpr, using the mod 2 result to determine odd or even (similar to the above). Of course, now I'm stuck on returns in the table layout: If I use {{User:Masem/infobox/base|title=Page title|1|1|||3|3|4|4|||6|6|7|7|||9|9|||||||13|13|es=background:#f0f0f0;|os=}}, I get the result shown to the right. The extra spaces on the table are a result of having two (or more) empty rows between those that are defined (based on using Special:ExpandTemplates, which seem to automatically create a <br/> and I can't figure out how to not prevent an unneeded row from creating a blank line. I've tried HTML commenting to link lines together with no luck. Any pointers here? --MASEM 07:04, 5 February 2008 (UTC)

See the second part of the section m:Help:Table#Conditional_table_row.--Patrick (talk) 09:33, 5 February 2008 (UTC)
Perfect, that works beautifully. All template issues resolved. --MASEM 15:18, 5 February 2008 (UTC)
For what it's worth, I've taken the advice above to create a new template {{Alternating rows table section}} that creates a section of a table (not a complete one!) that is more generic for anyone other purpose for generating such via templates. This was primarily aimed for infoboxes that used the "no borders" style, but I'm sure there's other uses. --MASEM 22:39, 5 February 2008 (UTC)

posts on talk page without any signatures

Maybe this isn't a bug, but I have noticed this on many talk pages. (I don't want to add clutter, but could give examples if necessary). In these cases, there are parts of a section, or an entire section, of a talk page that is not followed by a signature (including the "unsigned" default). Is this a problem? —Preceding unsigned comment added by Z1perlster (talkcontribs)

Like yours, just now? People forget to sign with ~~~~. There are bots that try to catch these comments and sign in their name, or users may sign them with the {{unsigned}} template (like I just did with your post). It is preferred if people signed, but sometimes they just forget ;-) Technically, MediaWiki won't sign if you forgot to do that. -- ReyBrujo (talk) 03:03, 5 February 2008 (UTC)
I was talking with a non-Wikipedian (non-idiot, but non-technical) friend, and she actually found the requirement to sign comments intimidating enough to avoid commenting altogether: User:Omegatron/Interviews with NormalsOmegatron 04:20, 5 February 2008 (UTC)
It would be more user-friendly if we could use a real forum system like phpBB instead of the wikicode. Maybe someday... —Remember the dot (talk) 05:18, 5 February 2008 (UTC)
Someday we may have mw:Extension:LiquidThreads. PrimeHunter (talk) 10:42, 5 February 2008 (UTC)

Namespaces as subdomains

Would it be possible/desirable to make namespaces into subdomains? Like Image:something.jpg's address would start with http://images.en.wikipedia.org and Wikipedia: namespace pages would start with wikipedia.en.wikipedia.org? The only reason I have in mind is that it would make it easier to Google site: search within a namespace.  :) Any other benefits/disadvantages you can think of? — Omegatron 04:06, 5 February 2008 (UTC)

  • Late comment, since I only saw this question after it archived: {{Google custom}} does exactly what you want (in terms of searching). See the examples on the template page. --Teratornis (talk) 20:49, 9 March 2008 (UTC)
The MediaWiki software really isn't designed to do that... —Remember the dot (talk) 04:47, 5 February 2008 (UTC)
And I think there is some deal/method we have to screen out certain namespaces from google, so a person searching George Bush will get an article, as opposed to the Talk: George Bush page, which might be more active and have more external links. MBisanz talk 04:52, 5 February 2008 (UTC)
You could probably do that with Apache rewrite rules, but that would be insanely computationally expensive for a site of Wikipedia's size, and I'm not really sure about the benefit. Titoxd(?!? - cool stuff) 04:59, 5 February 2008 (UTC)
You should be able to search for any namespace with Google (say, site:en.wikipedia.org/wiki/Wikipedia: for Wikipedia namespace, and site:en.wikipedia.org/wiki/Image: for image namespace). You may get a few false positives, though. And while talking about search, I noticed the MediaWiki search has been downgraded (before I could search for *.play-asia.com/SOap, but now it does not take into account the directory, it would also return *.play-asia.com/paOS results, which were previously filtered out). -- ReyBrujo (talk) 05:17, 5 February 2008 (UTC)
It is my experience that Google searches all namespaces except article talk.
google:site:http://en.wikipedia.org/wiki/ - millions of hits including articles
google:site:http://en.wikipedia.org/wiki/Talk: - no hits (except the article Talk)
google:site:http://en.wikipedia.org/wiki/User: - many hits on this and all the following
google:site:http://en.wikipedia.org/wiki/User talk:
google:site:http://en.wikipedia.org/wiki/Wikipedia:
google:site:http://en.wikipedia.org/wiki/Wikipedia talk:
google:site:http://en.wikipedia.org/wiki/Image:
google:site:http://en.wikipedia.org/wiki/Image talk:
google:site:http://en.wikipedia.org/wiki/MediaWiki:
google:site:http://en.wikipedia.org/wiki/MediaWiki talk:
google:site:http://en.wikipedia.org/wiki/Template:
google:site:http://en.wikipedia.org/wiki/Template talk:
google:site:http://en.wikipedia.org/wiki/Help:
google:site:http://en.wikipedia.org/wiki/Help talk:
google:site:http://en.wikipedia.org/wiki/Category:
google:site:http://en.wikipedia.org/wiki/Category talk:
google:site:http://en.wikipedia.org/wiki/Portal:
google:site:http://en.wikipedia.org/wiki/Portal talk:
Searching old sentences on article talk pages doesn't give Google hits at http://en.wikipedia.org/wiki/Talk. Wikipedia:Searching#Google says: "Google indexes all namespaces". It looks like this should be altered. I don't see a reason for Google to omit article talk in http://en.wikipedia.org/robots.txt or the rendered talk pages. Does anybody know why they do it? Could it be their own choice independent of us? Yahoo includes article talk in search results. PrimeHunter (talk) 11:18, 5 February 2008 (UTC)
See Template:Google custom#Problem with the Talk: namespace for some notes about this. My understanding, which may not be entirely correct, is that Google seems to have stopped indexing the Talk: namespace because some people have used article talk pages to linkspam as a way to inflate the PageRank of the linked-to Web pages. Articles themselves are harder to "game" this way because our WP:EL guideline keeps articles relatively free of linkspam. However, inappropriate links can persist for a longer time on article talk pages, since we normally do not edit other people's talk page comments. Evidently this trick of using article talk pages as a Search engine optimization technique does not work as well on our non-article talk pages, because Google still indexes them. --Teratornis (talk) 17:20, 9 March 2008 (UTC)

Can someone please restore the 'what links here' link to the post-move Special:Movepage screen

Drives me nuts when that disappears. Thanks Shawis (talk) 08:21, 5 February 2008 (UTC)

I restored it. See also wp:VPR#Suggest removal of 'please fix double redirects' message after page move..--Patrick (talk) 09:54, 5 February 2008 (UTC)

New Archive Box

I have created a new archive box to replace the current {{archives}}, {{archive box}}, and {{archive box collapsible}} templates. It incorporates all of the necessary parameters. Could I please get some comments/suggestions on it. And how would I go about implementing it? Thanks, MrKIA11 (talk) 14:57, 5 February 2008 (UTC)

Userspace template issue

This is probably a mistake on my part, but does anyone have a suggestion as to why the small gray note at the top of the User talk:Voyagerfan5761#User talk:89.244.182.22 section is showing the default text specified in the template (User:Voyagerfan5761/talknote) rather than the markup specified? I didn't have problems before the new preprocessor; am I being affected? What do I need to do differently/change in the template or call? Tuvok[T@lk/Improve] 01:47, 6 February 2008 (UTC)

The note text contains "=" characters, so the template didn't recognize the text as a numbered parameter. Easily fixed by prepending the text with "1=". This is not pre-processor related; numbered parameters always had this behaviour. EdokterTalk 02:34, 6 February 2008 (UTC)
Thanks! Why didn't I remember that...? Brain spaced out, I guess. Tuvok[T@lk/Improve] 04:52, 6 February 2008 (UTC)

I have found that Template: Infobox Former Country runs into problems where there is a condominium between two countries rather than a colony of one country - as in the New Hebrides and the Anglo-Egyptian Sudan. The template does not recognize it. I need help finding a way to include it on those pages. Can anyone help? Btn551 (talk) 04:03, 6 February 2008 (UTC)

Going from a redirect's talk page to the redirect without getting lost

Admittedly, there have been several improvements over the years in respect with redirects. One no longer gets carried to the redirect's target when saving an edit to the redirect, and one can also avoid this abrupt journey when following a link from a category, and I think from a couple of other places (as usually, my memory serves me badly here).

I think there is another transition to be improved. The cleverer of you might have guessed it already by looking at the message's title. Personally, I find it stupid to write a message in a redirect's talk page (a protected one, in my case; I was asking for an administrator's help), then try to go back to the redirect and have it... redirect you.

To make a long story short, I now do solemny propose that the upper-left corner tag of a redirect should incorporate the "redirect=no" function (or something like that), at least when returning from a talk page. As I think that the tag's main function is to refresh the page as is, or, again, something like that, I have few hopes of this getting done. I do think, however, that it would save many editors from an unnecessary journey back and forth. Waltham, The Duke of 01:05, 31 January 2008 (UTC)

Hello? Yes, I am right here. Waltham, The Duke of 22:17, 2 February 2008 (UTC)
I do not like moving posts around, but I will not be ignored. I should like to know exactly how ridiculous my idea is or is not before having it archived. Waltham, The Duke of 11:46, 4 February 2008 (UTC)
Good idea.
By the way, for a redirect in a category, the link on the category page redirects you to the target, which I think is good (except in the case of a category just for redirects, intended to keep track of them).--Patrick (talk) 12:06, 4 February 2008 (UTC)
Moral support Sir Duke, do you mean when you go to WP:ANI, click on "discussion" to see the :Talk page, click on "project page" to go back to where you just were, then find you've arrived at WP:AN? If that is what you refer to, I agree completely. It's most confusing... Franamax (talk) 12:18, 5 February 2008 (UTC)
The situation you describe is even worse, because both the project page and the talk page are redirects, and in this case to different pages. Indeed, I just had a look at it and realised that it can get you lost quite easily. However, I am referring to a more general problem: in my case, I left an "editprotected" template in Wikipedia talk:FAITH, then clicked on the "Project page" tab, and found myself in Wikipedia:Assume good faith. In my opinion, there should be an additional link or something similar allowing one to go from a redirect to its talk page and back without the redirects activated. I know it can be done by adding "?redirect=no" to the end of the URL, but it is slow and not all editors know it.
As far as the categories are concerned, I think I will agree with Patrick. Actually, I have made no reference to the matter; redirects constitute a great chapter in the Wikipedia Book of Life, too extensive to be analysed here. Waltham, The Duke of 13:46, 5 February 2008 (UTC)
I find this a problem as well. I tend to find though, that clicking "back" in the browser takes me back to the redirect, or I just click twice. It is annoying, but not enough for me to do anything about it. Carcharoth (talk) 14:30, 5 February 2008 (UTC)
I suppose you are right; technical changes are never made for such small things. I am not disappointed, however; I have another idea, with more potentials (still redirect-related). Waltham, The Duke of 11:52, 6 February 2008 (UTC)

Odd behavior around removed image

At Wikipedia:Copyright problems/2008 February 2/Images, the first item was reported by me. I checked back now to see if anything had happened. The image link was not red, so I tried clicking on it to confirm the image was still there. It wasn't. (So far so good.) The link took me to a page reading:

No file by this name exists; you can upload it. For more information, please see the file's log.

Here "log" was a link to this page, which shows "No matching items in log." It is exactly as if the image was never there (unless you look at the history of the one page where the image was being used).

I then went to Wikipedia:Copyright problems, where the specific page I edited is transcluded. On that page, the link to the image in question is red, so again it is as if the image was never there.

Why should it be different when I access the full "Copyright problems" page vs. the specific subpage? Is it intended that I should be able to "see the log for more information", or not? If so, why is there nothing there? And if not, why was I offered it?

--207.176.159.90 (talk) 01:50, 6 February 2008 (UTC)

Not sure why it was like that, but a page purge fixed the Wikipedia:Copyright problems/2008 February 2/Images page. Gimmetrow 02:15, 6 February 2008 (UTC)
Thanks. But say I hadn't been checking up on my own copyvio notice. Am I supposed to be able to find a log that tells why the image was deleted or not, and if so, how? --207.176.159.90 (talk) 00:13, 7 February 2008 (UTC)
In this case, the image was located at Wikimedia Commons (see the deletion log there), hence why it did not appear in the Wikipedia deletion logs. --- RockMFR 01:06, 7 February 2008 (UTC)
Ah! Thanks. --207.176.159.90 (talk) 02:11, 8 February 2008 (UTC)

anti-aging

the health of the body depends on acid, alkaloid, and base; in other words, pH-balance. this has to do with anti-aging.

on the internet, already, it says that if the cells of the body stay away from the acid side of pH-balance, then most bad diseases can not form in the body.

it is also known that the nucleus of the cell commands the rest of the cell, except, maybe, the mitachondria(s) of the cell. combining pH-balance with the nucleus command states that a less acidic nucleus will knock the cell into its best shape, unless the mitachondria(s) need to be less acidic, too. In that case, if we knock the nucleus and the mitachondria(s) to be less acidic, this will knock the cell into being its best shape.

assuming the best shape of a cell or a body is about 25 years old, it may be best to keep the pH-balance of the cells, their nuclei, and their mitachondria(s) to be the same or around the pH-balance of 25 years old cells.

there may be DNA in the nucleus that command the pH-balance of the nucleus which will also command the pH-balance of the rest of the cell. if this same DNA is in the mitachondria(s) or if the mitachondria(s) nuclei follow the cell's nucleus's DNA, then the entire cell's pH-balance will be commanded by the certain pH-sensitive DNA expression. The trick will be to get this certain pH-sensitive DNA expression to be around the pH-balance of a 25 year old cell's nucleus.

how to find this certain pH-sensitive DNA of the nucleus and mitachondria9s0 is

Huh?
You may be looking for the Reference Desk. Waltham, The Duke of 15:57, 6 February 2008 (UTC)


looks like he was writing an article right here. Since we are on the subject, acidity or acidic-states cause (chronic) inflammation, which has been identified as one cause of premature aging (including of mitachondria) and a factor and cause of many degenerative diseases (such as cancer).  :) The Transhumanist 07:47, 7 February 2008 (UTC)

AWB

I've posted some AWB questions at Wikipedia talk:AutoWikiBrowser. Since you guys are technophiles, I figure you might know the answers. I'd appreciate your input on them. Thank you. The Transhumanist 07:36, 7 February 2008 (UTC)

Template problem

A number of templates have stopped working, looks like the common theme is that they pass a parameter on to another template thus:

{{fact|{{#if:{{{date|}}}|date={{{date}}}}}}}

Maybe this is to do with the new parser, but regardless, is there a workaround?

Rich Farmbrough, 13:20 7 February 2008 (GMT).

It is, the pre-proecessor . Looks like a solution such as
{{fact|{{#if:{{{date|}}}|date|dummy}}={{{date}}}}}
is needed, because the = can no longer be interpreted in a passed parameter in that way (it sets default param 1 to "date=January 2008" instead of date to "January 2007"). Pity, because the new syntax is even more convoluted than the old. Rich Farmbrough, 14:14 7 February 2008 (GMT).
Expected difference #5 of Migration to the new preprocessor. "The equals sign between parameter and value can no longer be generated (by transclusion, parameter, parserfunction, etc) as a delimiter in the template parameters, it is interpreted literally. This requires moving the equals sign outside the generator." --TheDJ (talkcontribs) 15:12, 7 February 2008 (UTC)
Many templates, including "fact", do not distinguish between empty und undefined; then one can use date={{{date|}}}.--Patrick (talk) 16:20, 7 February 2008 (UTC)

Worrisome database locked message

For some reason, this strikes me as particularly worrisome:

Database locked The Wikipedia database is temporarily in read-only mode for the following reason: $1

What in the world does the $1 mean? Is this just a glitch? Silly rabbit (talk) 15:52, 7 February 2008 (UTC)

It means you need to pay to edit.
$1 is a placeholder -- it should be replaced automatically with the reason. Clearly there was some error. Sam Korn (smoddy) 15:55, 7 February 2008 (UTC)
He's kidding about the you need to pay part. (though that *would* curb vandalism :-P )
$1 is a variable or placeholder. Apparently someone forgot to fill it in? --Kim Bruning (talk) 23:15, 7 February 2008 (UTC)

Ping

I'd appreciate some input on Village_pump_(technical)#Category_title_searching, which is a question about improving wikimedia's category searching. - Neparis (talk) 21:22, 7 February 2008 (UTC)

"Save" button does not work in Special:preferences

For some reason, the "Save" button does not work when I am using the Classic skin. It works fine with the other skins. Also, this seems to be an IE-only problem; the button works perfectly when I am using Firefox.

Could this have something to do with the user scripts that I am using? --Ixfd64 (talk) 00:24, 8 February 2008 (UTC)

Userscripts are not called on Preferences page (security precaution) so no. IE6 works fine for me in Preferences in Classic skin. You have to be more specific: you cannot press Save or it doesn't save your modifications? ∴ AlexSm 01:03, 8 February 2008 (UTC)
The button is there, but nothing happens when I press it. In fact, the button doesn't even depress like it is supposed to. --Ixfd64 (talk) 01:08, 8 February 2008 (UTC)
I and others have the same problem. As the FAQ at the top of this page says: If you changed to another skin and cannot change back, use this link. PrimeHunter (talk) 01:13, 8 February 2008 (UTC)

Watchlist - Is there a way to automatically watch every page you view

Watchlist - Is there a way to automatically watch every page you look at, not just edit. Alexsanderson83 (talk) 09:57, 7 February 2008 (UTC)

Add to your monobook.js:
//Auto page watcher, activates the ajax-watch function after 1 second delay
//Note that a delay is required as the ajax watchlist setup script loads after addOnloadHooks.
//If you want to watch *all* namespaces, change wgNamespaceNumber==0 to wgNamespaceNumber!=-1

if(wgNamespaceNumber==0) addOnloadHook(function() {
  var foo = setTimeout('watchThis()',1000); //milliseconds. Increase if on a slow connection.
})
function watchThis() {
  var wli = document.getElementById('ca-watch');
  if(!wli) return;
  var wlink = wli.getElementsByTagName('a')[0];
  //we can't .click() an href, and the .onclick is a mess, so lets clone it
  var wbutt = document.createElement('button');
  wbutt.onclick = wlink.onclick;
  wbutt.style.display = 'none';
  wlink.parentNode.appendChild(wbutt);
  wbutt.click();
}
Quick and dirty. --Splarka (rant) 09:12, 8 February 2008 (UTC)
Watch every page you view? Is this really practical? I have trouble imagining how. Doczilla (talk) 09:15, 8 February 2008 (UTC)
Why not? It might be useful for admins working through Special:Unwatchedpages. Pegasus «C¦ 14:08, 8 February 2008 (UTC)

Change approval

We use mediawiki at work. I know a change approval option has been discussed, but has it ever been implemented? They'd like this feature at work. RlevseTalk 14:48, 7 February 2008 (UTC)

There is something being developed at mw:Extension:FlaggedRevs (and being tested at test.wikipedia). It doesn't prevent the unapproved changes being viewable (in the history or recent changes for example), they just don't show as the "top" edits (by default). --Splarka (rant) 08:34, 8 February 2008 (UTC)

Comment not eaten

The comment on the first line of ZHX1 causes additional vertical space. I thought it would normally be eaten by the parser (or that's what I gathered from Migration to the new preprocessor), so its presence shouldn't make a difference at all. (It seems to interplay with the invisible template, since it doesn't occur on simpler pages.) Is this a bug, or am I missing something? -- Lea (talk) 10:52, 8 February 2008 (UTC)

Okay, so what happens is this:
<!-- foo -->
{{PBB Controls}}
<!-- bar -->
test
gets transformed to "\n\ntest", since the invisible template leaves a blank line and there is something in front of it. Ouch. Could the line eating algorithm be fixed somehow, so that comments at the beginning truly don't matter? -- Lea (talk) 10:56, 8 February 2008 (UTC)
You have newlines before and after the comments, so the result is as expected.--Patrick (talk) 11:22, 8 February 2008 (UTC)
It seems that in the case of one or more comments separated by single newlines, the newlines before, between, and after these count together as not more than one. If such sequences are separated by e.g. two newlines, this rule applies separately for each sequence. In the example the template call has no effect, only the newline before and after it.--Patrick (talk) 12:16, 8 February 2008 (UTC)
Hm. Would it be reasonable and possible to change the parser so that it eats lines that contain invisible templates? Several invisible templates in a row would certainly cause unwanted blank lines (unless you avoid the line breaks), so this would seem like an improvement. -- Lea (talk) 22:03, 8 February 2008 (UTC)
In my experience, that kind of pseudo-intelligence is usually a bad thing. What's wrong with
<!-- foo
--> {{PBB Controls}} <!--
 bar -->
test
instead? —EncMstr 22:16, 8 February 2008 (UTC)

Timeline Question

Hi. For some reason, on List of Justices of the Supreme Court of the United States (edit | talk | history | protect | delete | links | watch | logs | views), this edit broke the time line, but this previous edit was fine. (The first edit actually did break the time line when I tried to pipe the link to a non-existent page, but after I fixed that it was fine. No such luck with the problem edit - the page I am trying to link to definitely exists.)

Any help on this would be much appreciated! Thanks, --Sam Chase 17:48, 8 February 2008 (UTC)

What's broken about it? I'm not seeing any problems with that version. Algebraist 18:03, 8 February 2008 (UTC)
Okay, that's really, really, weird - when I first edited it, and when I posted the diff, the whole time line had completely disappeared. But now it's back...I guess I'll redo the edit, and hope that you won't think I'm ripe for the loony bin. --Sam Chase 18:06, 8 February 2008 (UTC)
Yep, just redid it and it looks fine. Sorry for causing all the trouble. --Sam Chase 18:10, 8 February 2008 (UTC)