Wikipedia:Village pump (technical)/Archive 23

From Wikipedia, the free encyclopedia

Village pumps: PolicyTechnicalProposals (persistent)AssistanceMiscellaneous

Contents

[edit] Is this possible?

Hello all,

Is it possible to run a bot/js script etc to carry out the following:

  1. Trawling the list at Tramcars of the National Tramway Museum#Complete list under the "Status" section
  2. Compiling a list at the bottom to give the number of each type, i.e:
    1. Operational = foo
    2. Out of Service = foo
    3. etc (it would need to work by colour, as different editors put different things (In Service/Operational!)
  3. Add the Operational Number to Template:NTM operation cars, so that it appears on all pages using the template.

Comments would be appreciated, as would anyone with the technical know how to actually write/run the thing!

Thanks,

BG7 17:29, 26 February 2008 (UTC)

Sounds like a bot request; the place to ask is WP:BOTREQ. -- John Broughton (♫♫) 18:53, 28 February 2008 (UTC)

[edit] (not yet written)

I have no earthly idea where else to discuss the new red link feature (within Wikipedia, at least), so I am posting this here.

I have one simple point to make: useful as it may be to newcomers, "not yet written" is often inaccurate because many red links are a result of deletions, deleted pages having been (quite obviously) written at least once. I believe that the label should change to something more inclusive.

Does anyone agree with me on this? Waltham, The Duke of 13:31, 27 February 2008 (UTC)

The message is from MediaWiki:Red-link-title and can be discussed on MediaWiki talk:Red-link-title (which amusingly says not yet written at the moment). PrimeHunter (talk) 13:46, 27 February 2008 (UTC)
Just noticed that it also appears on Image namespace links, which is also kinda "awkward" at times. The coolest image.jpg Might be a good idea to introduce a {{#switch:{{NAMESPACE}} for a "(not yet uploaded)" message ? --TheDJ (talkcontribs) 19:51, 27 February 2008 (UTC)
I agree that the message ought to be clearer. I haven't thought of an improved wording, though. --ais523 20:29, 27 February 2008 (UTC)
I changed it to "page does not exist".--Patrick (talk) 01:46, 28 February 2008 (UTC)
I did see the page before posting here, but was not sure whether these were ever used for discussion. Stupid of me, I know, but, after all, the talk page was "not yet written" and the decision ought to have been discussed somewhere (or at least that's what I thought at the time). I did see, at a later time, the article in the Signpost making mention of a mailing list—but I barely know what a mailing list is, so...
In any case, the current message is much better. Thank you for considering my petition, and may the Unicorn (mhhhnbs) keep you safe from the evils of our world. Waltham, The Duke of 09:50, 28 February 2008 (UTC)

[edit] Set Twinkle to stop speedys showing as minor edits

how can I stop speedy deletions showing up as minor edits when I use twinkle?Williamwade (talk) 04:33, 28 February 2008 (UTC)

In your monobook.js add:
TwinkleConfig = {
        markSpeedyPagesAsMinor          :       false,
};

seresin | wasn't he just...? 05:05, 28 February 2008 (UTC)

Actually that's not 100% correct, it should be:
TwinkleConfig = {
        markSpeedyPagesAsMinor          :       false
};
The last element should never be followed by a comma in JavaScript. Opera bugs out bigtime if you do this wrong. But I'm guessing only very few people use Opera. --TheDJ (talkcontribs) 19:49, 28 February 2008 (UTC)

[edit] editing the MediaWiki namespace

Not sure how long this has been happening, but when l look at the source of any of the MediaWiki: pages such as MediaWiki:Protectedtext for instance, I am getting an extra <p> right above the "This page provides interface text for the software, and is locked to prevent abuse." I've been unable to track where that is coming from. Anyone have an idea? --TheDJ (talkcontribs) 12:30, 28 February 2008 (UTC)

Doesn't seem to be at MediaWiki:Protectedinterface. I guess it might be a small bug in the new parser. • Anakin (talk) 14:06, 28 February 2008 (UTC)
As an admin, I get a different header, and no extra tag. Does this help localise the problem? Happymelon 19:01, 28 February 2008 (UTC)
As an admin, you're still able to log out (or simply use another browser) and then check how it looks. —AlexSm 19:27, 28 February 2008 (UTC)
I was only trying to help, no need to bite. Happymelon 22:13, 28 February 2008 (UTC)

There is another problem, which I've been observing for several days. All pages in MediaWiki namespace pretend as if they were manually protected by admins: they have var wgRestrictionEdit = ["sysop"] in JS variables section in HTML source and they display MediaWiki:Protectedpagetext message on "view source". This should not happen, as MediaWiki namespace pages are protected by default. As far as I can see, this happens only on English Wikipedia, other projects I checked still behave as usual (example on de). —AlexSm 19:27, 28 February 2008 (UTC)

[edit] WIKIPEDIA DEMOGRAPHICS

I was wondering if there is any wiki-tool to check the demographics of the hits to a Wikipedia article. I am thinking in where is the computer located or how long a particular user stays reading the article. Would Google Analytics work?

Also, is there a way to survey the reader to see if he/she found the article useful? Something like a pop up or a question at the end of the article. I am thinking in something more survey oriented than the the dicussion page.

Thank you!

anunezsanchez (talk)

The short answer is no. Providing information on who read what article would be a privacy violation. You can't use Google Analytics because you can't get access to the weblogs. More relevant, Wikipedia is the 8th-most read website in the world; the weblogs are kept only for a very short amount of time. And because Wikipedia doesn't sell advertisements, it's not particularly interesting to the Wikimedia Foundation to see (a) where readers came from or (b) how long they stayed, so such analysis simply isn't done (internally).
If you're interested in page views, you can find them here. But that has no demographic information, just a total count per day for any given article. -- John Broughton (♫♫) 18:42, 28 February 2008 (UTC)

[edit] HotCat.js

Something that I know several people have been waiting for. I have started a port of HotCat.js hopefully later to be turned into one of the available gadgets. The code is currently here in my user space. Comments can be left on its talkpage. A short list of important changes from the original:

  • blacklisting code is disabled.
  • all code for the uploadForm has been removed
  • autocommit is disabled
  • will be enabled on pages without categories so that you can easily add them
  • uses javascript:void() as a dummy value for href in order to avoid a conflict with popups.
  • checks for {{Uncategorized}} and removes it if a category is added
  • does not use JSconfig for configuration options like its Commons original

People are invited to test and give feedback. --TheDJ (talkcontribs) 17:32, 28 February 2008 (UTC)

[edit] Possible bug???

Hi, I just nominated Pork Stories for deletion using Twinkle and the afd nomination page has been created and posted at afd, however the link in the article still looks as if there is no discussion there, even though when you press it there is. Is this deliberate, or is it a bug? -- Roleplayer (talk) 21:55, 28 February 2008 (UTC)

WP:PURGE or something called "browser cache" Not a bug, not deliberate, just one of the caches being out of date with reality. --TheDJ (talkcontribs) 22:04, 28 February 2008 (UTC)
Ooh it worked, thank you. -- Roleplayer (talk) 00:01, 29 February 2008 (UTC)

[edit] Pages still taking forever to load

Pages are still taking a long, long time to load. I have access to two PCs and it's happening on both of them. Nothing's changed since I reported this a couple of days ago. Corvus cornixtalk 23:34, 28 February 2008 (UTC)

The previous report is #Pages are loading really slooooooooooowly? -- SEWilco (talk) 23:52, 28 February 2008 (UTC)

[edit] Wikipedia:Template messages/Image namespace is in too many categories

If you click on the above link, and scroll to the bottom of the page, you'll see that this list of messages is included in over a hundred categories. For example, it's in Category:Department of Transportation images while it is obviously not itself a Department of Transportation image. (I had noticed this problem when I saw this list of messages appear in Category:Contested candidates for speedy deletion, and I was pretty sure it was not a candidate for speedy deletion). Luckily no quick-fingered admin has yet seen fit to speedy delete it.

This problem doesn't occur for a similarly-named page, Wikipedia:Template messages/User talk namespace, which is only in one category. Does anyone know if this is a fixable problem? EdJohnston (talk) 19:18, 25 February 2008 (UTC)

The page includes samples of lots of templates. Those templates are meant to be used on pages in a corresponding category. They add a pretty little box explaining that this image is a so-and-so and add the page to the category. Since this page uses all of those templates, it is placed in all of those categories. JackSchmidt (talk) 19:25, 25 February 2008 (UTC)
There are ways to change templates so they don't do this. One way I've never seen implemented is to have a parameter for "display" being on or off. The default would be off, and the template would work as normal. The "display=on" parameter would be used when just displaying the template on a list page like the above, and the categories would be switched off in that "display" mode. But i don't know enough about templates to do that. Carcharoth (talk) 01:35, 26 February 2008 (UTC)
I could probably whip something up, if it would be useful. The main issue is just, oh my goodness, that is a lot of templates. What would you think about about the syntax:
{{add cats|Articles needing attention|Even-toed ungulate stubs|1973 in sports|display={{{display}}} }}
where you would add this to any of those tempaltes that add categories? You would just sort of list categories between pipes, and then include the magic words display={{{display}}} somewhere to make sure it knew whether it was being called to display an example or not. In display mode, it could just sort of give a short note
Better syntax? Worth doing? JackSchmidt (talk) 02:09, 26 February 2008 (UTC)
What I meant by "display" mode was a mode that showed what the template looked like (ie. same messages as normal), but just omitted the categories. Does that make sense? Carcharoth (talk) 02:23, 26 February 2008 (UTC)
Sure. I was suggesting a way for template writers to do this without learning complicated template syntax. Inside the template, instead of listing the categories normally as [[Category:Articles needing attention]], the template writer would use {{add cats|Articles needing attention|display={{{display}}} }}. Then when someone wanted to just show off the template as an example, they would say {{my favorite template|text=What it is|year=2008|display=on}} instead of {{my favorite template|text=What it is|year=2008}}. Then the template would display normally, except instead of including the categories on the page, it would (1) do nothing, or (2) make a pretty little informational box thing saying what categories it would have added the article to had it not been in display mode. JackSchmidt (talk) 02:31, 26 February 2008 (UTC)
Sounds great. I find with things like this, if you do it for some templates, it will catch on and spread. Well, hopefully. Or you could try and do it for all templates... Carcharoth (talk) 02:52, 26 February 2008 (UTC)
(←) I created the template in my user space at first, to make sure the idea is clear. The internal template for template writers is at User:JackSchmidt/TL_AddCats, and a sample template is at User:JackSchmidt/TL_AddCats/testcases, and a sample article using the sample template is at User:JackSchmidt/TL_AddCats/sandbox. If all looks good, I'll move these over to Template:Add cats tomorrow, and start doctoring up the image templates. JackSchmidt (talk) 04:01, 26 February 2008 (UTC)
A large number of templates already use the parameter |category= to override display of categories. Inside the templates, just replace the code [[Category:SomeCat]][[Category:SomeOtherCat]] with {{{category|[[Category:SomeCat]][[Category:SomeOtherCat]]}}}. To override the display of categories (or to manually specify a category), call the template as eg {{unreferenced|category=}} rather than {{unreferenced}}. I would beseach anyone considering adding category exemptions to templates to use this system rather than a new one. Happymelon 10:31, 26 February 2008 (UTC)
I agree. But whatever system is used, please make sure it is widely documented. It is no good if people don't know about it!! Carcharoth (talk) 11:59, 26 February 2008 (UTC)
If a system already exists, I am happy to use it. However, no such system exists for {{unreferenced}}. You can see at Commutant, that adding {{unreferenced|category=}} still adds categories to the article. Note that {{unreferenced}} is aware of what namespace it is used in, and only has effect really in the article namespace (or article talk). JackSchmidt (talk) 14:21, 26 February 2008 (UTC)
I just wrote down the first categorising template that came to mind (shows my gut reaction on the state of Wikipedia :D); I've added the necessary code to {{unreferenced}}, so have a look at my two edits to see how to implement this. {{db-g1}} would have been a better choice - all the CSD templates use this opt-out mechanism AFAIK. My knowledge of this system was derived from MediaWiki:"Unused" templates.css, which needed a method of preventing the page from being listed in every possible speedy deletion category. Category:Category suppression supporting templates holds a (very messy) partial list of templates supporting one or more methods of category opt-out. I would strongly support any attempt to standardise category opt-out, and would advocate this technique as the simplest and most elegant method of doing so. Happymelon 15:26, 26 February 2008 (UTC)
I prefer using #ifeq to detect namespace, as transcluders don't have to worry about remembering the parameter name. –Pomte 15:40, 26 February 2008 (UTC)
I think the parameter expansion is incredibly natural, and far superior to using a separate template (as I had earlier suggested). I also like the #ifeq to handle obviously bad placements of the tag. I suspect in most cases the #ifeq is enough, since lists of templates should probably not be in the article namespace. My guess though is that the article Talk: namespace could both reasonably expect to use templates "on purpose" and "for display", so that the category= method is useful to have too.
Thanks very much for the references. I have an interest in helping with template and mediawiki coding, but wikipedia is such a huge place it is hard to know where to start. Selectively including categories is quite easy, and I thought the ambox made a pretty display, so it seemed like a good project. JackSchmidt (talk) 16:10, 26 February 2008 (UTC)
"Intelligent" templates which switch the categories included based on which namespace the template is in are extremely useful, and I'm not suggesting that they be removed or replaced. I don't really count that as "category opt-out", as I see that as a pure "manual override". Currently we have some templates using the |category= system, some using various other override parameters. I would like to see these manual overrides standardised, and would support the "category" parameter as the means of doing so. Happymelon 19:01, 26 February 2008 (UTC)
JackSchmidt has edited the troublesome template {{Di-replaceable fair use disputed}} which I found to be causing the unwanted inclusion of the message list in a speedy deletion category. The message list is now out of that category, though it remains in some others. He used Happy-melon's method with the 'category=' to suppress unwanted inclusion in categories. I hope he will add more to this thread about how his fix works. EdJohnston (talk) 13:59, 29 February 2008 (UTC)

[edit] Pages are loading really slooooooooooowly

This started yesterday and is continuing. I see the top half of a page immediately, then the whole rest of the page takes a long time to load. I can't scroll until the load is complete, so I just sit there and twiddle my thumbs until the page finally finished loading and releases it hold on the scrolling. Corvus cornixtalk 03:25, 27 February 2008 (UTC)

Get rid of the dial-up connection.--The Fat Man Who Never Came Back (talk) 03:37, 27 February 2008 (UTC)
It's probably your browser's fault. Browsers don't display the HTML of a page after a JavaScript script tag until after the JavaScript loads. This could be fixed by moving all the script tags to the bottom of the page, but that might break several existing user scripts. GracenotesT § 03:46, 27 February 2008 (UTC)
Why would this just start happening, and on two different computers, neither one of them a dialup connection? Corvus cornixtalk 03:50, 27 February 2008 (UTC)
It does look like it's an IE7 problem. When I use Firefox, I don't have the problem. Corvus cornixtalk 03:53, 27 February 2008 (UTC)
I've been having this problem all day, and I use Firefox most of the time. (It was occurring in IE as well, though - I checked both browsers on two systems.) Really long delays in loading, only for Wikipedia pages, and I've already tried restarting my computers, router, and modem. Mid-day today, I even encountered a "Wikipedia database error" message, along with some malformed edits (duplicates, text dropped, etc.) and long delays when saving edits. --Ckatzchatspy 05:55, 27 February 2008 (UTC)
This is exactly what I'm having happen. I'm using Firefox, on two separate systems, and have been dealing since yesterday with very slow load times on most pages (though this does fluctuate), one attempt to load this page that ended with just a blank page, and at least two edits where the page I was editing suffered a major chopping for no apparent reason. Witness Dave Grohl yesterday. oops. Tony Fox (arf!) 16:54, 27 February 2008 (UTC)
The same (top half ok, rest of page took the proverbial forever) on Safari for about an hour yesterday. — Athaenara 11:20, 27 February 2008 (UTC)

Do you have any elements on the page that fail to load at all ? There are several scripts hosted on toolserver or other foundation servers, if those error out, you could experience things like this and it might effect some browsers more than others. But element load failures should be easy to check with almost any browser. --TheDJ (talkcontribs) 10:23, 27 February 2008 (UTC)

I've never noticed any failute to load completely, and never a blank page or an error message. But it's still happening on IE7 today. Corvus cornixtalk
(copied from "Wikipedia slowing down?" below) Is anyone else experiencing this? For the last two weeks or so, I've been getting frequent outages, where I either can't get into the site at all ("server not responding"), or where it's taking five-ten minutes for pages to load, especially around 10 pm UTC. Preview is so slow, it has become practically unusable. Even when the site is loading normally, everything is noticeably slower. I went on IRC to ask if this was affecting anyone else, and a couple said yes, but most said no. I'd guess it started about two weeks ago. SlimVirgin (talk)(contribs) 21:48, 29 February 2008 (UTC)

[edit] Articles created

What's the easiest way to see or compile a list of all the articles a particular user has created?

The Transhumanist    17:53, 28 February 2008 (UTC)

That would be this tool. It's a bit slow, but you can just let it run for five or ten minutes while you do something else. -- John Broughton (♫♫) 18:31, 28 February 2008 (UTC)
Perfect! Thank you. The Transhumanist    19:01, 28 February 2008 (UTC)
I was looking into why this was slow. Apparently whether the edit created the page is not stored in the user contrib/page history, but rather in the recent changes (which is cleaned out to only have a month or two of changes), so more or less for every page you've edited, the tool has to check the page history to see if you were the first. Does that seem reasonable? I hadn't noticed the big difference between recent changes and page history before, but I think this explains why some tools are slow. JackSchmidt (talk) 16:23, 29 February 2008 (UTC)
There is no efficient way with the current database schema to get a list of pages created by a certain user. A database schema change (coupled with a maintenance script) would fix this. The key is finding someone willing to do it. ; - ) --MZMcBride (talk) 20:28, 29 February 2008 (UTC)

[edit] Protection against creation without log entry

I just deleted Creative web studio after deleting and protecting Creative Web Studio.

While I am not entirely displeased with the result I am puzzled though at the protection of Creative web studio as there are no log entries for it. All I can think is that the alternative capitalization does automatically protect. Agathoclea (talk) 16:09, 29 February 2008 (UTC)

It does protect against the creation of all capitalizations. Unfortunately, the logs will only show it if you get the capitalization exactly correct, which can make it harder to find out exactly why a page was protected. E.g., Creative Web Studio, not Creative web studio. • Anakin (talk) 16:42, 29 February 2008 (UTC)

[edit] Bot flag on edits

Is it true that the bot flag b on edits is only kept for recent changes? In particular, for page histories or user contributions, there is no such thing as a "bot edit" (they are just like other edits)? JackSchmidt (talk) 16:15, 29 February 2008 (UTC)

Yep it is. The "recentchanges" table in the database has columns "rc_minor", "rc_bot", and "rc_new", for the m, b, and N letters, respectively. The "revision" table only has "rev_minor_edit", so once things go off recent changes, information about whether it was a bot edit is lost unless you check who made the edit. Bug #11181 is a request to change this. Development on it seems to have stalled due to the required database change being a nuisance. • Anakin (talk) 16:54, 29 February 2008 (UTC)
Excellent, thanks for the clear and thorough answer. Looks like changing the behaviour would be difficult. Right now I'm just working on changing my perception. JackSchmidt (talk) 17:01, 29 February 2008 (UTC)

Assuming this change is desired for long-term statistical research on bot edits, the data would be useless unless it was 100% complete. Retroactively adding a "bot flag" to revisions in page histories would be complicated (though not impossible) as it would require examining each edit, identifying the user who made the edit, then checking not whether the user is currently a bot, but whether the user was a bot (at the time of the edit). The latter would involve comparing timestamps in the user rights log both here and on meta. Is there some reason this would be worth the trouble to obtain complete data (or some reason that partial data would be useful)? — CharlotteWebb 19:52, 29 February 2008 (UTC)

To be clear, I'm not suggesting a change in mediawiki. I'm just observing something that had not been so clear to me before: the pages that use "recent_changes" have access to a lot more information, but over a shorter time frame, than those that access "revision". JackSchmidt (talk) 20:22, 29 February 2008 (UTC)

[edit] How to change font size in a table?

I want to change all of the text in a table to small font ... how do I do that?

I have read through the Help on Tables three times and I can't find any mention of this. I know that I could add the small tags in each cell of the table, but that is a lot of tedious work. Can it be done globally in a table? - mbeychok (talk) 18:53, 29 February 2008 (UTC)

you can change the style attribute of the table
like this Pomte 19:30, 29 February 2008 (UTC)
Try putting the following in the first table line:
{| style="font-size: 85%;"
(...rest of table)
EdokterTalk 19:31, 29 February 2008 (UTC)
Since you specifically mentioned <small> tags, I believe using "font-size:smaller;" will have the same effect (likewise "font-size:larger;" and <big>) though the actual change in size will vary from one browser to the next. Probably better to stick with percentages as a general rule. — CharlotteWebb 19:39, 29 February 2008 (UTC)
Thanks, everyone. Using the style attribute in percentages does exactly what I wanted. - mbeychok (talk) 20:02, 29 February 2008 (UTC)

[edit] Wikipedia slowing down?

Is anyone else experiencing this? For the last two weeks or so, I've been getting frequent outages, where I either can't get into the site at all ("server not responding"), or where it's taking five-ten minutes for pages to load, especially around 10 pm UTC. Preview is so slow, it has become practically unusable. Even when the site is loading normally, everything is noticeably slower. I went on IRC to ask if this was affecting anyone else, and a couple said yes, but most said no. I'd guess it started about two weeks ago. SlimVirgin (talk)(contribs) 21:48, 29 February 2008 (UTC)

See also: #Pages are loading really slooooooooooowly --TheDJ (talkcontribs) 21:59, 29 February 2008 (UTC)
Thanks. I've copied this to that section too. SlimVirgin (talk)(contribs) 22:09, 29 February 2008 (UTC)

[edit] formatnum

Someone asked on one of the Reference Desks about a population shown in E-notation at List of countries by population, and someone else "corrected" it via this edit.

Evidently, someone thought it was smart to have the page model the current population of China as increasing linearly and update itself automatically. (Seems dubious to me, but this is not the place to discuss that.) But they used something called "formatnum" to do it and round off the results, and this produced scientific notation.

Except it doesn't for me. When I view the previous version of that page, I see the population as an integer, rounded as requested: 1,322,720,000. When I view the edited version, I see the unrounded integer: 1,322,721,419. Is this number being formatted in my browser or what? That doesn't seem to make sense—I would have expected it to go into the Wikipedia cache in already-formatted form. And if I view the HTML source I just see <td>1,322,720,000</td> for the entry. So why are these people seeing scientific notation?

Also, where is this "formatnum" documented? Not at WP:Formatnum or Template:Formatnum or Formatnum, that's for sure.

--207.176.159.90 (talk) 01:30, 29 February 2008 (UTC)

Formatnum: a magic word that inserts decimal seperators. See Help:Magic words#Formatting. The E notation however was caused by the "round -4" part that was embedded in the #expr: function. EdokterTalk 01:41, 29 February 2008 (UTC)
I suppose it does not depend on formatnum specifically. See also m:Help:Calculation#Numbers as output, the format of the result of any computation may depend on the server, even within Wikimedia.--Patrick (talk) 01:59, 29 February 2008 (UTC)
Okay, it seems to me that this should be considered a significant bug to be fixed (presumably by rewriting formatnum so it does not depend on system-dependent behavior like exactly what sprintf(buf,"%g",val) gives you). What is the point of having a utility like formatnum if it may produce output formats that are wildly inconsistent (in terms of human-readability) for the same value?
Thanks for the pointer to Help:Magic words. I've been contributing to Wikipedia for a couple of years, and this is the first I can remember even hearing the term used.
--207.176.159.90 (talk) 00:10, 1 March 2008 (UTC)
I find it confusing too. However, rather than using an improved formatnum as a workaround, it would be good if the result of #expr would be in a format independent of the server.--Patrick (talk) 10:03, 1 March 2008 (UTC)

[edit] Special:Recentchangeslinked

How does the "Special:Recentchangeslinked" work? Or rather, will "Recentchangeslinked" show changes to either the article or to it's talk page? -- SatyrTN (talk / contribs) 06:41, 29 February 2008 (UTC)

It shows all recent changes for all pages linked to, from the current page. For category pages, it shows recent changes for pages in the category as well. --Splarka (rant) 08:30, 29 February 2008 (UTC)
But does it show recent changes to the talk pages of those that are listed? -- SatyrTN (talk / contribs) 14:04, 29 February 2008 (UTC)
No, have to add those separately. • Anakin (talk) 14:31, 29 February 2008 (UTC)
Um. Is that something that can/should be changed? -- SatyrTN (talk / contribs) 15:19, 29 February 2008 (UTC)

Probably better to leave the default behavior alone, but an &includetalk=1 parameter in the "related changes" url would be helpful for users who want it to behave exactly like Special:Watchlist. Of course this same option should also work the other way (to show changes of the corresponding article even if only the talk page has been linked to). — CharlotteWebb 20:03, 29 February 2008 (UTC)

One odd technical thing here is that the "watchlist" SQL table has rows for both the regular page the talk page. When SpecialWatchlist.php queries the database, it has no special code for talk pages. That means if someone does write an includetalk patch, it will have to have what is more or less a new idea. The standard namespaces all come in Main/Talk pairs, and DefaultSettings.php instructs one to do this for custom namespaces too, so it should not be too hard, but it might take longer to review the patch.
Does this feature sound interesting enough to implement? JackSchmidt (talk) 20:30, 29 February 2008 (UTC)
Partially done as bugzilla:13202. I might need some help to make it work for Special:Recentchangeslinked/Category: since categorylinks does not include the page title or page namespace for pages in the category, just the page id. The obvious answer would then require joining a new table into the query, and I've never done SQL programming at the wikipedia.org scale, so I'm not sure how reasonable that is. JackSchmidt (talk) 20:57, 29 February 2008 (UTC)
Two more things someone could help with:
  • Someone familiar with Help/Interwiki: Where does one document this new flag? Is it in the source code, on mediawiki.org, both?
  • Basic PHP coder: One should probably add a little checkbox on the default Special:Recentchangeslinked page so that people can just click away to find the feature.
Thanks, JackSchmidt (talk) 21:07, 29 February 2008 (UTC)
It would probably be just a regular hypertext link to go back and forth between the two modes (just like "hide minor edits" and any of the other options), not a check-box. — CharlotteWebb 22:07, 29 February 2008 (UTC)
Yup, very true. I've done this as you suggested, and I think documentation is taken care of. The main thing left is handling categories. I'll give that a try tomorrow, but I don't have a dev account to test the SQL query on a giant thing like wikipedia. JackSchmidt (talk) 04:11, 2 March 2008 (UTC)

[edit] Moving on watchlist

I watch several thousand articles (almost all being small communities in the USA that don't get edited very often), and I pay attention to my watchlist quite often. It's somewhat annoying, however, that a page is never listed in the watchlist simply for having been moved to a new name. Such a thing shows up in the article history as a modification (look at Middleburg, Logan County, Ohio for an example), but not on the watchlist. Is there a way to have my watchlist display when a page is moved? The only solution that I have is occasionally to look through my raw watchlist, and note pages that have improper titles (for example, US communities are supposed to have their state's name in their title, and occasionally people remove it), if I can see that. Nyttend (talk) 21:22, 29 February 2008 (UTC)

Right: deletion, moving and protection are not shown in watchlist. I have a half-finished script that loads raw watchlist, loads recent logs and then shows you which watched pages were recently moved. I guess could find some time to finish and publish it. —AlexSm 22:11, 29 February 2008 (UTC)
Yes please. Or any other solution to this problem. I share this frustration, with not being made aware of page moves. -- Quiddity (talk) 21:58, 1 March 2008 (UTC)
And I would very much like to see when a page on my watchlist is deleted. A deletion is about as big as a change could be. The watchlist shows other changes but not the really big change of deletion. Sbowers3 (talk) 23:17, 1 March 2008 (UTC)

[edit] Wikiproject newsletter - existing tool or would require new bot?

One WP I'm in is considering creating a "newsletter" to drop to all members of the project via talk pages to let them know where help is needed (as well as to let them know how to either remove themselves from the project or to just stop getting the newsletter). Delivery would be similar to the WikiSignpost that exists already. My question is, is there a generalized bot or existing tool that can handle such deliveries, say after some expected page is created or the like, or would a new bot be needed for this (and if so, likely we'd have it do some of the prep work for the newsletter). --06:24, 29 February 2008 (UTC)

WP:AWB is an automated tool that any user with more than 500 edits can use. If you used an opt-in list as the absis for distributing the newsletter, that would be a permissible use of AWB. You can either create a page and have AWB subst:transclude it or just have AWB append the entire content of the newsletter to a new section in the user's page. MBisanz talk 06:39, 29 February 2008 (UTC)
Dur, of course , I completely forgot about that! The question that then comes up is, for it to be used in a bot-like manner, I would likely need to apply for a bot-account to use it, correct? If I have a list of 200+ users that I want to deliver to, even if it's just hitting the "process" button 200+ times, it can be a bit annoying. --MASEM 14:34, 29 February 2008 (UTC)
There are quite a few bots that have approval for this. I would ask at Wikipedia:Bot requests for anyone available to do it. It wouldn't even need a new BAG request I don't think. Woody (talk) 16:28, 29 February 2008 (UTC)
Yea, to use AWB in that manner you'd need a bot-account. But given the bots we already have, I'd say to find one of them. MBisanz talk 20:07, 29 February 2008 (UTC)
You can find a long (but possibly incomplete) list of bots that do such a delivery in the Editor's index. Just be sure to check their user contributions to see if they're still in use. -- John Broughton (♫♫) 22:39, 2 March 2008 (UTC)
Masem Ill be glad to deliver your newsletter. βcommand 22:44, 2 March 2008 (UTC)

[edit] Image not appearing

Can anyone see Image:Funeral Marches and Warsongs Cover.jpg? I cannot, and neither can anyone I talk to. I have reuploaded it and it still is not appearing - the wikimedia server it is hosted on gives me a 404. Ideas? αѕєηιηє t/c 09:45, 2 March 2008 (UTC)

I have the same problem. No idea? 79.112.24.4 (talk) 09:47, 2 March 2008 (UTC)
Also happening here. αѕєηιηє t/c 09:50, 2 March 2008 (UTC)
Another one. αѕєηιηє t/c 09:52, 2 March 2008 (UTC)
See also on ro.Wiki http://ro.wikipedia.org/wiki/Imagine:Grigore_Ghi%C5%A3%C4%83.jpg . What's happening? 79.112.27.163 (talk) 09:54, 2 March 2008 (UTC)

Fixed. -- Tim Starling (talk) 11:09, 2 March 2008 (UTC)

[edit] Hidden category bug

When a page is only in categories that are hidden, the category box still appears. For example: [1]Remember the dot (talk) 01:43, 1 March 2008 (UTC)

I noticed this as well when I was porting HotCat.js, It simply looks "wrong". The big problem is that even if Cats are hidden, you need to be able to force them to show (and you want them to be in the HTML for that). Perhaps the best thing would be to add an empty "Categories:" text. It wouldn't be ideal, but at least you would then know what that "weird empty box at the bottom" is for. It wouldn't be consistent with pages that have no categories at all, but I cannot think of a better solution. --TheDJ (talkcontribs) 11:07, 1 March 2008 (UTC)
Perhaps have a bot look for all pages in only hidden categories and tag them with {{Uncategorized}}? –Pomte 03:07, 2 March 2008 (UTC)
?? {{Uncategorized}} is an admin cat and is thus a hidden category itself! What is solved? Hmains (talk) 04:23, 2 March 2008 (UTC)
How about if we ask the developers to make the "Categories" box invisible if it contains no visible categories? —Remember the dot (talk) 05:10, 2 March 2008 (UTC)
The problem here (based on discussion with devs on IRC), is that the actual box is part of the monobook skin, so solving this issue poses some problems. Basically <div id=catlinks>(added by monobook) and <div class=catlinks>(skin core) need to be merged into one div, or there need to be 2 seperate blocks for the 2 types of categories, but this has obvious problems of course with existing tools/skins etc that might break if we change things like that. I was thinking that alternatively, we could add some simple JS to the monobook (and monobook compatible) skins that would work around it, but it's not really a fix of course. Still, it would be easy to do.
d1 = document.getElementById ('catlinks')
d2 = document.getElementById ('mw-normal-catlinks')
d3 = document.getElementById ('mw-hidden-catlinks')
if( d1 && d3 && !d2 ) {
  // Following should work for IE6, IE7, FF, Saf2, Saf3 and Opera)
  d1.style.cssText = "display: none;";
  d1.setAttribute( "style", "display: none;" );
}
Less than ideal, but something that might be considered if this issue remains on the table. --TheDJ (talkcontribs) 12:52, 3 March 2008 (UTC)
Uncategorized seems like a reasonable exception to the hidden admin category emerging consensus. I think it shouldn't be hidden. We want all users to know that the article is uncategorized, so it is the one case where a self-reference makes sense. I'm going to un-hide it. -- SamuelWantman 08:37, 3 March 2008 (UTC)
I didn't realize how many there are. I have to go to bed. Maybe someone else can un-hide them all... -- SamuelWantman 08:42, 3 March 2008 (UTC)
Did anyone finish this. Where is the list? Is it all the subcats of Category:Category needed, or are there more than this that need fixing to show up? Carcharoth (talk) 13:50, 3 March 2008 (UTC)
Still needs to be done. All the descendants of Category:Category needed. -- SamuelWantman 16:11, 3 March 2008 (UTC)

[edit] Deletion log glitch

I speedily deleted Eric crespo using twinkle, but there's nothing in the deletion log. Is there something wrong with the MediaWiki software? Spellcast (talk) 06:56, 3 March 2008 (UTC)

Hmm. Strange. Not a clue, but that is worrying. Carcharoth (talk) 17:36, 3 March 2008 (UTC)
I'm not that familiar mediawiki database, but, I believe this has been known to happen sometimes, so you shouldn't worry about it much. Oysterguitarist 05:31, 4 March 2008 (UTC)
It happens occasionally with blocks and deletions. It has to do with using two transactions (for performance reasons) instead of one. What that actually means is beyond me. ; - ) Bugs have been filed in the past; known problem. Cheers. --MZMcBride (talk) 05:58, 4 March 2008 (UTC)

[edit] Uploads image dimensions, but not image

I've had this problem a few times. I will decide to upload a new version of an existing image to wp, but when I hit the upload button, instead of uploading the new version of the image, it just appears to upload the dimensions of the new image but leaves the old image in place, resulting in a distorted image. Am I doing something wrong? Thanks, Gatoclass (talk) 11:16, 3 March 2008 (UTC)

The upload works when you do that, it just doesn't purge the cache of the old thumbnail versions right away - append '?action=purge' to the image page URL. —Random832 14:55, 3 March 2008 (UTC)
Even purging probably won't do it, it's usually the cache. In Firefox, press CTRL+F5. αѕєηιηє t/c 19:33, 3 March 2008 (UTC)
Instructions for all major browsers. --TheDJ (talkcontribs) 19:42, 3 March 2008 (UTC)

[edit] WP keeps signing me out

Is anyone else having problems staying logged in? WP will only keep me logged in for about 60 seconds. --207.206.137.71 (talk) 18:58, 3 March 2008 (UTC)

Over on WikiSource, it isn't logging me out, but it's only allowing me to make edits sporadically. The rest of the time, it says "Sorry! We could not process your edit due to a loss of session data. Please try again. If it still doesn't work, try logging out and logging back in." Of course I've reloaded and logged out/in multiple times, but it only allows an edit to succeed rarely. – Quadell (talk) (random) 19:06, 3 March 2008 (UTC)
I had that too. About 30 seconds then, *poof* logged-out. But it's fine now. - 52 Pickup (deal) 19:08, 3 March 2008 (UTC)
Nope, now it won't keep me logged in long enough to even make an edit under my username!--207.206.137.71 (talk) 19:12, 3 March 2008 (UTC)
OK, I shut down all of my browser (Firefox) windows and cleared the cache and it seems to let me stay logged in now (I logged out to make this edit, I may be dumb, but I'm not stupid).--207.206.137.71 (talk) 19:26, 3 March 2008 (UTC)

[edit] MediaWiki:Titleblacklist

I posted a comment on MediaWiki talk:Titleblacklist about whether it would be possible to use this page as a more elegant solution to the problem we have of people uploading images with titles like Image:Picture.jpg. Currently we have images uploaded and protected at titles like this, which is a pretty clumsy solution. Is this possible? Or is the title blacklist restricted to the mainspace? Happymelon 11:22, 27 February 2008 (UTC)

MediaWiki:Filename-prefix-blacklist provides the functionality you need. MER-C 11:35, 27 February 2008 (UTC)
Does that page work? Looks fake to me. --MZMcBride (talk) 11:31, 28 February 2008 (UTC)
Oop. I guess it is real. Dunno when / how that got added into the software.... --MZMcBride (talk) 11:39, 28 February 2008 (UTC)
"Hermy"? "Hagger"? Whoever's idea it was, they have forgotten to black-list Grawp as well. :-D Waltham, The Duke of 11:50, 28 February 2008 (UTC)
Why is it that DSCF, DSCN, and DSC_ were listed but not DSC (which is what every digital camera _i've_ seen uses) —Random832 15:29, 4 March 2008 (UTC)

[edit] Help with Signature?

The code I have is: <font color="green" font face="Comic Sans MS">[[User:Stepshep|§tepshep]]</font>{{•}}<font color="green">[[User talk:Stepshep|¡Talk to me!]]</font>

Which makes: §tepshep • ¡Talk to me!

But when I put it in the four tildes I get: <font color="green" font face="Comic Sans MS">[[User:Stepshep|§tepshep]]</font>{{•}}<font color="green">[[User talk:Stepshep|¡Talk to me!]]</font> (talk) 19:45, 27 February 2008 (UTC)
When I try to set it to a raw sig, I get the red error.

Any help? Thanks. §tepshep • ¡Talk to me!

You have to use "raw sig" option, just remove extra "font" word immediately after "green". And please do not use templates in your signature. —AlexSm 20:08, 27 February 2008 (UTC)
It's a protected template and it'll be automatically substed when he signs anyway. —Random832 15:24, 4 March 2008 (UTC)
First, the template is already used in the (manualy added) signature above. Second, he might find a way to get around automatic substing in the future. —AlexSm 15:37, 4 March 2008 (UTC)

Thanks, I removed the brackets too. §tepshep¡Talk to me! 20:59, 27 February 2008 (UTC)

[edit] Help with 3RRR report

Would someone check over my 3RRR report for correctness of formatting :[2]?

Every time I have filed one it has been rejected as "malformed" because, for the life of me, I cannot understand the directions. Please help! Thanks, Mattisse 17:00, 2 March 2008 (UTC)

You didn't change [http://VersionLink VersionTime] to an actual version. On the history page of the article, each time/date represents a version; in most browsers, you can right-click one of these to copy the URL to your computer's clipboard, then paste it in, over the "Versionlink" text.
Also, you cannot use warnings posted on article talk pages as proof that the editor was warned; you have to post a warning on the offending editor's user talk page. If you post on an article talk page, there is no reason for admins to believe that the editor actually saw the warning. In fact, you should never post warnings about potential 3RR violations on article talk pages; they belong only on user talk pages.
And, again, technically, in the text Diff of 3RR warning: [http://DIFFS DIFFTIME], you should replace the URL with the diff for your posting on the user talk page. What you posted wasn't a diff, it was a link to a section on a page.
You should really compare the text that you create versus the text that other editors are posting on the page; for example, they are using a piped link to create diffs like this:
You're doing something like this (which does provide a link, but isn't as easy to read:
  • [3] 22:00, 28 February 2008
Hope that helps. -- John Broughton (♫♫) 23:12, 2 March 2008 (UTC)
I don't understand what you mean. Do you mean a diff from the article's history? (I'v crossed out the report, but I would still like to know). Thanks, Mattisse 23:18, 2 March 2008 (UTC)
I am still interested in getting an answer to my last question - the link that I did not provide that I should have -- I do not undertand the jargon wording in the 3RR directions. Thanks! Mattisse(Talk) 15:03, 3 March 2008 (UTC)
Answered on user talk page. -- John Broughton (♫♫) 15:12, 4 March 2008 (UTC)

[edit] Sock library

From time to time I'll see a user pop up in several places at once in ways that seem strange to me (newbies who know to much, people missing obvious POINTs, etc). While we all recognize the top sockpuppeters by virtue of their mass production, a lot of the more minor individuals continue to be problems. It would be great if we had a sock puppet database that listed the MOs of all Socks. I know WP:BEANS so maybe make it a linked wiki that only admins can access (like the mediation/workinggroup wiki). For instance, if I see a user making certain types of proposals and being overly detailed, I could search "Proposals AND Wordy" and see socks with that profile. Or if I see someone !voting multiple times in Hindu-related articles. Just an idea. MBisanz talk 18:34, 3 March 2008 (UTC)

Personally, I think a much easier way would be to clean out the sock categories, deleting all the old and useless ones. I would bet that most of them only have a handful of accounts created by people who have moved on months or years ago or who may have only made one sock account. Its mostly just clutter that will never be used. Alternately, we could abandon sock categories altogether and use manually updated lists. People would be less likely to waste their time creating a list for 1 sock and we wouldn't have to keep all the old userpages around. In list form they would also be searchable. Mr.Z-man 06:14, 4 March 2008 (UTC)
Thats not a bad idea. Unless there is a way to search Cats in a comprehensive sense (push-thru), they aren't really that useful at IDing new socks. Maybe some sort of search engine that only searches SSPs? MBisanz talk 07:05, 4 March 2008 (UTC)