Wikipedia:Village pump (technical)

From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by Mdennis (WMF) (talk | contribs) at 17:58, 17 June 2011 (→‎augmenting article history: reply). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

 Policy Technical Proposals Idea lab WMF Miscellaneous 
The technical section of the village pump is used to discuss technical issues about Wikipedia. Bugs and feature requests should be made at BugZilla.

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. Questions about MediaWiki in general should be posted at the MediaWiki support desk.


Problem with box padding

Hi,

On my user page, I'm getting extra left-padding in the box on the right.

I created a simpler test case in my sandbox. Here is the code:

{{Boxboxtop|About|backgroundcolor=lightblue}}
{{Userboxtop|}}
{{User en}}
{{User Wikipedian for|year=2005|month=12|day=5}}
{{Userboxbottom}}
{{Boxboxbottom}}

Here is what I see. The left margin is far bigger than the right margin. Does anyone know how to fix this?

Tommyjb Talk! (12:57, 30 May 2011)

 Done Wasbeer 02:12, 11 June 2011 (UTC)[reply]

Hiding top edits

I used to be able to hide items on my contrib list that were most recently edited by me (via a doohickey in my monobook) - since the contrib lists have changed to enable an editor to only see their top edits, this no longer works, which is a real pain. Is there any way of being able to again just see contribs which have been edited since my last edit (other than adding several thousand new items to my already-bloated watchlist)? Grutness...wha? 01:04, 4 June 2011 (UTC)[reply]

User:Markhurd/hidetopcontrib.js is still working for me with the Vector skin - any good? -- John of Reading (talk) 06:45, 4 June 2011 (UTC)[reply]
I'll try that, thanks. Grutness...wha? 19:11, 4 June 2011 (UTC)[reply]
Grutness, since you're using my top-edit hiding script in nothingthree.js (at least in your vector.js), could you please let me know what browser/OS/etc. you're using? The script still works perfectly for me, so the bug that breaks it for you is probably trivial to fix. I've had no problems with it in Safari, and I debugged it a little for IE fairly recently. {{Nihiltres|talk|edits|}} 16:17, 4 June 2011 (UTC)[reply]
Safari 5.0.2 on the machine I'm currently on... I suspect the other computer I sometimes use has an earlier version of Safari. Grutness...wha? 19:11, 4 June 2011 (UTC)[reply]
My dilapidated old laptop is running Safari 3.0.4, and it's stopped working on that, too. Grutness...wha? 00:52, 6 June 2011 (UTC)[reply]
I'm so far unable to replicate the problem. Try clearing your cache, in case the problem is a fluke. {{Nihiltres|talk|edits|}} 02:25, 5 June 2011 (UTC)[reply]
Seems to be working again on 5.0.2 - may have been a cache problem. Grutness...wha? 00:55, 7 June 2011 (UTC)[reply]
Uh...no. It worked again once on 5.0.2 - and hasn't worked again since. Doesn't seem to work at all now in either 5.0.2 or 3.0.4. I'll try the Markhurd one, I think... Grutness...wha? 10:47, 10 June 2011 (UTC)[reply]

CentralNotice about Board of Trustees election

Every Wikipedia page I visit says: The Wikimedia Board of Trustees election has started. Please vote. When I click that link it says "Sorry, your account on the English Wikipedia does not meet the voting requirements for this election." Right underneath the request to vote! Can we please modify this to only be shown to the users it applies to? It's really a slap in the face to the users to say "regardless of what you've contributed so far, it's not enough to count for us, so we don't want you as part of this community."

Thanks,

WBTtheFROG (talk) 02:23, 9 June 2011 (UTC)[reply]

I completely agree with what you said. I understand that I haven't contributed a lot to Wikipedia, but do I really need it rubbed in my face? Ybrik222 (talk) 17:10, 12 June 2011 (UTC)[reply]

augmenting article history

Sometimes an article has a history which is not adequately represented by the list of editors in the history tab. This creates a licensing, especially when automated tools such as PediaPress use the history information. The simplest example of this problem is when a section of article A is used as the basis for new article B. Another example is a paragraph from article A being added to article B. The usual practise is to describe this is the edit summary when saving the edit to article B, but an automated tool can't reliably parse this edit summary. However if that isnt done, how do we 'amend' or 'augment' the history to reflect this?

We sometimes place {{Copied}} on the talk page.

Another case to consider is where article A is later deleted; currently we usually keep the history of article A public, or we record its history on the talk page.

However a note on the talk page doesn't appear to comply with the GFDL's expectation that there is a history 'section' (which we interpret as the history tab). I've placed this on the technical VP page as I am wondering if we can create a solution where the amended history is in an easily detectable location and consumable format, and this is noted on the history tab when needed. If we create pages named "Talk:[Article]/History notice" (like we do with FAQs) or "Template:History notice/[Article]" (similar to page notices), automated tools can look for its presence and include it, and a message (such as mediawiki:histlegend) can do the same. John Vandenberg (chat) 13:28, 9 June 2011 (UTC)[reply]

I'm not trying to immediately go off track here, but before getting in to this can I ask a honest question? In all seriousness... why does this matter? Don't get me wrong, I'm all for giving credit where credit is due, but... what are the details of the potential licensing issue, here? I don't think that this can be adequately addressed without firmly establishing the details of the problem. Besides... according to Wikipedia:Reusing Wikipedia content, the page can't have been undated since November 2008 in order to export text under a GFDL license, right?
— V = IR (Talk • Contribs) 16:43, 9 June 2011 (UTC)[reply]
I'll leave the technical issue aside (just passing through), but our content is dually licensed. Reusers can export content under CC-By-SA or GFDL. If content cannot be exported under GFDL, we are required to indicate as much in the footer, history or discussion pages, but that should not apply to content copied from other Wikipedia articles, unless so indicated. (Just to note: this is not legal advice or an "official Wikimedia Foundation" statement. :)) --Maggie Dennis (talk) 19:07, 9 June 2011 (UTC)[reply]
I cant see where Wikipedia:Reusing Wikipedia content mentions November 2008.
As Maggie has pointed out, if an article can not be re-used under the GFDL, we are required to inform potential reusers when they cant use the GFDL. Currently we say that potential re-users will need to read a) the footer, b) the history and c) the discussion pages to determine whether they can copy a Wikipedia page. Even for a single Wikipedia page, that could involve hours of reading. If a human wants to copy more than a few pages, the amount of reading they must do quickly becomes ridiculous. Automated tools can't read; even if we used consistent templates for automated tools to parse, needing to parse the article text and article discussion pages is a big job if they only want to determine whether or not additional attribution is needed. (I think I need to find some examples where an automated tool has bungled the attribution by relying on the list of usernames...)
Wikipedia:Verbatim copying under the GFDL#History Section currently can't provide a single description of what the 'history' section in the GFDL refers to. If we can include notes on the history tab, we can migrate our existing notes, recommend that contributors record credits in this way, and eventually inform consumers that they can a) link to the 'history tab' as a single record of the attribution for that article, and b) automatically construct an attribution block by combining the history information and the text on the associated "History notice" page if it exists. John Vandenberg (chat) 23:53, 9 June 2011 (UTC)[reply]
I think that this is part of an overall problem of attribution of any article text that was not original text written for that article. This could include the following.
  • Text copied from another article on the same language wiki.
  • Text translated from another WMF wiki.
  • Text copied from a website that uses CC-BY-SA (or GFDL before Nov 2008)
  • Text submitted and verified by an OTRS ticket.
  • Text in the public domain.
Unfortunately there doesn't seem to be a standard method (or even place) for indicating that some text has been copied from elsewhere. This is particularly true for translations which I have seen attributed in the article text itself, on the talk page and occasionally via history comments. I was toying with the idea of proposing a more standard mechanism whereby these attributions would be put inside a (collapsible) template on the talk page but something that could appear on the history page would be even better IMO. However, doing this on the history page would obviously require an MW software change.
I think that any potential re-users need to have a fairly good idea of the origin of all the article's text. Another reason that this may be important is if some of the text was published before 1923 but written by someone who died less than 70 years ago. Any re-user in a country that follows the PMA-70 rule could not legally reprint that text. Boissière (talk) 17:13, 12 June 2011 (UTC)[reply]
I think we can transclude a note onto the history tab without any MW software change - we can use mediawiki:histlegend as a temporary solution at least.
Thank you for expanding on the types of attribution issues which we are not addressed in a standardised manner. I think we should build a list of good examples below. Feel free to add examples, and replace any where you have a better example of the same problem.
I dont think we need to include OTRS tags verification.
Your final comment about "pre-1923 text which is not 70pma" is an interesting problem which frequently arises on English Wikisource, where we put a very distinctive notice on the bottom of the text. Do you know of any instances where such text has been reused in English Wikipedia? If so, I agree that should be prominently mentioned. English Wikipedia has a similar problem with images; although the images are clearly tagged, they are used in articles without any way for the casual reader to know the article includes images which are illegal in their country. John Vandenberg (chat) 11:54, 14 June 2011 (UTC)[reply]
I don't currently have any examples of pre-1923 text though I did wonder if, for example, any of the poems of the First World War poets were reproduced substantially but I can't readily find any (poetry's not really my thing though). So this is perhaps more theoretical at present. By the way I have added a couple of examples of translation attribution to the list below. Is this the sort of thing that you are looking for? Boissière (talk) 22:32, 14 June 2011 (UTC)[reply]
Thank you. I've added The Sniper, Ode of Remembrance, Suicide in the Trenches and On Scratchbury Camp below. John Vandenberg (chat) 14:14, 15 June 2011 (UTC)[reply]
Some examples of internal moving I'm aware of that created attribution problems (this is quite common when an article is split): Talk:History of Greek and Roman Egypt has the main notices, with Talk:History of Ptolemaic Egypt and Talk:Egypt (Roman province) having the corresponding notices. Sometimes the problem goes the other way (with a merge), but in this case it was a split after a long history at one location. I'm not sure it is entirely practical to attribute very old fragments of text that have been moved back-and-forth and up-and-down and all around, but certainly the movement of large chunks of text should be tracked and links made to or copies maintained of the editing history. Carcharoth (talk) 00:14, 17 June 2011 (UTC)[reply]
But again... I'm not trying to be flippant here or anything, I'm honestly curious: why? What exactly is the interest in attributing specific bits of text in an article to specific users?
— V = IR (Talk • Contribs) 00:36, 17 June 2011 (UTC)[reply]
The GFDL requires that each new author is listed. See §4.I. We state that our text is available under the GFDL. For pages which are only available under CC-BY-SA, we must say so. Which means we need some consistent way of recording that a page isnt available under the GFDL. John Vandenberg (chat) 05:20, 17 June 2011 (UTC)[reply]
but... it doesn't say that specific bits of text need to be attributed to specific authors. Does it? We already do list each author, one way or another (as has already been outlined here), so... I guess that I'm just not clear on what the problem is. I have to be honest and admit that I don't see that there actually is anything to worry about here regardless, but it appears as though yourself and others are saying that there is some sort of legal issue here. Why should we, as editors, be concerned about issues that responsible reusers and the Foundation should consult a competent lawyer about regardless of what any editor here says or does?
— V = IR (Talk • Contribs) 06:03, 17 June 2011 (UTC)[reply]
If you don't care issues like this, why are you bothering to commenting here? Most editors don't hang around VPT. Some of us do care about current and potential re-users, especially when they are wanting to comply with the license and want advice. If you think this problem can be solved by a lawyer, then you havent understood the problem. If re-using Wikipedia content requires consulting a lawyer, we've failed to achieve the goals of http://freedomdefined.org/. I am often asked questions about reusing content, and I currently need to say that people should not re-use Wikipedia content unless they analyse the full edit history of the page and the associated talk page. The most 'freedom' supporting lawyer in the world will say "the advice of Wikipedia is sound, which is that you can not reuse Wikipedia content unless you manually eyeball every revision of the page and its associated talk page" and their client will say "oh, well that is impractical".
Do you think it is practical for anyone to reuse Wikipedia content only after reading every revision of page and talk page? John Vandenberg (chat) 07:01, 17 June 2011 (UTC)[reply]
Well, I was trying to understand where you were coming from, and what it is that you wanted to be done (I have a small interest in working on the page history functionality in the software for a slightly different reason), but... no big deal. I'll leave you to carry on here without getting in your way.
— V = IR (Talk • Contribs) 12:05, 17 June 2011 (UTC)[reply]
In addition to §4.I, if you're still interested, you might want to see §5 of Wikipedia:Text of the GNU Free Documentation License, specifically where it talks about combining documents released under GFDL, indicating that "In the combination, you must combine any sections Entitled 'History' in the various original documents, forming one section Entitled 'History'...." Since our goal as editors is to create content which can be freely distributed anywhere, we do want to do our best to make sure that we are keeping it free under the licenses that we've selected. That both licenses are important is suggested by the fact that we will not accept text licensed only under CC-By-SA if the contributor is the sole copyright owner (which means that if you've published something somewhere else first, you have to dually license it. The only exception is if you co-own copyright, in which case we will accept CC-By-SA-only licenses. Of course, we do accept CC-By-SA-only licenses authored by others; this complication arises when the person placing the content here is the copyright owner.) (Again, I'm not talking for the WMF here. :)) --Maggie Dennis (WMF) (talk) 17:58, 17 June 2011 (UTC)[reply]

Examples
  • Optics, which has a note on the talk page and a separate history at Talk:Optics/Wikisource history. The results of the "contributors" tool can be seen here and [1], and of course there is no mention of ScienceApologist/128.59.171.155 or user:Awadewit. The "Article Sources and Contributors" section of Book:Physics is six pages of usernames, but ScienceApologist doesnt get a mention.
  • Kubla Khan was imported from simple.wikipedia.[2] and the attribution is a note on the talk page: Talk:Kubla_Khan#update_from_Ottava_Rima. The author of the rewrite was banned user User:Ottava Rima. He had made some edits to that page before being banned, so he does appear in Contributors
  • Elegy Written in a Country Churchyard is also an import of User:Ottava Rima's work on simple wikipedia, however this time there is no mention of him in the Contributor list, and there is no obvious attribution note on the talk page (comments mention Ottava Rima, but nobody says the article was written by Ottava Rima).
  • Nucleosol has a {{CCBYSASource}} note at the bottom which says that it includes text from wikidoc. The WikiDoc page has two editors: user:Marshallsumter is also an editor of the WP article however user:Zorkun hasnt edited the Wikipedia page, so they are not attributed.
  • Colin_Gunton includes {{Theopedia}} at the bottom, however it does not contain a link to theopedia or any authorship information, so I am 100% confident that is violating their copyright.
  • David Douillet contains a (very small) template on the talk page saying that it is a translation of the corresponding fr.wiki article (but no details of which version of that page it is translated from and certainly no author info).
  • Hélène Dorion says (at the bottom) that it "contains information from the equivalent article on the French Wikipedia" with wikilinks to the relevant article and also to the fr.wiki mainpage for some reason.
  • The Sniper (poem), Ode of Remembrance and Suicide in the Trenches include a PD-1923 text which is not PD in the country of origin which requires 70pma. (Does On Scratchbury Camp fit within PD-1923? If not, it is a copyright infringement)

WikiProject category templates

A large number of WikiProject categories have been CFDed to use "importance" rather than "priority" - the list is at Wikipedia:Categories for discussion/Working/Manual#Other. However I can't locate the precise point to adjust the templates that populate them - is anyone able to find and initiate the change? Timrollpickering (talk) 07:11, 10 June 2011 (UTC)[reply]

TW SPI reporting not working

The sock reporting component of Twinkle is not working. Dr.K. λogosπraxis 07:34, 10 June 2011 (UTC)[reply]

Please see WT:TwinkleDoRD (talk) 15:35, 10 June 2011 (UTC)[reply]

Huggle

I'm trying to configure Huggle so that no talk pages are included. I can't seem to get the command right. The command I found Here seems to be (namespaces:"alltalk") If I read the parameter correct wrapping (alltalk) with ("") disables Huggle from reading and loading talk pages. What is happening is talk pages are still being loaded and when exiting Huggle it reverts the command off my huggle.css page. What am I doing wrong ? Mlpearc powwow 15:19, 10 June 2011 (UTC)[reply]

Probably best asked at Wikipedia:Huggle/Feedback. – ukexpat (talk) 16:15, 10 June 2011 (UTC)[reply]
I think you need to add namespaces:-alltalk to your CSS page. You'd probably have to ask on the Huggle talk pages why it's removing stuff though. — Bility (talk) 16:17, 10 June 2011 (UTC)[reply]
Thank you, and Hello Ukexpat, long time no read . Mlpearc powwow 16:36, 10 June 2011 (UTC)[reply]

Last login - "touched field"

In the discussion about inactive admins, User:Ohms law mentioned here a "touched field" in the user table, and mentioned that there are likely extensions that already make use of it. I was watching that post but no responses were forthcoming. Do any such extensions exist? How feasible would it be to give certain users (crats, CUs, etc.) access to last-login information? I'm not trying to weigh the merit of a function like this, just the feasibility. ▫ JohnnyMrNinja 16:42, 10 June 2011 (UTC)[reply]

I've done some looking around, MW:Extension:LastLoginTime seems to allow users to see their own login time, and MW:Extension:UserSnoop seems to allow an array of functions and requires a special right. Could these be easily modified for the purpose? ▫ JohnnyMrNinja 03:46, 12 June 2011 (UTC)[reply]

How does one colour a an entire column in a table?

Eg. In the List of metropolitan areas in the Americas, I want to seperate the cities located in North America from those in south America so I was thinking of making them different colors. Can anybody help? — Preceding unsigned comment added by ThisguyYEAH (talkcontribs) 19:20, 10 June 2011 (UTC)[reply]

I think you'd have to color each cell individually, but please don't - we shouldn't convey information using color alone. I think the county is sufficient, or you could include a symbol to state which continent. --Golbez (talk) 19:41, 10 June 2011 (UTC)[reply]

Sorting talk pages with a template

Please assist {{OnThisDay}} adds talk pages in the form of (e.g.) Category:Selected anniversaries articles (March 2009)|{{PAGENAME}}, which means that a page such as Talk:The Beatles gets sorted as "The Beatles" rather than "Beatles"—is there a way to fix this? Thanks. —Justin (koavf)TCM☯ 20:53, 10 June 2011 (UTC)[reply]

Yes, you can wrap {{PAGENAME}} in a custom sort parameter, like {{{catsort|{{PAGENAME}}}}} so that it uses the parameter but defaults to the page name so it can be safely omitted if you don't need it and won't affect already-transcluded OnThisDay templates. — Bility (talk) 21:21, 10 June 2011 (UTC)[reply]
Quite a lot of templates which categorise a page will force the sort key to {{PAGENAME}}, which has always puzzled me. This would make sense if the normal action in the absence of a sortkey was to take the namespace into account (i.e. that putting [[Category:Test]] onto Talk:Foobar would sort that page under T, not F); but it doesn't work that way - Talk:Foobar sorts under F. It's a pain because a forced sortkey like this defeats the action of {{DEFAULTSORT:}}; and on talk pages, the {{DEFAULTSORT:}} is normally simulated by adding the |listas= parameter to one of the WikiProject banners (usually the first one).
I was actually looking at this specific case quite recently - the template concerned isn't {{OnThisDay}} itself, but {{OnThisDay/link}}, which contains the following:
[[Category:Selected anniversaries ({{#time:F Y|{{{date}}}}})|{{PAGENAME}}]]
It beats me why this particular template should require a default sort key to be defeated. Personally I think that it could be simplified to
[[Category:Selected anniversaries ({{#time:F Y|{{{date}}}}})]]
I suppose this should be brought up either at Template talk:OnThisDay or at Wikipedia talk:Selected anniversaries. --Redrose64 (talk) 13:51, 11 June 2011 (UTC)[reply]
{{PAGENAME}} is always the base page name, not the full name including the namespace. Therefore, your suggested edit would place all pages in that category under the T of Talk:. Ucucha 12:22, 14 June 2011 (UTC)[reply]
Your first statement is correct; the full name including the namespace would be {{FULLPAGENAME}}. But your second is incorrect: this is easily demonstrated by locating a talk page which has no banners (because they might cloud the issue by invoking a hidden {{DEFAULTSORT:}}), and onto that, add a category without specifying a sort key, and after saving, look in the category to see where it's been sorted. See Talk:Sioux Webserver and the only category that it's in - Category:Unassessed Computing articles. For me, I see it sorted under S, not under T. --Redrose64 (talk) 13:54, 14 June 2011 (UTC)[reply]
You're right. I'm almost sure it previously was as I wrote; the change might have to do with the new sort collation algorithm. Ucucha 13:56, 14 June 2011 (UTC)[reply]

New messages bar

The bar, no matter what colour it's in is bloody annoying. I propose that we use a central notice style notice, that would be far less intrusive and would be just as effective. —James (TalkContribs)3:45pm 05:45, 11 June 2011 (UTC)[reply]

  • I think part of the reason for the conspicuousness (is that a word?) is so that you're alerted in case someone is trying to reach you urgently. For example, someone might be trying to tell you that a string of edits you're making is causing a problem. It's doubly annoying in AWB, where you not only get a window telling you that you have messages, but the whole process gets stopped until you address it. Not that I'd change that, I'm just sayin'. --Auntof6 (talk) 05:53, 11 June 2011 (UTC)[reply]
  • Indeed. It strikes fear into into the most established users :P You're not going to succeed in changing it for everyone (it's too engrained now), but you could probably restyle yours to be more like a central notice. - Jarry1250 [Weasel? Discuss.] 10:34, 11 June 2011 (UTC)[reply]
    • I use a lovely light blue and even then it's still bloody annoying :P I'm not so confident with CSS, I wouldn't know how to make it look like the central notice, though it would be nice if this were the default. I've come across quite a few users who'd rather get emailed. —James (TalkContribs)9:46pm 11:46, 12 June 2011 (UTC)[reply]
      • Well now of course you do get emailed (by default). I guess you could probably just hide it? Or make the background white? Those are relatively simple operations. - Jarry1250 [Weasel? Discuss.] 13:18, 12 June 2011 (UTC)[reply]

"Mark all edits minor by default" is now disabled for all users

Just a heads up, "Mark all edits minor by default" is now disabled for all users, this is for the completion of bugzilla:24313. For users that want this facility, you can use the WP:Userscript that can be found here or several from the original VPT discussion. Peachey88 (T · C) 07:38, 11 June 2011 (UTC)[reply]

I've gone ahead and fixed the heading title; I changed "not" to "now". Graham87 07:50, 11 June 2011 (UTC)[reply]

Redirect capitalization tool

I have just discovered that "Google Phone" and "Google phone" redirect to different places. When a redirect is being created, maybe it would be useful if there were a built-in Wikipedia tool that checks alternative capitalizations and prompts the contributor if they already exist; if alternative capitalizations do not exist, then the tool could ask the contributor if he/she wants to create them. LittleBen (talk) 14:05, 11 June 2011 (UTC)[reply]

Feature request should be made in bugzilla:. —TheDJ (talkcontribs) 13:57, 12 June 2011 (UTC)[reply]

DNS alias for moble wikipedia

http://en.m.wikipedia.org/ is the URL of the mobile-phone/small-device optimised version of wikipedia. However, the similar URL http://m.en.wikipedia.org/ is not supported. It would be nice if they latter was a DNS alias to the former. Or http://m.en.wikipedia.org/$FOO is a HTTP 301 (Moved Permanently) redirect to http://en.m.wikipedia.org/$FOO. -- CS Miller (talk) 16:44, 11 June 2011 (UTC)[reply]

I am fairly certain that is a bugzilla sort of request. Killiondude (talk) 22:48, 12 June 2011 (UTC)[reply]
I filed bugzilla:29364 about this. --MZMcBride (talk) 00:48, 13 June 2011 (UTC)[reply]

Image thumbnail purge request

The thumbnail for the image File:Barclays_Cycle_Hire_bike_handlebar.jpg, which is used on Barclays Cycle Hire is corrupt (has black bands through it); the full-size image sees to be fine. I've followed the instructions at Wikipedia:Purge#For_images to no avail. Could someone force a purge for me? CS Miller (talk) 16:53, 11 June 2011 (UTC)[reply]

The image itself seems corrupt; it generate bands at any but the original resolution. Edokter (talk) — 18:30, 11 June 2011 (UTC)[reply]
I'm not sure whether it's the image or the scaling algorithm. And the band are visible only at size of 482 px of smaller. User<Svick>.Talk(); 22:42, 11 June 2011 (UTC)[reply]
This is bugzilla:24854, an upstream bug in the image conversion software ImageMagick that Wikipedia uses. Work around is to convert the image from CMYK to RGB and reupload. —TheDJ (talkcontribs) 13:56, 12 June 2011 (UTC)[reply]
Shouldn't all images be uploaded in the RGB colourspace rather than CMYK? CMYK is for printed images not for screen display. – ukexpat (talk) 20:28, 15 June 2011 (UTC)[reply]

Category search / Related search

The search box in Wikipedia doesn't search categories for a search term by default. Would it be helpful to have an "(include) categories" option check box next to "search"? It seems that a "topic name" can appear as a category name but not as an article name, making it difficult to find.

A related idea: sometimes there are many nearly-identical articles about the same topic. If they have not been placed in the same category then they may be difficult to find. Google is perhaps the best tool for finding such articles. Would a "use Google to search" option be useful (if both Google and Wikipedia agreed to permit this)? LittleBen (talk) 19:55, 11 June 2011 (UTC)[reply]

Both exist already. In the first case, don't enter any search criteria at first, but instead click the magnifying glass, and then click on Advanced. This produces a series of checkboxes headed "Search in namespaces:", so you can select and deselect any one or more of the twenty namespaces before entering your search criteria. As regards Google search restricted to Wikipedia: see here. --Redrose64 (talk) 20:15, 11 June 2011 (UTC)[reply]
At Special:Preferences#preftab-6 you can choose to include categories in searches by default. PrimeHunter (talk) 00:09, 12 June 2011 (UTC)[reply]
Thanks for pointing out that you can get "Advanced Search" options by doing an empty search. Maybe most ordinary users would never work this out. Wouldn't an "Advanced" link next to the "Search" label above the search box on Wikipedia pages make it more obvious that advanced search options are available? LittleBen (talk) 03:16, 12 June 2011 (UTC)[reply]
Nah, link clutter is bad. There are already way too many things to click in the Wikipedia interface. Most people will never need this either. Help:Searching explains all this and is linked from the Search results page. —TheDJ (talkcontribs) 13:51, 12 June 2011 (UTC)[reply]
I don't know, with the instant search thing now (Open Search?), getting to the "advanced search" functionality isn't exactly intuitive. I agree that link clutter is bad, but... I had no idea about the empty search technique, for example (thanks for the tip Redrose64, by the way).
— V = IR (Talk • Contribs) 14:04, 12 June 2011 (UTC)[reply]
Surely it's unacceptably unintuitive. Google would surely not display a "search options" link next to its search box (on Google.com) if that really were "unacceptable link clutter". LittleBen (talk) 14:44, 13 June 2011 (UTC)[reply]
Perhaps we should add an "Advanced" element below "containing" in the dropdown. —TheDJ (talkcontribs) 22:37, 15 June 2011 (UTC)[reply]
Sounds like a good idea, to me. :)
— V = IR (Talk • Contribs) 23:53, 15 June 2011 (UTC)[reply]

Javascript banner lag is driving me insane

Every link I click on, the page loads and then (after a half-second delay) shifts down 40px to accomodate the "Wikimedia Board of Trustees" banner or some other crap. There has to be a way to incorporate these banners into the source code rather than making them load dynamically. I'm constantly clicking on wrong links because the banner will move things around as I'm trying to click them. —Designate (talk) 20:40, 11 June 2011 (UTC)[reply]

Nope, pages are cached, so adding them into the sourcecode would invalidate ALL cached pages, when enabled or disabled. (Which would bring down wikipedia). —TheDJ (talkcontribs) 13:45, 12 June 2011 (UTC)[reply]
Agreed, but is there anything else that could be done to stop the content jumping up and down as the page loads? --DanielRigal (talk) 20:24, 12 June 2011 (UTC)[reply]
The CSS could be changed to leave a blank padding area while the content loads. That depends on it being a consistent size. —Designate (talk) 20:44, 12 June 2011 (UTC)[reply]
For those of us that don't have JS running a blank space that never get populated would be a bad idea. HumphreyW (talk) 20:48, 12 June 2011 (UTC)[reply]
That sounds less annoying than having the page jump around. —Designate (talk) 23:07, 12 June 2011 (UTC)[reply]

Need to have a more details box.

I am not sure if this can already be done in Wikipedia or not.

Not all readers of Wikipedia share the same intellect level. Some need more explanation to get the point. All these details on the other hand will put off more intelligent readers and will make the article too long. One way to get around this is to provide (say) a small '+' icon in places which contain more details. Readers who need to read that can click it and reveal a floating tool tip. IMHO this will allow to maintain article's brevity but at the same time will be accessible to all readers.

--Apple Grew (talk) 06:22, 12 June 2011 (UTC)[reply]

This is what links are for. If you don't understand a concept, you can click a link that explains it in more detail. User<Svick>.Talk(); 10:42, 12 June 2011 (UTC)[reply]

Template fix needed for new user pages

Could an expert please look at Template talk:New user bar - the template is currently showing an incorrect "[Edit]" link on lots of new user pages. This is a fully-protected template, so ideally it needs someone who is an expert and an admin. -- John of Reading (talk) 11:29, 12 June 2011 (UTC)[reply]

Done. —TheDJ (talkcontribs) 13:48, 12 June 2011 (UTC)[reply]

Could someone with the necessary technical skills help out here. Citations to a baseball stats resource are causing anomalous footnotes. Thanks. David in DC (talk) 23:11, 12 June 2011 (UTC)[reply]

You can't remove the bullet from the template, so it's best to just use plain text, like this. Gary King (talk · scripts) 23:59, 12 June 2011 (UTC)[reply]
Thank you for starting me off. I've fixed all of the anomolous entries. David in DC (talk) 17:01, 13 June 2011 (UTC)[reply]

Stuff not showing up for users who are not logged in

Hi, I tried to describe the issue at the help desk here, they pointed me in this direction. And I tried a bit of debugging after I saw it reported initially by someone else here.

I'm pretty sure it's not a local cache or a particular browser issue (I've tried IE8 and Mozilla) - and I've seen it multiple times, despite frequently clearing my cache (as I have a fairly old machine - still running xp).

I had a quick search in these archives - and the closest article I could find was this

Would I be right in thinking that it's something to do with a backlog for the 'squid' (whatever that might be?) not updating to the 'full data' rather than the 'fairly recent copy data' for people who are not logged in? And trying to look at the implications would that mean all IP addresses (rather than logged in users) trying to edit something have a risk that their edit is trying to be applied to something older than the full data ? EdwardLane (talk) 08:34, 13 June 2011 (UTC)[reply]

Page load never finishes (again) - is it the Wikimania banner?

The issue previously described in WP:Village pump (technical)/Archive 88#Page loading never stops has started occurring again. I am on Firefox 3.6.17 (I won't use FF 4 because it is so slow as to be unusable). With IE 8 the page eventually loads, but with a significant delay - this delay seems to wait for the Wikimania 2011 banner to appear. Note the banner never gets to appear in FF.

Google Chrome seems to behave properly (though for unrelated reasons I don't wish to use it). Anyone know what's going on? Jan1naD (talkcontrib) 10:55, 13 June 2011 (UTC)[reply]

FF has now started behaving (and the banner does not appear). Jan1naD (talkcontrib) 10:58, 13 June 2011 (UTC)[reply]

Rollback from watchlist

Resolved
 – ...sort of. The script apparently works, but something in my setup is causing it to fail. —DoRD (talk) 14:16, 15 June 2011 (UTC)[reply]

I have had User:Zvn/confirmwatchlistrollback.js in my .js for quite a while to prevent accidental rollbacks from the watchlist. Unfortunately, it has stopped working and I accidentally rollbacked an edit while browsing on my iPad this morning. I tested it a while back using FF4 and it doesn't work there, either. I don't often use rollback from the watchlist, so I don't know when it stopped working.

When I click on rollback, the script gives a confirmation popup as designed, but then goes ahead with the rollback without waiting for an answer. Perhaps someone who knows about these things can figure out what's wrong. P.S. I'm using Vector. Thanks! —DoRD (talk) 12:25, 13 June 2011 (UTC)[reply]

Or, if another solution is available, I'd be glad to consider it. —DoRD (talk) 15:53, 13 June 2011 (UTC)[reply]

You could remove it altogether or you could use a different account on your iPad that doesn't have the rollback privilege (using a watchlist token to access your main watchlist). –xenotalk 16:01, 13 June 2011 (UTC)[reply]
I also found this more specific hack, but I like(d) the ability to abort a rollback from the watchlist even using FF. I have another unused account, but occasionally do admin stuff on the iPad. —DoRD (talk) 17:28, 13 June 2011 (UTC)[reply]
I'm sure someone more script-savvy can fix the actual root cause for you. –xenotalk 22:17, 13 June 2011 (UTC)[reply]
Could you bypass your browser cache and try again? Amalthea 13:36, 14 June 2011 (UTC)[reply]
With FF4, I'm still getting the same behavior. With Safari 5 and mobile Safari, the popop waits for a response, but goes ahead with the rollback even if you click cancel. —DoRD (talk) 13:57, 14 June 2011 (UTC)[reply]
Well, works for me, in both these browsers (Windows versions). Might be a conflict with one of your other scripts or gadgets? Amalthea 14:43, 14 June 2011 (UTC)[reply]
I'm at a loss, then. I disabled everything except the above script and still encountered the same behavior. I didn't change any of my gadgets, though, so the conflict may be there. I give up and will instead disable links on mobile Safari. Thanks for trying, though! —DoRD (talk) 16:04, 14 June 2011 (UTC)[reply]

Secure server login

Over the past 24 hours or so, I've noticed that the screen for logging in no longer offers a link to the login page that uses the secure server. Is there still an option of logging in via the secure server? If so, how should one do it? Thanks. --Tryptofish (talk) 20:17, 13 June 2011 (UTC)[reply]

It seems to come and go. --Tryptofish (talk) 20:24, 13 June 2011 (UTC)[reply]
The main page is here and the login link is here. For most URLs on Wikipedia, you just alter the first bit from http://en.wikipedia.org/ to https://secure.wikimedia.org/wikipedia/en/ --Redrose64 (talk) 21:43, 13 June 2011 (UTC)[reply]
OK, thanks. What I notice, though, is when one goes to the login page for the regular server, there is usually some text just below the box where one logs in, under the title "Secure your account:", and beginning with the line "Consider logging in on the secure server" with "secure server" blue-linked to the login page you link above. Over the past 24 hours or so, that text has been disappearing (just leaving the login box with blank space below it) and reappearing. I'm getting the impression that this disappearance may be a bug rather than a feature. (By the way, I use Firefox 4.0.1.) --Tryptofish (talk) 21:57, 13 June 2011 (UTC)[reply]

Office hours to discuss Article Feedback Tool

Hi everyone, I just wanted to announce that this Thursday the 16th at 18:00 UTC, there will be an IRC office hours concerning the Article Feedback Tool (full documentation) which is currently in experimental partial deployment. I'll be moderating mainly for Erik Möller, but hopefully we'll be joined by most of the Foundation staff who've contributed to this feature. Just to clarify, we want to stick to two general topics:

  1. The strategic goals the feature aims to address. In other words, its purpose.
  2. Plans for developing and deploying it further.

If you have bugs to report or specific design feedback, as always Bugzilla and MediaWiki.org are respectively the best places to discuss those two things. For the office hours we'd like to stick to a broader explanation of the feature and its future. Time conversion links and other documentation for IRC office hours are on Meta. Looking forward to chatting, Steven Walling at work 22:17, 13 June 2011 (UTC)[reply]

Logs posted. Thanks to everyone who attended. :) Steven Walling at work 20:01, 16 June 2011 (UTC)[reply]

Problem resetting password

Hi, I've got an account "dez" that I haven't used for a while, and I have time between jobs to catch up on some edits etc, and when I tried to login my password didn't work, so I used the "email reset password" button but the email never came through ( to dez@blanchfield.com.au ) for some reason, I've chatted to a bunch of great folk ( SudoGhost / Hersford ) and they recommended I come here for help. Is there someone here with access who can check that my "dez" account is linked to my "dez@blanchfield.com.au" email address as I'm doing a "tail -f /var/log/maillog" and nothing from wiki has come through yet? Help please ;-) Many thanks, Dez Blanchfield. — Preceding unsigned comment added by 203.16.243.213 (talk) 04:10, 15 June 2011 (UTC)[reply]

The login dez (talk · contribs) doesn't seem to have ever been used. However, it is registered; so you really have two choices - (a) create a different login or (b) usurp the Dez account.
BTW - per the talk page guidelines, please start new discussion threads at the bottom of discussion pages; and always remember to sign your postings with four tildes ~~~~. --Redrose64 (talk) 10:03, 15 June 2011 (UTC)[reply]
I'll add that there isn't a way to check what, if any, email address was used when you registered that account, so I'd also suggest that you simply register a new one. However, there are a bunch of accounts starting with Dez, so be sure you're trying the correct one. —DoRD (talk) 14:13, 15 June 2011 (UTC)[reply]
"dez" has no email set, so you can't reset the password. –xenotalk 14:20, 15 June 2011 (UTC)[reply]

erroneously prompted new EL captchas

I've inquired about this issue over at Wikipedia:Help desk#New external links false positives and was pointed here. To reiterate:

I frequently get the non-autoconfirmed new EL captcha, although I haven't added any new external link. It happens seemingly at random, and may happen even on the simplest of edits (e.g. correcting a typo). One example diff where this happened to me is here.

I've searched through BugZilla but couldn't find a ticket about this. Is it a known issue, or should I file a new bug report? Asking since there appears to be no markup/editing-related explanation. --213.196.218.59 (talk) 15:59, 15 June 2011 (UTC)[reply]

List identifiers in small font

When a list is created in small font, the list numbers or bullets appear in regular size font. Any workaround suggestions, besides numbering the list manually? Thanks. — Preceding unsigned comment added by Yuckhil (talkcontribs) 00:55, 16 June 2011 (UTC)[reply]

  1. First numbered item
  2. Second numbered item

  • First bullet list item
  • Second bullet list item

The bullet cannot be changed (it is an image), but the numbers can. The best way to do so is using a <div>. Edokter (talk) — 01:32, 16 June 2011 (UTC)[reply]
  1. First numbered item
  2. Second numbered item

JavaScript disappearances...

I seem to be having problems with JavaScript add-ons in the last couple of weeks. First, the script I was using to hide top edits in my contrib list stopped working - Nihiltres managed to fix it, briefly, but it's since stopped working again - and now I seem to have lost the HotCat function. Bearing in mind that I am completely an "end user" and wouldn't be able to tell a variable from an if/else function: 1) what has happened - is it something I've done, or a change in WP functionality, and 2) what can I do to get those things working again? Note that I'm using the monobook skin on Safari 5.0.2, ands I have tried all the standard IT consultant things (purging cache, logging out and back in, turning computer off and on, etc.) Grutness...wha? 10:56, 16 June 2011 (UTC)[reply]

CSS margins, width help

I am trying to make template:weather box play nice with right floating infoboxes, but I am afraid I do not know enough CSS to make this happen. The basic issue is that the current weather box sets the width to 90%, which appears to be based on the browser window width, so this causes lots of white space (see here). What I would like to do is have it push itself up and reduce its width to sit nicely in this open space. I was able to do this by removing the "width:90%" statement and adding "margin-left:5%" and "margin-right:5%" in the version shown here, if the sandbox is still the same as how I left it. The problem is that now the box does not expand and contract when you change the width of the browser window, so it is more narrow in the case when there is no floating infobox on the right (see here vs. here. So, I am wondering if there is a way to have it always at a width of 90%, but have that width computed after subtracting the right floating elements, or if the only solution is to just reduce the width. Thank you. Frietjes (talk) 20:48, 16 June 2011 (UTC)[reply]

Tough one. Ambox does what you want; it has a 10% margin set, but no widht. It's trick is to expand the right-side table cell to 100%, to cause the ambox table to take all available width (minus the 10%), even next to a floating element (see Template:Ambox/testcases). But where ambox only has two cells, the weather box has many. And trying to set it's header to 100% didn't work either. Edokter (talk) — 21:20, 16 June 2011 (UTC)[reply]
Yes, I got the idea for the "margin-left, margin-right" trick from ambox. A way around this might be to wrap the entire box inside another element (e.g. a div or a table) which expands the way we want it. I will do some more testing. Thank you. Frietjes (talk) 22:50, 16 June 2011 (UTC)[reply]
One method that works, using your hint, is to just add a very short blank row with set dimensions and at least two cells, like this. To not make it so obvious, I then removed merged it with the row above. It's sort of a hack but closer to what we want. The only issue with this is that I had to specify the relative width for the far left cell, which isn't that bad, but is a slight change in the default format. Frietjes (talk) 23:17, 16 June 2011 (UTC)[reply]

"Google wikipedia" not working

The useful facility {{google wikipedia}} has stopped working - Search Wikipedia with Google now returns a message: "We're sorry... but your computer or network may be sending automated queries. To protect our users, we can't process your request right now." Can anything be done to revive it? JohnCD (talk) 15:47, 17 June 2011 (UTC)[reply]

Fixed, although the page displays with some text overlapping (the radio buttons) for some reason, but only if you link directly to a search, e.g. Search Wikipedia with Google for: search term. - Kingpin13 (talk) 16:05, 17 June 2011 (UTC)[reply]
Thanks, that's much better than nothing, but I still usually (but not always) get the first item in the list overlapping the radio buttons even if I use the blank search form Search Wikipedia with Google, and then the hyperlink to the first item doesn't work. (Firefox 4.0.1) JohnCD (talk) 17:25, 17 June 2011 (UTC)[reply]