This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.
I'm trying to change the font for PRE tags, but I can't since it's currently using !important to override anything I set, and the code for that loads after my monobook.css. Could someone suggest a workaround so that I can change the font used for PRE tags? The discussion is here: MediaWiki_talk:Geshi.css#Text_size_fix. Gary King(talk · scripts)03:28, 6 August 2010 (UTC)
The solution is simply to increasy CSS specificity of your code, e.g. body pre { font-family: ... !important; } . — AlexSm05:30, 6 August 2010 (UTC)
Since July 31, I have been getting proxy errors when using secure.wikimedia.org. It's the same proxy errors as described in a past Village pump post from March (I got the same errors in March too). According to the Server admin log, I noticed squid (proxy software) was played with on July 31.... It just took me 15 minutes to save an article because I kept getting proxy errors. Bgwhite (talk) 05:21, 6 August 2010 (UTC)
SUL issue
I know this is probably not the correct place to address this, but someone here should be able to point me in the right direction.
I am User:Horologium on all WMF projects except for de.wp, where another editor grabbed the username about the same time that I registered it on en.wp. (He's User:Horologii on en.wp). I was not able to fully unify my login because de.wp will not usurp accounts with valid contributions, even if the account has been dormant and the e-mail feature disabled (which the user had done on both the de.wp and en.wp accounts). However, due to the wonders of importing articles from other projects, now the last four contributions listed at http://de.wiki.x.io/wiki/Spezial:Beiträge/Horologium (not my account) are *mine*, originally made here on en.wp. (The edit summaries gave it away; they're in English, and I recognized them because I made them here.) Isn't attributing my edits to someone else a violation of the GFDL and/or the CC-BY 3.0 license? Along those lines, wouldn't this be a justification for usurping a username which hasn't made an edit since October 2006, since it's now screwing up attribution? FWIW, I have valid edits on 40 different projects, and my SUL is linked to 10 more. Horologium(talk)00:42, 3 August 2010 (UTC)
I'm not sure that there's any place that you can petition other than locally at de.wp. Even m:Steward requests/SUL requests will only handle requests for wiki without active bureaucrats. There ought be a bugzilla: filed on this though, if there isn't already one. –xenotalk00:58, 3 August 2010 (UTC)
Also, only a Wikimedia sysadmin could separate your edits imported into de.wp from those made by the original user on the German Wikipedia, IIRC. A way to get around this problem would be to rename the user on de.wp, delete your imported edits then re-import the edits. Graham8705:09, 3 August 2010 (UTC)
Your best bet is to ask the user on de:Benutzer Diskussion:Horologium if he would be willing to change his username to "Horlogii" on dewiki as well. I've had success with this on fiwiki myself, so it's always worth a try. Zhwiki also renamed another account for me who hadn't logged in for several years after a small number of edits, but most projects have a policy that they won't forcefulyl rename users with nontrivial contributions without their consent. --Latebird (talk) 08:27, 3 August 2010 (UTC)
That won't happen. He had a big chunk of edits (over 650) in a six-week spurt in July and August of 2006 and then disappeared. I dropped a template on his user page (on de.wiki) addressing the name change idea, with no response for over a year. I got a very hostile response from an IP editor 13 months later on my talk page here, but I don't know if it is the same guy or not. In any case, Raymond, a de.wp steward, flatly told me that there are no non-consensual name changes there if there are valid edits, but I can't get consent because there is no way to contact him. What I am concerned about is the attribution issue (not for my case so much as from a legal standpoint; my edits were trivial on of the articles in question—single edits for copy-editing or other minor tweaks). Looking at the edit histories of the articles in question (two are in user space, and two are in main space), it's clear that the articles were imported from the English Wikipedia (history and all),and then translated into German. Any editors with the same "username in use" issue is going to run into the same problem, and it needs to be addressed. What happens if they import an article on which I have a substantial number of edits? There are six mainspace articles to which I have more than 100 edits (and about a dozen decent-size articles to which I am the only substantial contributor); if any of them were to be ported over, I would expect the attribution to be correct, as would most contributors. Horologium(talk)11:13, 3 August 2010 (UTC)
In that case there's little that can be done locally. The issue with imported article histories resulting in misleading attributions may need to be addressed on meta. Actually, I would really classify that as a bug. --Latebird (talk) 08:08, 6 August 2010 (UTC)
One, what is the "Editor" userright as seen here but not even listed in the drop-down? All those users were "autopromoted" in mid-June, but I cannot find any documentation of this right anywhere onwiki. Does this have to do with pending changes, and what does the right do?
Two, on Firefox 3.6.3 in Jolicloud (Linux-based), I don't get any borders for certain templates, particularly image licenses and CSD tags. Everything else, including tables, show up fine, and they show up in Commons and in both the Chromium and Opera browsers, but not on Wikinews. For an example, see here. I'm still trying to figure out how to update my Firefox on this new OS, but the issue doesn't happen in Firefox 3.6.8 in my Ubuntu 10.04 setup. Any ideas?
1, this was due to the introduction of Pending Changes. Accidentally, the usergroup editors, which is used in some Flagged Revisions configurations, was also added (or rather the autopromotion routine had not been disabled). This was since fixed.
On your Edit toolbar, on the right hand side, you should see the word "Cite" with an arrow. Click on that. It triggers a drop-down list of templates on the left-hand side. Click that arrow to find the Cite Web. Maile66 (talk) 22:18, 7 August 2010 (UTC)
I can't see the word cite. When I go into edit the top toolbar has Bold, Italic, other symbols, and then the Advanced, Special characters and Help drop-downs, but I can't see 'cite' there either? Eldumpo (talk) 23:09, 7 August 2010 (UTC)
I created subsections of the Cast section in the article That '70s Show. I want at least the subsection header parentheticals, like "Seasons 1-7", to be unbold, and possibly the entire subsection header. I've looked all over but can't find a way to do this.--Bbb23 (talk) 16:53, 8 August 2010 (UTC)
I think you're approaching the problem from the wrong direction. Information like seasons, in which particular character appeared, shouldn't be part of the heading, they should be in the section itself. I also think that the actor shouldn't be in the heading either. And headings shouldn't contain links. Svick (talk) 19:38, 8 August 2010 (UTC)
LOL. You put it back to essentially the way it was before I started. My purpose in creating subsection headers was to make it easier to edit each character rather than to have to edit the entire section just to change one character's summary. Based on the caveat about links in headers, I'll have to think about this some more. BTW, you don't happen to know the answer to my original question? :-) --Bbb23 (talk) 20:08, 8 August 2010 (UTC)
You could use code like this: ===bold <span style="font-weight: normal">not bold</span>===, but I really don't think it's a good idea to use it. Svick (talk) 22:53, 8 August 2010 (UTC)
The text “bold” will be rendered bold, and the text “not bold” will be rendered not bold. The rest is markup. You have to substitute those two pieces of text with the text you want. Svick (talk) 23:34, 8 August 2010 (UTC)
I know it has something to do with wgCanonicalNamespace, however I am not sure exactly where/how to add it, and trial and error has not helped. The script in question is at User:Avicennasis/nowcommonsreview.js. I would like to hide the tabs added unless the page is in File space. Any help would be appreciated. :) Thanks! Avicennasis @ 18:48, 8 August 2010 (UTC)
Wow! I spent too much time trying to add "wgCanonicalNamespace" to the actual script, I never once thought about adding it to my monobook! Such a simple yet elegant solution. Thanks! Avicennasis @ 20:45, 8 August 2010 (UTC)
Hello again. :) I am wondering if the following is possible. This script (User:Bradv/AjaxPatrolLinks.js) uses AJAX patrol links once you are on a new article, so you can click on "mark as patrolled" without leaving the page. This script (User:Mr.Z-man/patrollinks.js) adds patrol links to all unpatrolled pages when viewing in Special:NewPages, however clicking the link will load a new page. The first script seems to find the "mark as patrolled" link by div class (which is always 'patrollink' on new pages), which is not used by the second script. The second script generates the links without a div class, instead using a span id that changes with each link. (i.e., <span id="plink0">, <span id="plink1">, etc).
The million dollar question: is there any way to combine the scripts so I can have AJAX patrol links on Special:NewPages? This would useful on other projects like Commons as well. Avicennasis @ 21:31, 8 August 2010 (UTC)
You can make an edit request on the talk page (an edit request link should appear in the protected notice). That should be in the documentation. ---— Gadget850 (Ed)talk23:57, 8 August 2010 (UTC)
Thank you. I'm familiar with {{Editprotected}}, but that seems a long-winded and unnecessary hurdle. It also means that nomination rationale may not be posted until several hours after the notices are - which would lead to confusion and possible the removal of the notices. {{Editprotected}} is not mentioned in the {{Tfd}} and {{Tfm}} documentation. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits00:07, 9 August 2010 (UTC)
When the above template is added to a page where "<pagename>/Archive" exists, the template outputs "Sections with no replies in X days are automatically moved here", with the word "here" linking to the "/Archive" page; this is also true when only one archive exists, which is not a problem. The former behaviour was probably a legacy of the pre-October 2007 method of archiving the village pump, where messages were moved to a "/Archive" page for one week then discarded. I did a Google search for the phrase "are automatically moved here" in Wikipedia, and all the results returned pages where the "/Archive" page was either a redirect or a misplaced talk page archive. I am aware that the Google results might not tell the whole story, so I'm going to add a category to Template:Auto archiving notice to signify all the pages marked with that template that have an existing "/Archive page. If this category shows no legitimate examples of archiving *only* to a "/Archive" page as happened with the old village pump within two days, I will remove the functionality from the template. I will also note this discussion on Template talk:Auto archiving notice. Graham8714:18, 6 August 2010 (UTC)
To provide evidence to support work on the TIME element in HTML5, it would be useful to know how many instances of {{Startdate}} haev a YYYY-only or a YYYY-MM-only value entered. How can this best be determined - tracking categories? A database grep? Something else? Can someone assist, please?
—Preceding unsigned comment added by Pigsonthewing (talk • contribs) 00:05, 7 August 2010
I don't see you couldn't read the documentation yourself; looks like this query shows all cases with the 2nd parameter (column "1") being empty, and you can do another query for the 3rd parameter. — AlexSm18:46, 9 August 2010 (UTC)
I've had the same problem. Also, I don't know if this is related, but the article traffic count link has disappeared from my left-column tool box (W7, Ff3.6.6). Lampman (talk) 06:45, 9 August 2010 (UTC)
Committed identity
Just a quick question to reassure myself, this might not be a technical one either. My home (main) account is here at en-wikipedia. I have a Unified login over 38 project sites. I have my Committed identity posted here and a userbox on my user page. There is no need to post the SHA-512 string on any other project sites since they are all tied to this home account, Correct ? Mlpearcpowwow22:30, 8 August 2010 (UTC)
Crossposting from the helpdesk as there may be a technical issue here. Currently this article is about 110KB. It takes around 10 seconds on my browser (Firefox 3.6.8) to start loading and another few to finish rendering. For comparison, Line of succession to the British throne which is currently 397KB loads completely in a couple of seconds flat on my connection. What's going on here? Exxolon (talk) 01:00, 9 August 2010 (UTC)
User:Exxolon/Colorado is a no image version. Colorado has a huge number of high resolution images in it, a fair number ranging from 1MB to 2.5MB so they could've been the cause, but my version is no faster. Exxolon (talk) 01:24, 9 August 2010 (UTC)
I'm guessing using categorytree isn't particularly helpful in this case. I have removed it and as far as I know, we don't integrate things like that into the content anyways usually. —TheDJ (talk • contribs) 02:23, 9 August 2010 (UTC)
That edit seems to have fixed the problem - thanks. However we now need to work out what the problem is in this code you removed - if it's used elsewhere other articles could be affected.
I'm not over familiar with this markup - is there any documentation on it? Could it be affecting other articles in the same way? Exxolon (talk) 17:54, 9 August 2010 (UTC)
Like has been said many times before, complex articles and non-standard formatting (triggered by non-default preferences or something like categorytree) will always cause long displaytimes, because the servers have to 'calculate' a separate page for you and that simply takes long. Efforts are always ongoing to improve that, but that is simply the way the technical infrastructure works atm. —TheDJ (talk • contribs) 23:07, 9 August 2010 (UTC)
Block expired but user still gets block message
Loosmark (talk·contribs) was blocked for 55 hours at 19:22, 6 August; that should have expired hours ago, the block log shows no later block, and the usual block notices don't appear, but user is still getting "This account or IP address is currently blocked" - see unblock request on his talk page. JohnCD (talk) 10:22, 9 August 2010 (UTC)
I see that there are no redlinks on the page, with the exception of the hat-note, which I've just removed. What browser/OS are you using, and does this problem resolve itself if you use a different browser/OS? Are there any scripts you are using that may be causing the problem? Intelligentsium23:51, 9 August 2010 (UTC)
I'm an admin at the Afrikaans Wikipedia and since the cross over to the new interface, our version of Template:Featured article (af:Sjabloon:Voorbladster has not been displaying correctly. I think it might be solved by copying some of the syntax in your template, but, if I'm not mistaken, we will have to define the "metadata topicon" class referred to in the template. However, I can't find where it is defined here at the English Wikipedia. Anrie (talk) 08:10, 6 August 2010 (UTC)
Thanks for the quick reply! I didn't expect anything so soon, which is why I only had a look here again today. Works fine now, thanks for the help! Anrie (talk) 08:43, 10 August 2010 (UTC)
I opted in for Soxred93's edit counter, put Logo, checked to see if it worked. See my optin page. It worked. Is that code capable of comprimising my account? Besides, should I opt in for other users? Us441 (talk) 17:29, 10 August 2010 (UTC)
It won't compromise your account. You can't opt-in for other users because you can't edit their .js pages (it would also be discourteous...). –xenotalk17:31, 10 August 2010 (UTC)
Has anyone else encountered a situation with Wikipedia where the lines used to strike-through text appears far below the actual lines the tag is used on? I just started seeing this once I upgraded to Firefox 3.6.8 (on Ubuntu Linux), using the Classic skin for Wikipedia. I think I encountered a bug, but I don't know if it is in the browser, the skin, or some default configuration I need to fix on my computer. I'd like to first know this problem isn't unique to my computer -- or that I'm not the first to see this -- before I open a bug report. -- llywrch (talk) 18:59, 9 August 2010 (UTC)
I have no problems with Firefox 3.6.8, although I am running Windows 7, and use the Vector skin. I took a look at AN/I with the classic skin, and I still don't see a problem (except that the classic skin is hideous, but I digress...) Horologium(talk)20:22, 9 August 2010 (UTC)
testingtesting these work fine for me on ff3.6.8 on ubuntu with (modified, but not in this regard) Classic skin. Can you cook up a sample sandbox page o cite a ref that this does it for you, so I can verify that one too. -- Finlay McWalter • Talk20:27, 9 August 2010 (UTC)
I have yet to see with my current browser a use the strike-thru tags which don't exhibit this odd behavior. Have a look at your use of <s> above: on my computer the strike-thru lines appear in the first follow-up paragraph in the thread below. -- llywrch (talk) 17:19, 10 August 2010 (UTC)
The strikethru is displaying correctly for me (ff3.6.8, Ubuntu 8.04, classic and vector skins).
1) Is the strikethru appearing at a fixed height below it's intended placement? ie. If you resize the browser window to make it narrower, does the strikethru remain at the same number of pixels/lines below, or does it change relative-location (always appearing in the wrong thread)?
2) Just to confirm, have you tried completely clearing browser cache? Ctrl-F5 in firefox.
1) Yes. The interesting thing is that if I scroll up or down so that the stricken text is off the screen, the strikethru will vanish -- just as an artifact that doesn't get overwritten when you scroll up/down too fast will vanish if you click on "page up"/"page down" then return.
2) Just tried that: no difference. I don't think this is due to garbage in the cache: this effect is too consistent.
There are a few more details about my configuration I need to share. The version of Ubuntu I'm using is an old one -- 7.04. And the way I installed Firefox was to download the package from Mozilla -- I was given a tarball, not a .deb -- which I unpacked into a directory in my home folder. From there I invoked the browser by "$HOME/firefox/firefox %u". Somehow all of my existing settings & browser history was imported to this new binary, although I also have a lot of the bog standard settings which the Mozilla people add but aren't in the Ubuntu Firefox package. In other words, I have a different undocumented config file in the mix than you are Quiddity -- but the only way I know that is where I need to investigate is by asking questions like this. -- llywrch (talk) 05:54, 11 August 2010 (UTC)
Why not delete all that and install it properly? Firefox is in Ubuntu's package management system. apt install firefox if you don't like the GUI. OrangeDog (τ • ε) 08:50, 11 August 2010 (UTC)
Because I consider Debian apt black magic. I started fooling around with Linux back in the early days, when the install process was: download source, unpack tarball, type "make", & if there wasn't any problems then type "make install". And all of that stuff done to make Ubuntu "user friendly" only frustrates those of us who know how Linux works in trying to troubleshoot problems. But if you really want to hear my opinions on different operating systems & why each of them suck, contact me by email & we'll spare everyone Yet Another Replay of that topic. -- llywrch (talk) 05:31, 12 August 2010 (UTC)
Bottom link on the top
Hye. I miss a bottom link (#bottom) in allmost every Wikipedia page, not just talk pages. It could be placed on the top beside Read, Edit and View History links. --Janezdrilc (talk) 16:47, 11 August 2010 (UTC)
I think it would be useful to put the Hide user option in the Recent changes options. It's because some of them are very active and spend to much space in the Recent changes page. It would be like this: Hide user: XY (you write down in a empty field). Maybe it's possible to hide even more users. --Janezdrilc (talk) 16:55, 11 August 2010 (UTC)
(Cannot contact the database server: Unknown error (10.0.6.26))
Is anyone else having trouble - edits are going through but action=submit is not taking you back to the page? (Takes you to the "This wiki has a problem" page?) –xenotalk20:00, 11 August 2010 (UTC)
For some reason the database that hosts Commons went down and rebooted. Because Commons is used a lot by the other servers, the other wiki's will have a lot of problems as well. Mark and domas are already looking into the problems. —TheDJ (talk • contribs) 20:05, 11 August 2010 (UTC)
Can't watch video files on Wikipedia from a Mac running OS X version 10.6.4
But... I can watch all video files perfectly well on Wikimedia Commons. This has been true for maybe 6 months or so. In the last 2 days I have tried to follow the instructions on this page Wikipedia:Media help (Ogg) without any success. On this page [1] it says "MplayerOSXB8r5 is fixed" but I can't seem to download that "fixed" version successfully, only older versions from May of this year. In particular I am trying to view this video: [2]. The editor who created the file did not load it up to Commons, so I can't access it there. Thanks for any help you can give me, but bear in mind I am not super tech-savvy on this kind of stuff. All best wishes and many thanks, Invertzoo (talk) 18:21, 8 August 2010 (UTC)
Thanks Cybercobra, I did try that suggestion just now but could not get it to work for whatever reason. I also installed VLC Media Player and can now open ogg files using it, including the one I wanted to look at, but not as part of Wikipedia. I have to drag a file onto my desktop and then open it in. Any more suggestions? Invertzoo (talk) 00:47, 9 August 2010 (UTC)
I usually use Safari. I suppose I should try to get used to using Firefox instead? However, right now Wikipedia videos are not working when I am on Firefox either. Invertzoo (talk) 13:07, 9 August 2010 (UTC)
Thanks for trying to help, I really appreciate your kindness. This morning I uploaded v5.0.1 and yes Java, JavaScript and plug-ins are enabled in the Preferences. Unfortunately there has been no change: I can still watch videos on Commons just fine, but not at all on English Wikipedia. Invertzoo (talk) 19:50, 10 August 2010 (UTC)
You might try clearing your cache and deleting cookies. You may want to try to play the file in a different browser to see if it is a problem with your copy of safari. Snagglepuss (talk) 15:53, 11 August 2010 (UTC)
File:101 California Street plaza and garden.JPG was here and on Commons; I checked the bot upload to Commons, and it looked OK, so I deleted it here under CSD:F8. Normally when I do this, there is a bluelink at the top of the "deleted" screen, which I click to check that the local page refers to the Commons image. In this case, however, it stayed red, so I undeleted the local copy to check, and yes, it looked OK and so did the Commons version. It's not a great issue since it's orphaned here, but I can't see any reason why this file should be any different to the hundreds of others I've done. Any ideas? Cheers. Rodhullandemu17:33, 12 August 2010 (UTC)
Scholarometer is a cool new app to assess the performance of scholars. It would be great to upload their data to the pages of scholars on Wikipedia. As far as I can see, this cannot be done automatically. Please correct me if I'm wrong.
MediaWiki has an extension that does allow one to load data from an external site and publish them on Wikipedia.
Normally when you are typing in the name of an article, the most viewed articles will show up as suggestions in a dropdown box. But recently, it has been showing alphabetically instead, which is a small nuisance. Tell me if you need me to clarify. Access Deniedtalkcontribseditor review01:38, 13 August 2010 (UTC)
Another thing I've noticed is that when searching through the Reference desk archives by using the search box at the top of the page (not the main interface search box) it gives unhelpful results example that include the "prefix:Reference desk/Archives" part as part of the search string, in other words, the prefix: command for searching within archive subpages is acting funny. Nevermind, just noticed the thread above. -- Ϫ02:17, 13 August 2010 (UTC)
Wrong categorizations
As it is known, the pages of each namespace have their own rules about categorization (the main system only for articles, "Wikipedia templates" for templates, "Wikipedians" for user pages, and so on) and shouldn't be mixed. My userpage shouldn't be at "Argentine people", and Jimbo Wales shouldn't be among "Wikipedians". And the case I have seen more frequently, navboxes of "Foo" categorized at "Category:Foo" instead of somewhere within the categories of navboxes. Is there some system around to locate or list such pages wrongly categorized, to work in fixing them into the right categories? MBelgrano (talk) 02:14, 11 August 2010 (UTC)
Are you sure trees for different namespaces should be completely separate? For example Category:Physics templates (obviously a template category) is a member of Category:Physics (article category) and I don't think this is wrong. In any case, I agree that pages in the wrong namespace category are a problem and it should be addressed. Svick (talk) 09:14, 11 August 2010 (UTC)
Actually yes, I'm pretty sure that Category:Physics templates is wrong in Category:Physics (it doesn't further our reader's understanding about physics, does it?) and should be removed from there. Maintenance infrastructure should be strictly kept out of the way of non-editing users. --Latebird (talk) 11:59, 11 August 2010 (UTC)
Although in principal I agree, I think this is one of the smallest problems we have. Is it really worth to spend a lot of time on this? —TheDJ (talk • contribs) 13:11, 11 August 2010 (UTC)
If you don't feel so, then don't spend any time on it. But plese don't tell other people what to do with their time. --Latebird (talk) 12:57, 13 August 2010 (UTC)
New sitenotice-style banners?
Resolved
So far, I have all previous global sitenotice cruft hidden with css. However, when loading a page, something that looks like a ladybird or somesuch briefly appears at the top briefly, then disappears too quick to see it properly. Is someone trying to hack around with sitenotice again? This hasn't happened with any of the previous banners (I only know when they're there as this board gets filled with complaints about how broken they are). OrangeDog (τ • ε) 21:15, 11 August 2010 (UTC)
I'm pretty sure it's just that your personal CSS is taking slightly longer to load. I get flashes as well, but I haven't worried about it. --Izno (talk) 21:17, 11 August 2010 (UTC)
There have been several threads about this. It started happening a month or two ago. What you can do is either use a CSS rule to permanently hide it, or when the page is loading, hit your ESCAPE key in time for the banner to still be on the page but so that the rest of the page hasn't loaded yet, then click the "Hide" button on the banner. It should hide the banner for a week or so. Gary King(talk · scripts)22:03, 11 August 2010 (UTC)
Hi I was just browsing and it came to my attention that John Terry is (was) listed at 187cm (6'2") in his infobox. However if somebody is listed at 194cm like Saša Ivanović they are affirmed to be 6'4.5", not 6'4" or 6'5". I was wondering why this is given that 187cm is a fraction over 6'1.5" and nearer to that than 6'2" which is 188cm. It should be corrected to 187cm (6'1.5") if the infobox converters allow 1/2 inch for other listings. Dr.Blofeld15:08, 12 August 2010 (UTC)
Both of these use conversion templates. Saša Ivanović uses {{height}}; John Terry uses {{convert}}. The two templates appear to use a different default for rounding:
Mmm is this a good idea to have rounding? I mean some people prefer it to be rounded but I think it should be consistent. As a cm converter it should really be more precise to a third of an inch. Isn't it a little redundant having both templates? I think we should have one using the precision of the height template within the converter template. Dr.Blofeld15:35, 12 August 2010 (UTC)
Just as an FYI, I think it was me who forced {{height}} to round in half-inch increments (but I may remember wrong; it was in 2007). The reasoning was that 1/2" is accurate enough for a person's height. {{Convert}} did not exist then in the developed form it does now, so apparently the differences have not been reconciled along the way. Which way of rounding is better, I have no opinion.—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); August 12, 2010; 16:58 (UTC)
1/2 inch would be perfectly appropriate. I agree a quarter of an inch would be a little too much. Can you see to it that the convert template follows the half inch thing? Dr.Blofeld21:42, 12 August 2010 (UTC)
You can set the precision for {{convert}} to a decimal:
{{convert|1.94|m|1}} gives 1.94 metres (6 ft 4.4 in)
The prefix search engine does not currently seem to be working. I either get irrelevant or no results at all. Does anyone else have the same issue? /HeyMid (contributions) 21:52, 12 August 2010 (UTC)
Yeah, that works, I mean the prefix feature where you can search inside articles and archives, etc. It currently doesn't seem to be working. /HeyMid (contributions) 22:03, 12 August 2010 (UTC)
For example, try this. Normally it should give four results, and the results should say "BjörnBergman" somewhere in the archives. I now get four irrelevant results. Depending on what keywords I search for, I may also get no results at all. /HeyMid (contributions) 22:04, 12 August 2010 (UTC)
May be down for a while due to server relocation: From Foundation mailing-list:
We are currently relocating some servers internally in the datacenter.
As a consequence, search snippets, "did you mean..." and interwiki
search are going to be turned off during this time, and only bare
results shown. This will affect all WMF wikis. I expect, if everything
goes well, that in around 4-5h things are going to go back to normal.
Does this server relocation explain why the apfilterlanglinks parameter doesn't work in API enquiries? I am currently experimenting with the API from a browser session (to try and understand it a bit better) but whenever I submit an enquiry that uses that particular parameter the session hangs and eventually gets a Wikimedia Foundation error page. Boissière (talk) 21:44, 13 August 2010 (UTC)
Question about writing templates with handheld devices in mind
I think I read something about this on this page but I can't find it. I working on an update for {{Infobox mountain}}. I want to expand the minimum table width so that coordinates don't cause wrapping. A user is concerned that, if the table is a bit wider than absolutely necessary, users of hand held devices might have problems.
Hi all - I've had to leave a template unsubst'ed because subst doesn't seem to be workingith it! Have a look at this edit. Is there any reason why the template shouldn't work properly in the references at the bottom? Grutness...wha?09:18, 13 August 2010 (UTC)
There's an open MediaWiki bug for the fact that it's not possible to subst within a ref tag (I don't know the issue number). Rjwilmsi11:14, 13 August 2010 (UTC)
(That message plus a deadline of 25/8/10 has just started flashing at the top of the watchlist for monobook users.) Like the bug that makes it impossible to click on "help us find bugs" or hide it? Speaking of which ... does anyone know if there's a gadget yet to move the search box in the new skin to the middle left, i.e. where it is in monobook? As I and others have said, Vector may be nice but it's a non-starter for those of us who are typing in the search box at the same time that we're going down our watch list; you can't do both at once (unless you have lizard eyes) if they're on opposite sides of the page. - Dank (push to talk) 15:48, 13 August 2010 (UTC)
This is really poor now. The info after the edit text appears overlayed in the editable text. Not sure if it's Safari or a change on Wikipedia but wants testing and sorting out. Unfortunately I'm busy elsewhere and to return to Wikipedia to find that editing is near impossible means I am less likely to stop by. Regards, SunCreator(talk)11:11, 10 August 2010 (UTC)
I noticed the same problem a couple of days ago and was just about to make a screenshot to post, but I am no longer able to reproduce the issue. Something apparently got fixed somewhere... —DoRD (talk) 13:12, 10 August 2010 (UTC)
I spoke too soon - the issue is only apparent after you start to type in the edit box. portraitlandscape I'm thinking that this is an iOS or mobile Safari issue. Once you type inside the edit box, the screen on the device zooms to the cursor and it seems like the edit box zooms at a different rate than the rest of the page. Rotating the device from landscape to portrait or vice versa sometimes clears the issue, so I don't think is our site. —DoRD (talk) 13:26, 11 August 2010 (UTC)
Yikes, looks like a bug in iOS Safari. Can you find the exact iOS or Safari version for me? I'll try to file a report at apple on this one. —TheDJ (talk • contribs) 14:44, 11 August 2010 (UTC)
I'll try to file something as well, if I can figure out where to do so. The results above are from an iPhone 4 running iOS 4.0.1 (8A306) and whatever version of Mobile Safari that includes. —DoRD (talk) 16:28, 11 August 2010 (UTC)
I proposed that we use some sort of dynamic sizing to display TFP given the user's monitor resolution. This would display (landscape) images in the most appropriate size for the given screen real estate, rather than having them drown in acres of whitespace as is currently the case. (Portraits are left out of this discussion because they'd become much too tall, also, they are usually left-aligned with text alongside whereas landscape are top centre-aligned with text below.)
Mediawiki software does not support a percentage-based size option (at least, not on Wikipedia) so other technical solutions may need to be considered. Any inputs? How could this be implemented? Zunaid21:37, 10 August 2010 (UTC)
Actually, only panoramas are centered with the text below. Standard 4:3-ish landscape orientation images are still left-aligned. howcheng {chat}00:33, 11 August 2010 (UTC)
I've got a solution which does this in CSS (no JS). To test it:
Edit your user css file and copy the section of definitions labelled "Finlay McWalter's hacky code for rescaling image block." from the bottom of User:Finlay_McWalter/standard.css into it
This works fine if added to MediaWiki:Common.css (at least on my personal test wiki). Right now there are several issues that should be addressed before using this on any Wikipedia page, never mind high profile ones:
It's a hack, and I've not tested it adequately, but you get the idea (works at least on Firefox for me). I doubt things will be plain sailing on IE.
Every visitor downloads the image at the size we specify; specify a larger size and you'll cost WikiMedia a bunch of extra bandwidth
It looks sucky on a big screen, where the scaled image is larger than the original we've specified. This can be addressed using the css max-width and max-height properties (which you'd set to those of the underlying thumb) but support for those properties is patchy
I've done the work for a thumb - it should really be a plain undecorated image, but it's late and I'm sleepy - the CSS will need to be tweaked for whatever DIV classes mediawiki inserts for a plain inline image
I've probably overkilled it with the !important, and it may not be necessary to specify all these values for the intervening DIVs.
Oh, and remember to purge your cache of the .css file, as you see the images giant if you don't. That's an issue for any such change when it's visible to everyone, and the main page a thousandfold so. -- Finlay McWalter • Talk02:48, 11 August 2010 (UTC)
That works brilliantly. Is it possible to code it into just the TFP template? TFP's are normally very high-res images so should never be resized upwards. As for bandwidth, at most I would guess a 1700x500 thumbnail is the largest landscape/panorama we would need on today's widescreen monitors, which isn't THAT much. Only concern is the aliasing that occurs when resizing (the falcon's wing against the sky for example), which means the majority of our viewers will see an aliased version of the TFP. Is there any way around this? Zunaid07:36, 11 August 2010 (UTC)
It shouldn't be difficult to code it for that template. While 1700x500 isn't huge, it's a cost that we pay for every visitor to the main page (which I'm going to guess is quite a large number). If you can give me an example of such an image, I'll do some calculations. Aliasing: yes indeed. The quality of the scaling depends on code in the browser, and most browsers do a rough-and-ready job. -- Finlay McWalter • Talk07:54, 11 August 2010 (UTC)
Oh, and people with small screens still download the large thumbnail, which is obviously annoying. Add another concern: I don't know at all how this will behave for folks with mobile browsers like Mobile Opera and iPhone/Safari. -- Finlay McWalter • Talk07:57, 11 August 2010 (UTC)
I wouldn't personally worry about the server cost. If it really makes problems the server admins will intervene. If you feel it necessary run it by the server admins but I'm guessing they'll just point you to WP:Don't worry about performance. I would be far more concerned about the impact on viewers. A 1700x500 image is probably going to be several hundred k at least. This could make problems for those with slow connections. If this were solely for those who need it then it may be okay since there is probably some weak correlation between monitor size and internet connection and those still stuck with modems and the like probably aren't using very large monitors. However if this is going to be needed for all visitors even for those who don't need such a large thumbnail then it sounds like a bad idea. In fact even those using mobile phones may not be happy. Sorta like the problem we used to have with animated gifs.
Of course if it's implemented so the image is never larger then what they will get now, then it's a moot point.
BTW, are you aware that at the current time in your sandbox, the central image is overlapping into the left menu bar, the left side image is not only overlapping the left menu bar but is actually off the screen and the right side image looks okay except the text is overlapping it (but not all the way to the right side). This is on Firefox 3.6.
Template:POTD/2010-08-09 is a recent example with acres of whitespace on a widescreen, and even just on a normal aspect ratio high-res screen. How about my original idea: create 2 or 3 different size thumbnails and automagically choose the most appropriate one for the user? If then required, we can still apply Finlay's code to this, more appropriate, thumbnail. This would save bandwidth for small screen users while ensuring quality for big screen users. Zunaid13:39, 11 August 2010 (UTC)
Yes, I much prefer doing it in javascript, where the image is replaced with a scaled version that is (width/50)*50 pixels wide. This is much better than the CSS solution:
we don't need to wait a month (!!) for everyone to see changes to common.css
for visitors without javascript, they get what they'd get now
the javascript can either set the width of the image to the exact size necessary (entailing suboptimal browser scaling) or we just live with it (the max whitespace on either side would be 25px, so 25/2px on average) - that's not an offensive amount of whitespace, even on a small screen.
my javascript isn't quite up to the job, but it's straightforward to write:
find a div with id=mainpage_fp_img
find the img child of it
calculate which thumbnail to use and change the src field of the img tag accordingly
we need to remember (bot?) to have a page somewhere that references all the different sized thumbnails (and actually visit that) so that mediawiki generates and caches the necessary thumbnails.
this way small screens don't have to download an unnecessarily large image
The WP sidebar kills the first 170px or so of width. Then the margins around the template and some whitespace for ease of reading take up a bit more. The picture doesn't have to be as wide as the text so that gives us even more leeway. As a rule of thumb I'd say we need 3 image sizes: 1500px (for 1920- down to 1440-width screens), 900px (1280- to 800-width) and 450px (for 640 and below incl mobile browsers), combined with some of Finlay's code which reduces (but does not enlarge) the image appropriately. How would this get done? Any other programmers in the house? Zunaid18:05, 11 August 2010 (UTC)
A script like that would then be added to every single page in the wiki. In general we don't have "page specific" scripts where we can avoid them, simply because it isn't efficient (we have scripts that work for large groups of pages). Just because we can do something, or because it is fun, is not a good basis for actually doing something, without considering the consequences. We have a couple of 100 Javascript extensions, yet only very few are actually part of the website for all the users. This is because more javascript usually equals a slower and more complicated website. Perhaps it is a better idea if someone developed a "lightbox" like extension for Wikipedia. Though then there are issues with linking to the description page that need to be handled of course. Anyways, it is something we could test out if someone contributes the code, but just remember that every feature also has unwanted side effects usually. —TheDJ (talk • contribs) 19:36, 11 August 2010 (UTC)
No, you can't have javascript in templates any more than you can have it in regular wikitext. Javascript can only be in the site-wide .js, which is loaded for every page. You can made sure it only works for given pages, but it's loaded for the lot (and as TheDJ notes, you have to worry hard about what happens if it somehow does end up trying to work on a page you hadn't intended). I believe there's a new feature coming that will have an image-description-page specific script (which doesn't help in this circumstance), but not for something like this. You can think of all kinds of ways MediaWiki could be extended to afford a degree of flexibility in this regard (it's entirely a MediaWiki issue), but of course with that comes the worry about doing so in a sane, usable, and secure fashion. -- Finlay McWalter • Talk18:10, 12 August 2010 (UTC)
I hope I can explain my question in a way it can be understood. I am wondering if there is a script or any means (or can a script be written) to help clear a message alert a little more easy. I have two en-Wiki accounts my main account here User:Mlpearc and a Public use account here User:Mlpearc Public. The talk page for the Public use account is redirected to my main account talk. If someone leaves a message at my Public use account, that account gets flagged with the "New Message" banner, of course. Now the posted message is at User talk:Mlpearc, so now that account has new input so it also has a "New Message". Now to clear the banners or new message alert, three pages need to be opened, 1. the talk page the message was originally assigned to User talk:Mlpearc Public 2. The main account where the message is actually posted at User talk:Mlpearc. You would think that after opening both talk pages the "New Message" banner would be canceled, but in order to clear that complete message alert I still have to open this page [5]. I also have this same problem at Simple English Wikipedia and at a project wiki. I hope this illustrates my situation and question. Mlpearcpowwow14:03, 12 August 2010 (UTC)
As far as I can tell, most users with public accounts such as yours simply point the public account's signature back to their main account's talk page. That way, messages should only be left in one location. —DoRD (talk) 14:37, 12 August 2010 (UTC)
Yes, thats exactly what I have done, and the message is left in only one location, but the message alert is flagged to both accounts and going to and opening the three pages is the only present way to clear the alert. Mlpearcpowwow15:19, 12 August 2010 (UTC)
(ec) The latest "public" sig I see is here. What I mean is to change it to read [[User:Mlpearc|Mlpearc Public]] ([[User talk:Mlpearc|talk]]), or some variation on that. But, if the actual redirect page (User talk:Mlpearc Public) isn't being edited, no message indicator should appear for that account. —DoRD (talk) 15:40, 12 August 2010 (UTC)
Stange, your alt's orange bar shouldn't get triggered unless its talk page is edited. I've never encountered the behaviour you describe, and I operate a fewalternateaccounts whose talk pages redirect to mine. –xenotalk15:40, 12 August 2010 (UTC)
Yes this is strange, this is not the first query I've made of this matter. The actual talk page with the redirect isn't being edited but once the message system on the public account (how ever triggered) is set then It takes, as I describe, opening of all three pages to cancel the message alert (to remove the "New Message" banner from both accounts). Mlpearcpowwow17:02, 12 August 2010 (UTC)
Or indeed Wikipedia, which has articles on a number of characters; ̯ redirects to Breve (which unfortunately doesn't say much about the inverted below form). --82.36.25.115 (talk) 02:08, 15 August 2010 (UTC)
Can you really not last three days without checking your edit count? This sort of obsession is precisely what X! is trying to illustrate. OrangeDog (τ • ε) 11:40, 14 August 2010 (UTC)
I found X!'s edit counter fun--in the "curious" sense not the "editcountitis" sense--and even useful at times. I liked how it broke down edits by pages and page types. Recently I've worked on some categories and was curious to see how large a percentage of my recent editing it had involved. I had also been making a lot more small, quick edits than I usually do and was curious to see how it compared to my usual editing. It was then that I discovered X!'s tool was disabled. I suppose there may be an obsession among some editors and X! is making a point that needs to be made. But not everyone is obsessed with edit counts. I know that edit counts are a poor way to judge much of anything. Different styles of editing produce different numbers of edits. Comparing your edit count to someone else's doesn't really tell you much, etc etc. I still found the tool useful and was sad to see it disabled. But if it did more harm than good then oh well, alright. It's too bad there are more editors obsessed with edit counts than editors just curious. Pfly (talk) 17:19, 14 August 2010 (UTC)
Search engine indexing of redirects
I was noticing today that if you google for "National Commission on Fiscal Responsibility and Reform" wikipedia gets into the first page results. But if you google for "Catfood Commission" which is a redirect to the same page, wikipedia is nowhere to be found. Not to say that wikipedia should be agressively SEOing, but I thought I'd throw that observation out so anyone might suggest with a brilliant idea better maybe than putting the "redirected from" in the TITLE tag, which sounds great but probably has undesireable side effects (?). While in my example, one is a nickname, there are likely some places where multiple important commonly used terms refer to the same subject and the ones that are redirects get low search engine billing as a result. —Preceding unsigned comment added by 76.118.216.33 (talk) 15:23, 14 August 2010 (UTC)
It should climb in the results if you give it more time; the redirect has existed for only a month. I've seen it happen plenty of times where I Google for something and Wikipedia is the first result, but what I searched for is nowhere to be found in the article, then I realize that it was due to a redirect to the page. The redirect you're talking about also has no internal links, and only one external link. It's currently on the sixth results page; it should be rising over time. Wikipedia has a special place in Google's heart; a lot of people want to arrive at Wikipedia when they search for something since it has most things that they need, and which is also why Google donates so much time, effort, and money to Wikipedia—because it improves the engine's search results by offering exactly what people are looking for. It's a massive win-win. Gary King(talk · scripts)17:24, 14 August 2010 (UTC)
Color Inconsistency Bug in Vector Skin
In the vector skin, the background of the search box in the upperhand right corner is white. The foreground color is from another source. This causes color inconsistency and in some cases illegibility.
In my case I am using a GTK browser (FireFox) and the foreground color appears to be coming from the GTK theme. To reproduce the bug, select a GTK theme with light foreground colors, such as "High Contrast Inverse". This example will give you white text on a white background.
I have a question, maybe even a puzzle. I am working with my top icons the markup I have in use is <div style="position:absolute; z-index:5; right:{{#if:{{{1|}}}|{{{1}}}|30}}px; top:{{#if:{{{2|}}}|{{{2}}}|8}}px;" class="metadata">
{| style="background-color:transparent;border: 0px"
|-
|{{click|link=Wikipedia:Account creator|image=Wikipedia_Accountcreators.png|width=26px|height=26px|title=This user is an Account Creator.}}. I like the display issues of this, It gives "Clickabilty" to the subject and "mouse over" also.
I've been trying to adapt to this markup <span style="position:absolute;top:-33px;left:797px;z-index:100">[[File:Wikipedia_Accountcreators.png|30px|This User is an Account Creator]]</span>. This markup allows me to place the image anywhere on the page and control of the image size, what it doesn't give is the "clickablity" to the subject and the mouse over is to the image file not the subject matter. Is there anyway to get both ? Mlpearcpowwow18:21, 14 August 2010 (UTC)
I've tried that and if your right, then I can't position the image as in the second example. the first example doesn't (or I don't know how) give me the image mobility as the second example.Mlpearcpowwow21:10, 14 August 2010 (UTC)
Wikipedia stalls while editing with Chrome on Windows 7
Hi, there's been an issue that's been bugging me for some time on this site. While I'm editing my browser routinely slows to a crawl. Just about the only thing I can do is close my tabs. Once I close all of the Wikipedia tabs (I can keep other websites open, so they are not the problem) my browser speed returns to normal. I am using Google Chrome 5.0 on 64 bit Windows 7. The same browser works just fine on Windows Vista and I can edit on this operating system using Firefox 3.6. My browser never stops like this on other websites. Is this an issue with Wikipedia, my browser, or my operating system? ThemFromSpace17:18, 13 August 2010 (UTC)
Press Shift+Esc to open Chrome's own task manager, or choose it from the developer tools in the page menu. Does the process with the Wikipedia tabs have an unusually high CPU usage number? It might be some runaway script, and you may be able to find out where it's happening with the Web Inspector (Ctrl+Shift+I, also in the dev tools menu), by switching to the Scripts view and pausing execution. --82.36.25.115 (talk) 01:26, 15 August 2010 (UTC)
Thanks for the help! I'll be sure to fiddle around with this next time I have access to that computer (I'm out of town and won't see it again until a week from today). ThemFromSpace12:39, 15 August 2010 (UTC)
I have a similar problem with Chrome on Windows 7, but it just ends up freezing my PC and I have to turn my PC off (manually, hit the button). I use Firefox for other things, usually with 80 to 110 tabs open, so I don't want to burden it with the 30 or so Wikipedia tabs I often have open. Of course, when it freezes I can't check anything. But I guess when it starts slowing down, that's the time to check. Dougweller (talk) 14:02, 15 August 2010 (UTC)
Indian rupee template
The Indian rupee sign was recently adopted, and a new template ({{INR}}) has been created to allow using the symbol articles (for example, {{INR}}250 crore displays as ₹250 crore, and a couple more examples are ₹123 and ₹5,000).
Problem: A web browser is likely to wrap lines between the symbol and the number (adjust the browser window width to experiment with how the above line wraps). Is there some reasonable way of handling this, perhaps by making the value a template parameter? Or, is the simple wikitext illustrated above better than something like {{INR|250 crore}}? Johnuniq (talk) 11:08, 14 August 2010 (UTC)
As long as it's an image I don't think there's any way to stop it breaking. A unicode character would have to be used, but I imagine most people won't have the required fonts. OrangeDog (τ • ε) 11:39, 14 August 2010 (UTC)
If you did use a syntax like {{INR|250 crore}}, it could easily enough include {{nowrap}} to product output like ₹250 crore. I don't know whether every browser *cough*IE*cough* will correctly not wrap that, but it might be worth a try. Anomie⚔13:24, 14 August 2010 (UTC)
Thanks, that works well. The INR template might end up being used a lot, so if we made that change, would it better to include the "span...nowrap" stuff directly in {{INR}}, or for simplicity, use {{nowrap}} inside {{INR}}? I know I'm not supposed to worry about server load, and it would be a very light load, but I'm inclined to think that since the "span...nowrap" is so simple that it might be better to include it directly in INR. Johnuniq (talk) 03:32, 15 August 2010 (UTC)
(For sure, do not introduce 'server load'. If a problem at all, they will come ask you. This guidance solved me a lot of mental problems. We, editors, do our thing, and they should produce our stuff. It works. -DePiep (talk) 22:14, 15 August 2010 (UTC))
Though users won't be abused, editors will. Is there a way to prevent trolls from "adding bullsh*t"&"revert immediately" and leave a mess in the article history? —Preceding unsigned comment added by SchwarzKatze (talk • contribs) 14:31, 15 August 2010 (UTC)
I certainly agree that edit histories can become cluttered by a series of edits for vandalism followed by their immediate reversion. I had previously thought to myself, and I believe I also posted this thought here, that it would be useful if the MediaWiki software allowed for users viewing revision histories to have the option to hide edits that have been undone or rolled back. Administrators do have the ability to delete revisions from the article history, but my understanding is that this is only used in extreme cases, rather than for standard vandalism. Rjwilmsi17:54, 15 August 2010 (UTC)
Cite shortcut when editing
When I use the cite web template shortcut from within page edit it often places the information away from where I want it (higher up the article), despite checking that the cursor is in the correct place when I press 'insert'. Also, is there any way to change the default fields that appear when the box comes up, and could it be this that is causing the positioning problem? Regards Eldumpo (talk) 17:59, 15 August 2010 (UTC)
Capitalisation discrepancies
I have just noticed in my Edit watchlist page that entries are followed by "(talk | History)" or, in the case of user pages, by "(talk | History | User contributions)". It is of little consequence, but shouldn't all the links in these parentheses be lower-case? I hadn't paid attention before, so I cannot tell when this changed (or, indeed, if this hasn't always been the case), yet this inconsistency drew my eye almost immediately today, so I'd wager on its being a recent development.
There looks to be very little consistency as to initial capitalisation at Special:AllMessages, not even between MediaWiki defaults and enwiki overrides, though the latter are more often lower case. But, for what it's worth:
MediaWiki:Contributions = User contributions (MediaWiki default "User contributions", no enwiki override; deleted on 28 November 2007 as "...identical to the default..." by User:Ais523)
Well, "User contributions" I can live with, but I am certainly very happy to see consistency between "talk" and "history". Thank you both for your prompt response. Waltham, The Duke of10:48, 16 August 2010 (UTC)
It's possible. Not sure if it is really useful though, because the class cannot be used without the template, and in bytesavings, handing the class out to all visiters probably won't outweigh much with the added byte count to a set of articles. —TheDJ (talk • contribs) 12:09, 14 August 2010 (UTC)
The template should be very widely deployed; often multiple times on one page. It's also intended that the style should be available to other templates, not necessarily by calling this one. Are the criteria for making such a decision documented, or arbitrary? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits18:29, 14 August 2010 (UTC)
I am trying to fix the cat pages for AGT/PRT and PM transport systems. There are two pages in particular that need to be fixed... automated guideway transit works as it should, but automated guideway transport is redired to people movers. I cannot figure out how to fix it, because there's some off templates that I simply can't figure out.
So, can someone redir transport to transit, and then add a see-also in transit pointing to people movers? I edited them as I would an article, but this did not work. Maury Markowitz (talk) 19:49, 16 August 2010 (UTC)
I've found problems in the rendering of articles containing Template:NZ population data in PDF files and on WikiReader. I've asked at Template talk:NZ population data if there is any modification which can be made to the template to solve this. I don't know if the problem occurs on many other templates as well. I know that location maps in PDFs are also broken. Is there a more appropriate place to discuss this?-gadfium22:23, 16 August 2010 (UTC)
Removing table cell borders in nested tables
Hello. I'm trying to do the following involving nested tables:
The outer table uses class "wikitable" as normal.
The inner table should have no borders whatsoever, not around the table, not around the cells.
I can't figure out how to do this cleanly. Basically, the "wikitable" class in common/shared.css specifies a one-pixel border for "th" and "td" cells. The "style" parameter on the inner table only specifies the border for the inner table as a whole, so if I say "border=0" I still get borders around the cells due to the "wikitable" specification. How can I override this without having to add a "border=0" spec on every single inner cell? Potentially this would work if there is an existing "borderless" class I can use on the inner table that overrides the "th" and "td" border value, but I can't find such a class; and I can't see how to insert my own CSS specs into a regular wikipedia article page.
It seems that the newer version of ImageMagick that we recently started using, is less efficient with memory usage, thus increasing the change that the thumbnail cannot be produced on large images. —TheDJ (talk • contribs) 15:25, 17 August 2010 (UTC)
Hei. I've been pulling hair for a while, trying to figure out the source of errors I get for a location map I created. I used Template:Location map Greenland Disko Bay as a model to follow, and created Template:Location map Greenland Upernavik Archipelago. Right there, Blofeld's usage of {{#switch:{{{1}}} didn't seem to work for the template I wrote. The template showed the map only if I substituted that for {{#switch:{{{Location-map-upernavik-archipelago.png}}}, using Location-map-upernavik-archipelago.png as value instead of *. I'm guessing I was on the wrong path there, but I have no idea why the latter didn't work. Now, I get expression errors (missing operands), when I try to use {{Location map|Greenland Upernavik Archipelago|lat_dir=N|etc..}} on a page, as I would use {{Location map|Greenland|lat_dir=N|etc..}}. Can you help? — Algkalv (talk) 15:03, 17 August 2010 (UTC)
I do not understand why you used the name of an image for {{{1}}}; I changed it back. Initially you were on the right track, but I presume you forgot to purge the page after you saved, to allow {{Location map/Info}} to update. It is also possible you input too many decimal places for the coordinates (I'm not sure as I haven't read through the entire template {{Location map/Info}}), which I have trimmed just in case, and for consistency with your model. In case you did not understand how location map works, permit me to explain (if you did, bear with me). The main template {{Location map}} find the template which you specify through the location in the first parameter and extracts the various values from it, through the {{#switch:{{{1}}}}}. If you change {{{1}}} to something else, the main template cannot extract the values and fails. Intelligentsium15:32, 17 August 2010 (UTC)
I concur. The precision in the coordinates most likely wasn't a problem. Anything reasonable should work with #expr, which is what the main location map template uses for the mathematics in determining where to place the pins. Plastikspork―Œ(talk)16:04, 17 August 2010 (UTC)
Thank you for replies. The template is working as expected now. However, I'd appreciate a hint as to how to 'purge the page'. I assume this doesn't refer to my own cache, which I've emptied after I adjusted the precision (the marker was off otherwise). The source is adjusted, but the template shows old values. Am I missing something obvious? — Algkalv (talk) 16:16, 17 August 2010 (UTC)
Apparently working again NOW. Just wondering what was the problem? It was not a browser issue, I tried with both Mozilla FireFox and Internet Explorer, same issue. /HeyMid (contributions) 22:21, 17 August 2010 (UTC)
The oldest logs (blocks and user rights logs) don't have any block time nor any user groups specified. It only says "§2 §3" (§2 is the number, §3 is the attribute like hours, days, weeks, months, etc), although the summaries for the blocks etc are still there. Why is it like that?
That's because the special pages for logs were new and buggy a the time. They use dollar signs ("$") rather than the "§" sign, because "$1", "$2", "$3" etc are placeholders in the MediaWiki messages. See the output at MediaWiki:Blockedtext for an example. Before the special page logs were introduced, the logs were in the Wikipedia namespace; these old logs are linked from Help:Log#Other sources of logged information.
BTW, for me (using Firefox) it's also showing in red "Cite error: <ref> tags exist, but no <references/> tag was found" at the bottom of page. Trafford09 (talk) 14:05, 18 August 2010 (UTC)
Hi, I can't fix this template: http://en.wiki.x.io/wiki/Template:History_of_Turkey_sidebar
There are many small states in history, so I want a option to hide them, that's why I added "collapsible" to this template but I could not fix itI want "under the Mongols" phrase to go to same line as "1243–1335". Now, it's at the center of left column now, but I want "under the Mongols" and "1243–1335" go to the same line and Anatolian beyliks table to go to just below this line.
Can you fix it? Kavas (talk) 16:32, 18 August 2010 (UTC)
Under "under the Mongols" header, left of Anatolian beyliks box is empty. Can we get rid of this huge empty area? Kavas (talk) 16:47, 18 August 2010 (UTC)
The only way to get rid of empty areas is to fill them with some content. What do you want to put into that area?—EmilJ.17:05, 18 August 2010 (UTC)
I see. I have added a colspan="2" which moved the box to the left and reduced the overall width, but I'm not sure whether it looks like the way you want it to look. My further attempts to somehow move the inner table failed, so if this is not satisfactory, you'll need someone with better table-fu.—EmilJ.18:23, 18 August 2010 (UTC)
I'm trying to make a template that you put coordinates into, and it produces an image. Clicking the image takes you to geohack with the coordinates entered.
The template works perfectly in my userspace, but when I transfer it to article space, the link parameter shows up as a mouseover caption.
Note that I am not looking for input on the pros and cons of linking only an image. This has been discussed to great lengths by the projects involved, and the decision is very much in favour of implementing it. The flipside was no coordinates at all on any road articles in North America. - ʄɭoʏɗiaɲτ¢17:32, 18 August 2010 (UTC)
Firefox 3.5.11 on OSX Leapord. This is what shows up for me on Ontario Highway 77 -->
I misread, I thought you were talking about the road signs (not the globe). For me, it doesn't even generate the link... It offers me a popup preview of the image. –xenotalk17:51, 18 August 2010 (UTC)
I'm guessing you use the preview plugin, and its showing you a preview of what the image links to (in this case, the image). - ʄɭoʏɗiaɲτ¢18:10, 18 August 2010 (UTC)
It has nothing to do with the namespace, and everything to do with spaces in the generated link url. Encoding the page name seems to fix it. Since geohack accepts MediaWiki's space-to-underscore encoding, we can use {{FULLPAGENAMEE}} instead of having to do {{urlencode:{{FULLPAGENAME}}}}. Anomie⚔19:25, 18 August 2010 (UTC)
They changed "FULLPAGENAME" to "FULLPAGENAMEE" (the final "E" stands for "encoded") and the problem manifested in the mainspace page (which had spaces), but not your sandbox (which did not). –xenotalk19:53, 18 August 2010 (UTC)
Dates on watchlists
The perennial proposal on VPP maybe me remember a question I tend to have on Monday mornings: Is there a way to get advanced watchlists not to group by date? I tend to find the date grouping relatively useless when trying to determine what happened to specific articles over several days; I'd rather the option to remove the grouping them together, moving the date headings to each article's explosion or including a datestamp with the timestamp. —Ost (talk) 20:48, 16 August 2010 (UTC)
I would like that too, when I check my watchlist after midnight it does not make sense that the changes before and after midnight are separated. The only workaround seems to be to change my time zone setting such that the last 24 hours are in one day.--Patrick (talk) 21:05, 16 August 2010 (UTC)
That seems hardly possible, unless not only you watch a huge number of pages, but also your setting of "Maximum number of changes to show in watchlist" is very low.--Patrick (talk) 10:41, 17 August 2010 (UTC)
Hm, I just realized that I can now set that to up to 1000 changes. Didn't it used to be 500? In any case, with 18,314 on my watchlist, that still only takes me back to 05:05 AM =) –xenotalk13:21, 19 August 2010 (UTC)
I understand that this feature wouldn't be for everyone, I was more wondering how easy it would be to have such a feature as perhaps a script. I generally am not online around midnight and I can be conservative about watching pages, so it would help for seeing the history of edits from when I was last online. —Ost (talk) 13:15, 19 August 2010 (UTC)
Someone broke template:convert
{{convert|0|F}} and {{convert|0|°F}} are currently generating bad output.:
0 °F (−18 °C) and 0 °F (−18 °C) are currently generating bad output.
The template is a warren of subpages, so if anyone knows who's likely to have edited one recently... Wnt (talk) 17:40, 18 August 2010 (UTC)
For some articles, I will click the "show" button and it will send me to a link such as http://en.wiki.x.io/wiki/Cognition# instead of expanding the template. I use Firefox. I'm not sure what is causing this. Any help or guidance is appreciated. NonvocalScream (talk) 21:11, 18 August 2010 (UTC)
Could yo give a wikilink (using like [[Example]]) to one article where it fails? i.e. after before the bad outcome you described above, where did you click? Thank you. -DePiep (talk) 21:48, 18 August 2010 (UTC) (before = more clear) -DePiep (talk) 21:49, 18 August 2010 (UTC)
Do you mean you click the second "show" link at Cognition#External links? The url on the link says http://en.wiki.x.io/wiki/Cognition# but I think that's normal. Other show/hide links on this and other articles also add # to the url. Clicking the link works for me in Firefox 3.6 on Windows Vista, and other tested browsers. I'm not a browser expert but you could try to clear the entire, look at Firefox settings or reinstall Firefox. PrimeHunter (talk) 21:56, 18 August 2010 (UTC)
Well, I don't do urls. imo, your first line to me read like I click on "show" in a template [to me: the show/hide option, template in an article, all clear to me, DePiep], I get a out-of-english-wiki link. That is what I can check & react to.
But why do you introduce urls here? Anyway, that's out of my league.
Without Javascript enabled, the [show] link wouldn't even appear (as it is added by Javascript in the first place). It sounds like NonvocalScream is getting a javascript error in the function that performs the expansion; check your error console? Anomie⚔01:48, 19 August 2010 (UTC)
Removing the monobook scripts have resolved the issue. There was so much code, I did not debug it, I just removed the entire monobook. Thanks for the help! NonvocalScream (talk) 13:39, 19 August 2010 (UTC)
It would be elegant to import this gadget, especially for those who have never installed any additional fonts. It just translates the interwiki links into English, you can test it by ticking it here. JackPotte (talk) 08:15, 19 August 2010 (UTC)
JackPotte (talk) 05:13, 17 September 2010 (UTC)
API apfilterlanglinks property timing out
I mentioned this in a previous thread last week about a database move/restructure thinking that it was related to that but as the move seems to have completed and I am still getting the problem I thought that I would raise it again.
Basically I am trying out the API using a browser and I find that if I submit a request that uses the apfilterlanglinks option (sample request here) then after 4 minutes I get a WMF error page saying "Our servers are currently experiencing a technical problem...".
The details that the error page asks one to report are
Request: GET http://en.wiki.x.io/w/api.php?action=query&list=allpages&apfrom=Ma&aplimit=30&apfilterlanglinks=withlanglinks, from <redacted> via amssq36.esams.wikimedia.org (squid/2.7.STABLE7) to 91.198.174.33 (91.198.174.33)
Error: ERR_READ_TIMEOUT, errno [No Error] at Thu, 19 Aug 2010 11:20:12 GMT
I have tried a similar request against the fr and de wikis and that seems to succeed (though the response time is several seconds). Boissière (talk) 11:41, 19 August 2010 (UTC)
I am working on a new article in a sandbox. I have two pictures in it, one which seems to have difficulity displaying all the time. Checked multiple different PCs and same results. Same problem when I try to display the same picture here - especially when I refresh webpage. It just draws a blank white image and not the picture itself, most of the time. I suspect something wrong with the picture itself as it happens on several PCs. Can someone fix as I would like to use this picture. I am using Vista.
Note that Commons correctly displays "full resolution", but not the preview. Maybe this should be taken up there also. Wnt (talk) 12:39, 19 August 2010 (UTC)
The problem is now solved, there were lots of stuck imagemagick convert jobs on the server causing this problem. Apparently something with animated gifs again, this time do to a broken type of error reporting or something. —TheDJ (talk • contribs) 14:53, 19 August 2010 (UTC)
Is it possible to download all the versions of an article?
Hello all,
I have been checking the Wikipedia dumps and I wonder if there is any way to download just all the revisions/versions/history of a single article. Is there any tool to do this? Is there any ready database I could use?
On looking at things further, there was something weird about the colour handling in that image. In Firefox it showed as horribly gharish, and it colour information was badly mangled if viewed in ImageMagick. Clearly the graphics lib used by mediawiki (I don't know if this install is configured for GD or Imagemagick) had similar problems with the image (which, in fairness, rendered fine in other programs like Eye of Gnome). I've loaded and re-saved in GIMP, and uploaded that over the original, and it appears to work okay now. -- Finlay McWalter • Talk15:09, 16 August 2010 (UTC)
The colour problem was a result of the original version using the CMYK colour space, rather than sRGB. CMYK is intended for printing applications, not viewing via a web browser. sRGB is the standard for web applications. – ukexpat (talk) 18:21, 16 August 2010 (UTC)
Currently, when I click on a redlinked file (not a file link with a colon in front, but something like File:Foobarbaz1234567890.jpg), it takes me to the upload page. This is annoying when I want to see who deleted a certain file, etc. Is there anything I can add to my skin's css page so such links will instead go to the file page itself? —fetch·comms22:46, 19 August 2010 (UTC)
Dumping a wiki with linked documents to an offline version.
Hello! Greetings. I'm a Peace Corps volunteer in Cameroon, W. Africa. I'm a member of the steering committee for the agroforestry program, and we've been working tirelessly to develop a wiki to collect and preserve agroforestry information, both general information and information specific to Cameroon. You can view the wiki here:
While the wiki is great for anyone that has access to the internet, some volunteers are upwards of twenty hours away from an internet access point, and even then there's little hope of downloading an 11mb file over a dialup connection in West Africa. For this reason, we want to make an offline version of the wiki to distribute to volunteers upon arrival here in Cameroon. I've looked into various offline wiki viewers, but I don't think that any of them are satisfactory for the following reason: they don't download the files that are uploaded to the wiki, nor will they download files that are linked off-site. Most all of the information that we have is in PDF format, either that we've uploaded, or that is on someone like the World Agroforestry Center's website. The wiki is just our way of organizing the links to the files.
My big question: Is there any sort of software that will allow me to dump an entire wiki onto an offline format, including all the files that are linked-to? If it will work if the files are uploaded to the wiki site, but not for external links, then this is no problem; I can work to get all the files uploaded locally.
Yeah I read both those pages before posting, but I'm not sure that they address my problem... From what I gathered, it is possible to dump the uploaded files on a wiki to a local source, but perhaps Wikipedia stopped doing that just because of practical concerns. But I'm still at a loss as to how accomplish this exactly... It seemed that there are scripts to do this for image files, but maybe not other types of uploaded files. Is it possible to modify these scripts to do the task I want? I have no programming experience, and I also only have internet access on borrowed computers in the capital. Egad I'm way over my head as regards these things. Thanks for any help you can offer.
Mediawiki turns [[google:"foo%20bar"]] into a search for "foo_bar", (google:"foo bar"). Nice, except, the underscore messes up the results. Is it possible that the devs can make it use %20 in the actual link, instead? If so, I'll file a bug (or someone else who wants to). —fetch·comms00:54, 21 August 2010 (UTC)
MediaWiki does this because google is at meta:Interwiki map although google isn't a wiki. Some of the other interwiki links will break with %20 instead of underscore. It seems messy if different interwiki links should treat the argument differently. You can use + for Google: google:"foo+bar". There is also {{Google}} which allows spaces: "foo bar". PrimeHunter (talk) 01:24, 21 August 2010 (UTC)
Hello, I'd like to bring up a very interesting issue that I have been having when it comes to editing Wikipedia. The problem is that almost every time I click on the edit button to edit an article while using Google Chrome v6.0.472.36, something odd happens: The page appears to fully load and appears for a split second (literally), and then it goes completely blank right after. There's nothing there; the page is completely white. The browser still functions, but the actual webpage viewer shows nothing. To give you an idea of how often this happens, I opened up ten pages to edit, and only two of them loaded without blanking out on me. The problem also occurs when I preview a page (I can, however, save a page just fine). I did some experimenting to see whether or not this event is confined to Wikipedia; The results show that this is happening to me only on the English Wikipedia. I tried several foreign-language Wikipedias and several other Wikimedia and non-Wikimedia wikis, and they all work just fine when it comes to editing, as opposed to here. What I find to be the most interesting, though, is that this only happens when I am logged into my account. When I'm logged out, I can edit any page just fine without it blanking out on me. Again, I am using Google Chrome v6.0.472.36 (and yes, the problem only occurs when I use Chrome; I can edit just fine in Firefox). I was wondering if anyone else who uses Chrome has had the same problem, or if anyone possibly knows what's going on? As you can imagine, this is very annoying :) Thanks, ~SuperHamsterTalkContribs05:14, 19 August 2010 (UTC)
Sounds like a document.write() in a script. Try disabling any gadgets you have enabled, one by one. See if you can pinpoint the problem. It could also be an error with code that only loads for logged in users, but I don't remember any changes in that recently, so that doesn't seem very logical. —TheDJ (talk • contribs) 12:00, 19 August 2010 (UTC)
I started out with doing what DJ said to do, and that actually solved the problem - luckily enough for me, the problem lied within the first gadget that I unchecked. It was "Add a sidebar menu of user-defined regex tools, with a dynamic form for instant one-use regex." Funny enough, I never use that tool, so I don't know why I even checked it (unless it's a default option). Anyway, a thank you to both of you. ~SuperHamsterTalkContribs17:09, 19 August 2010 (UTC)
I've also noticed the same blanking when using Chrome, though it seems to have stopped without making any changes in my preferences. older ≠ wiser02:15, 22 August 2010 (UTC)
Recent changes for unwatched pages.
On Wikiquote, I set up a dummy account to watch several thousand unwatched placeholder year pages. I listed all of them on one page, so I can just go that page and click 'related changes' and see if any changes have been made to any of those pages. I am wondering now whether we could get a button that lets admins see what recent changes have been made to all unwatched pages, and perhaps also to page that are only watched by editors who have not logged in for a certain length of time. For example, if there is a page that is on some watchlists, but none of those editors have logged in for the past three months, then the page is effectively unwatched. I would like to be able to get a quick sense of what is going on with those pages, since vandalism to pages like that would otherwise be hard to uncover. bd2412T20:09, 20 August 2010 (UTC)
With the Wikiquote pages, that's not a problem because they can be checked frequently. There are few enough of them that the recent changes list is almost always blank, and the recent changes pages is quiet enough that a change to a placeholder year page would stand out like a sore thumb. I don't propose to do that with other pages there, or with any pages here. What I'd like instead is a link I could click much like the current unwatched pages link, but one that would show me recent changes to those pages instead of just the list of pages. Looking at a list of tens of thousands of unwatched pages does me no good. Being able to see if any of them have been changed is what is called for. I'd like a recent changes page, accessible only to admins or perhaps some larger set of trusted users, for these unwatched pages. bd2412T20:29, 20 August 2010 (UTC)
What is that doing on user-targeted help page? It's only useful to sysadmins; that's why it's in MediaWiki.org's Manual namespace. Would anyone be opposed to that part being removed? Reach Out to the Truth20:59, 21 August 2010 (UTC)
Note that the links to the MediaWiki manual are recent additions - I added them after I saw Gary King's comment above. Without the links, of course readers misunderstood what "LocalSettings.php" is supposed to be. I'm fine with removing any mention of that file from the onwiki documentation, though, which should bring the number of defects in the helpfiles down to infinity. — Gavia immer (talk)21:38, 21 August 2010 (UTC)
Variation between Monobook and Vector
Is there any way (e.g. using perimeters) so that I can vary the HTML markup on my userpage based on whether the viewer is browsing the page using Monobook or Vector? When I browse Wikipedia, I use Monobook (nostalgia reasons), and I would assume that most people now would use Vector. However, for a fixed image that I use to appear properly, I must use <span class="noprint" style="position:absolute;top:45px;left:-150px;z-index:4"> for Monobook, but <span class="noprint" style="position:absolute;top:15px;left:-150px;z-index:4"> for it to appear correct in Vector. Is there a way I could get around this? Thanks, -- 李博杰 | —Talkcontribsemail09:04, 21 August 2010 (UTC)
And this shows exactly why you just shouldn't do stuff like this. Wikipedia isn't MySpace. Other skins will be in use, and using absolutely positioned elements will make your userpage unreadable for some users and unusable for others (think some mobile devices). It's not forbidden, but it isn't wise either. —TheDJ (talk • contribs) 13:01, 21 August 2010 (UTC)
I don't think these features are enabled here, why are these rights included in the reviewer usergroup? –xenotalk14:52, 21 August 2010 (UTC)
There's also a few quirks with what permissions the bureaucrat right can grant. It can not grant IP block exemptions, but it can revoke them. It can also grant and revoke the account creator and reviewer rights. Technically these abilities are for administrators, not bureaucrats. Of course, in strictly practical terms, all of our bureaucrats are administrators so it's not an pressing issue. --Deskana(talk)20:24, 21 August 2010 (UTC)
Creating a Template
Let me explain what I want to accomplish. In the article United States District Court for the Central District of California, there is a section called current district judges. It uses a template that displays a table. There is also a section called current magistrate judges. It uses the same template. However, some of the columns are either inapplicable or unnecessary for magistrates. So, I created a template to display a table more suitable for magistrates. I thought it would be easy and created Template:Start U.S. magistrateship, using the district judge template as a guide. However, when I then try to edit the article itself and use my template, it doesn't even come close to working. I'm fairly confident I'm missing some key components to doing this properly. I looked at the help for templates but got dizzy and was hoping I could get an answer here that is specific to my problem.--Bbb23 (talk) 23:23, 21 August 2010 (UTC)
The template you created needs to have the same number of columns as the table that you intend to use it on, otherwise the columns won't match up. Your best bet is to just copy the existing template (as you have done), but don't remove any of the columns; instead, just rename the ones you want to change. Gary King(talk · scripts)00:19, 22 August 2010 (UTC)
The new end template will make more sense to editors but because it redirects to a standard template, its new name won't affect how the page displays.
Then, in the relevant section of the article, change the template names and, to make things tidy, delete the unused |index=, |senior term= and |appointer= details.
Outstanding. Even with your very clear help, it took me a while to make it work, but I believe it's all correct. Thank you very much.--Bbb23 (talk) 01:13, 22 August 2010 (UTC)
Search by suffix?
The following is a thread that began at the Help Desk. I have copied it here in order to continue the discussion further, particularly with regard to there being a bug or reason for a feature request. Osubuckeyeguy (talk) 18:06, 19 August 2010 (UTC)
I would like to find articles that share the same suffix, specifically (fighter). Pages would have this suffix if there exists other pages by people of the same name, and there is likely to be a corresponding disambiguation page. Ultimately, I am trying to figure out how many disambiguation pages link to articles about fighters. If I search article titles, it returns all pages with fighter in the title (e.g, fighter pilot), which is far more responses than is helpful. Searches seem not to preserve the parentheses, even though they are a key part of the search request. Is there a way to get the search to preserve meaningful punctuation? Or is there some other way to perform this search that I haven't thought of? Osubuckeyeguy (talk) 16:34, 19 August 2010 (UTC)
I think it would be useful if parentheses would be included if they are in quotation marks (e.g. if the search term was (test) then the brackets could be ignored perhaps, but if it was "(test)" then the brackets should be included! Perhaps you could bring this up at Wikipedia:Village pump or at Bugzilla (as the ignoring of the brackets in a quoted search would probably count as a bug)? -- PhantomSteve/talk|contribs\ 17:50, 19 August 2010 (UTC)
Thanks. This is exactly what I was looking for. Is it nutty to think that this should already be a feature incorporated into WP? Osubuckeyeguy (talk) 18:42, 19 August 2010 (UTC)
Well they said "pointless", not useless, but similar effect nonetheless. Why didn't you rejoin that it would be pointful for finding pages disambiguated with the same suffix? –xenotalk13:40, 21 August 2010 (UTC)
Text is not wrapping around images in Chrome. I'm not sure when this issue started, but I'm guessing it's within the last few weeks. Text wraps fine in Firefox (2.0.0.20) and IE (6.0.2900). I'm using Chrome version 5.0.375.125 on Windows XP SP2.
—Michael (talk) 16:37, 22 August 2010 (UTC)
This template is generating lines of whitespace. Could someone with some template knowledge (which I don't really have) take a look at this? See the Willie Sutton article for an example. Garion96(talk)16:53, 22 August 2010 (UTC)
Okay it should be better now. The problem began with "nationality" all the way down to "children". Note to self: New rows should be created outside of the if conditionals, not inside, otherwise it'll just create an empty line when the condition is not met. And apparently the conditionals have to begin their own line. Someone should consider converting that template to use {{Infobox}}. Gary King(talk · scripts)20:26, 22 August 2010 (UTC)
and now it won't render as a PNG. Visiting the URL that should return a PNG file ([13]) instead returns a "Stub Start Error". The only change I made from the second revision to the third is move label number 3 a little to the left to make things a bit clearer. Attempting to view the file itself works as expected. Did I do something wrong or did I find a bug in the wiki software? Frotz (talk) 20:48, 22 August 2010 (UTC)
"Stub start error" means that all image scaling is broken, it doesn't depend on which image you look at or what size you use. The fact that some images appeared to be working is just because they were cached. It's due to an intermittent problem on ms4, cause unknown but restarting SJSWS fixes it temporarily. For the moment, it doesn't trigger our monitoring, so if you see it, you should report it on #wikimedia-tech on irc.freenode.net. -- Tim Starling (talk) 04:24, 23 August 2010 (UTC)
What has happened here? Also, can somebody fix the problem involving {{admin dashboard}}, when userpages transcluding the dashboard show up in Category:Administrative backlog whenever one of the items in the dashboard reports a backlog?
The newpages tranclusion is a known bug, and AFAIK we just have to live with it or not transclude the page in certain circumstances. I'm too tired to fix the second part, as it requires (I think) going through the transcluded pages and stuff. —fetch·comms02:40, 23 August 2010 (UTC)
Userpage problem
My userpage is set up to display a United States flag behind the wikipedia logo. Until a few minutes ago, it was working properly, but now the flag seem to be about 80 pixels below the logo. I am using the Monobook skin on IE 8 and I have not made any CSS modifications to monobook.css. What is the reason for this? Access Deniedtalkcontribseditor review04:42, 23 August 2010 (UTC)
It looks like this is caused by the MediaWiki:Sitenotice added due to the #Images not rendering bug above. That pushing down the start of the main text and the flag is positioned relative to that. You can click hide next to the site notice to hide the message, but this will still have the image temporarily misplaced for other users. --Salix (talk): 05:23, 23 August 2010 (UTC)
Images not rendering
forgive me if I'm in the wrong place or this has been raised before, but I just tried to switch the image on Template:In the news and it wouldn't display. At all. Just a big square block of whitespace.[15] I'm pretty sure this isn't me, because I earlier responded to a {{helpme}} request with much the same problem. Can anyone offer any assistance? HJ Mitchell | Penny for your thoughts? 00:47, 23 August 2010 (UTC)
Apparently it's software called ImageMagick that is playing up. There are bug reports on Bugzilla, e.g.[16], but the tech people seem to have no idea what is causing the problem. Fences&Windows01:29, 23 August 2010 (UTC)
So that means no sitenotice for the pending changes poll... yay, shuts up that debate. The thumbnail issue, however, is really annoying; I don't know if it's possible to revert back to the old version of ImageMagick or what, as I can't tell what a certain image looks like without loading the full 10,000px-by-2,000px version, thus crashing Firefox on my system :P. Sighs. —fetch·comms02:34, 23 August 2010 (UTC)
I seem to be encountering problems with Geohack. On a few articles in which i have tried to find the locations of, instead Geohack is not giving a list of available wqebsites as per the norm but all i get is a rather strange version of the Wikipedia Main Page, with Geohack and the article title mentioned at the top. Simply south (talk) 21:56, 23 August 2010 (UTC)
Linking to subaddresses
Hi. How can I find out the subaddresses of this Personal website from John Peter Oleson, specifically the "HUMAYMA EXCAVATION PROJECT" and "SKERKI BANK DEEP WATER SHIPWRECK PROJECT"? I want to link to them directly as been done in footnote 2. Thanks in advance Gun Powder Ma (talk) 09:28, 23 August 2010 (UTC)
At the time of your question footnote 2 [17] linked to http://web.uvic.ca/~jpoleson/#romacons. Based on this I guess you wanted to link directly to a specific place on the page http://web.uvic.ca/~jpoleson. Often there will be an internal link from near the top of the page like in the table of contents in Wikipedia articles. But not in this case. Instead you can try to view the page source in your browser, search it for text at the place you want to link, and then look for code defining a name or id. In this case the code <a name="Humayma"> means you can link http://web.uvic.ca/~jpoleson#Humayma. The SKERKI section has no defined name and therefore doesn't permit a direct link. PrimeHunter (talk) 02:25, 24 August 2010 (UTC)
Yes, I wanted to do it like footnotes 1-3 in the current version, but I did not find a link to the SKERKI section, either. For the next time...how can I view the "page source" in my Firefox? Gun Powder Ma (talk) 14:12, 24 August 2010 (UTC)
I made a simple image at the Monge's theorem article. If memory serves, I used the application KSeg to make it. One thing that has always bothered me is that there's this border at the bottom of the image when its displayed on Wikipedia that shouldn't be there. In fact, most other applications do not show the border. I am not really proficient at SVG and my attempt to fix it was blind and failed. (There is also an error in the svg's id element that causes the original version not to validate that ought to be fixed.) My aborted attempt to fix it sort of shows what I was trying to do (the lines should go right up to the edge of the display area). If somebody could help me, I'd appreciate it. Jason Quinn (talk) 03:10, 24 August 2010 (UTC)
You had a clip path set on the group which didn't match the dimensions of the actual image; your other applications must have been ignoring the clip path completely. I just removed the clip path; while this makes it a bit less pretty in Inkscape, it should work fine for viewing in the browser. Anomie⚔11:18, 24 August 2010 (UTC)
With the way the "career_highlight" parameter in that infobox works, it was ending up with misnested HTML something like <div><ul><li>...</div></li></ul>, which confused tidy enough to screw up {{Fix bunching}}'s table. Adjusting the value of that parameter slightly so it ends up correctly nested as <div><ul><li>...</li></ul></div> instead fixes the problem. Anomie⚔11:46, 24 August 2010 (UTC)
The problem there is just that "Former NLL Teams" and "Former MLL Teams" are rather long bits of text. If you can manage to shorten those up somehow, that would do it, or you could wrap them in <span style="white-space:normal"></span> to allow those labels to wrap (or remove the white-space: nowrap from |labelstyle= to allow any of the labels to wrap). You could also try removing some of the extra padding added in |labelstyle= and |datastyle= to give more room for the text. Anomie⚔15:53, 24 August 2010 (UTC)
I guess my point with the two examples above from other sports is that the teams have a full return to the far left in other infoboxes instead of a return limited by the field name.--TonyTheTiger (T/C/BIO/WP:CHICAGO/WP:FOUR) 16:28, 24 August 2010 (UTC)
Current watchlist notice displaying after being told to go away
As you can see at my userpage, I'm in the process of moving from one US state to another. Yesterday, I dismissed the watchlist notice inviting me to attend the New York City meetup or whatever it is; this was before I began travelling. Having arrived at the new place yesterday, I've gotten the Internet connected, and today I discovered that the watchlist notice was back. Is this because I'm now on a different IP address, or is some other cause likely to be the reason? Nyttend (talk) 13:07, 25 August 2010 (UTC)
I've done some experimenting and it appears to have something to do with my account. I am using the vector skin, with some customisation to turn visited links purple placed at User:Whpq/vector.css, and which has been there for some time.
I checked the above AFD while logged in both Firefox and Chrome browsers. I cannot see any table of contents.
I logged myself out and checked the above AFD in both Firefox and Chrome browsers. The table of contents now are visible and formatted as expected.
I removed the single line of code in User:Whpq/vector.css and repeated the above checks following instructions to bypass teh cache for each browser. It made no difference in the results.
User:Whpq/vector.css is the only customisation I've done. The remainder of my preferences are controlled through "My preferences". What else should I be looking at? Prior to today, the AFD has been showing table of contents. -- Whpq (talk) 18:09, 25 August 2010 (UTC)
Update: After checking some regular articles, it appears that I am missing the table of contents on all articles. -- Whpq (talk) 18:41, 25 August 2010 (UTC)
Is it possible to have a reference within a reference? In the article Mama grizzly, I am trying to put a reference within a footnote. Since the footnote is a type of a referece, the software does not seem to allow me to put a reference inside the footnote. For now, I just repeated the content of the reference inside the footnote, but that's not very elegant. Victor Victoria (talk) 21:11, 25 August 2010 (UTC)
Thanks. However I need something slightly more sophisticated. What I need is {{#tag:ref group="n" name="XYZ"|ABC<ref>ABC</ref>}}, but that doesn't work. I need the group="n" because it's a footnote, rather than a reference, and I need the name=XYZ because this footnote appears many times in the article.Victor Victoria (talk) 21:29, 25 August 2010 (UTC)
There is a text written right into the software, used when deleting this image on the most recent occasion. I am unable to find this text to propose a change in it (it's currently stating the policy incorrectly). Can you help me find it, please? See also: User_talk:Airplaneman#File:Ruwan.jpg which is where I first brought this up. Magog the Ogre (talk) 23:21, 25 August 2010 (UTC)
I want a way to force users to actually think when contacting me, instead of using robo-poster tools that allow them to mass-mail. I can do this for my telephone, so why not here? I am open to suggestions on how to accomplish this. I am not entirely clueless technically. Maury Markowitz (talk) 17:18, 25 August 2010 (UTC)
Looks like standard template messages, {{Idw-noncom}} to warn you about a pending deletion of an image you uploaded as the image licensing prohibited commercial use and hence is not "free" according to wikipedia standards.
It is fairly standard practice to use the standard warning templates as they do provide all the necessary details which would be hard to do if written by hand. It seems that the template was correctly applied, as the copyright on the Canadian Forces website restricts non-commercial use. If you want to avoid such messages in the future the only solution seems to make sure you know the ins and out of WP:COPYRIGHT especially WP:FAIRUSE. --Salix (talk): 22:44, 25 August 2010 (UTC)
No Salix, I get all sorts of these all the time from users who are doing things incorrectly and littering my talk page. I'm tired of it, and I'd rather not know about them. So I don't want to get them. It's my talk page, so how do I block these? Maury Markowitz (talk) 11:49, 26 August 2010 (UTC)
I think there's an option to ignore warnings when you upload an image, but I'm not sure if it does exactly what you want. I don't think there's any way to block messages from people in the way you suggest. Chase me ladies, I'm the Cavalry (talk) 12:56, 26 August 2010 (UTC)
Using {{nobots}} would prevent most automated processes, but semi-automated systems like Twinkle would still cause messages to be left for you. You realize that if there were a way to opt-out of these notes, you wouldn't get notifications about items you've uploaded or created being nominated for deletion? –xenotalk13:28, 26 August 2010 (UTC)
Page notice/edit notice for IP users
Recently the embedded notice for IP edits was changed to include a vast amount of red text. While I understand the desire to grab the attention of such users, I find it rather overbearing and distracting. I was referred here from ANI in the hopes of finding the discussion or perhaps even convincing somebody to change this back. It's already causing some eye strain. <squint> 69.181.249.92 (talk) 06:20, 26 August 2010 (UTC)
I've removed some emphasis there; there is little point in emphasizing every second sentence with scary colors and bolding. Ucucha13:42, 26 August 2010 (UTC)
Table format
At List of regions by past GDP (PPP) per capita, I would like to add to the bottom of the table "Europe 1830–1938 (Bairoch)" the heading row with the years, so that people can also sort the table from its lower half.
And there is a problem with the sorting I don't quite yet understand: It takes each time 4 clicks to sort the table from highest to lowest value and vice versa. But on the table " World 1–2003 (Maddison)" above it only takes two. I suspect it may have something to do with the "–" in the Bairoch table which lacks in Maddison's, but I don't know how to fix the problem.
I don't think it's possible to have two rows where you can click to sort a table. It's possible to have a row fixed at the bottom if you don't want readers to have to look at the top to see the column heading, but they still have to click at the top if they want to sort. See Help:Sorting#Excluding rows from sorting.
Currently the sort mode at List of regions by past GDP (PPP) per capita#Europe 1830–1938 (Bairoch) changes between numeric and alphabetical because it becomes alphabetical when an en dash '–' is in the top row. You can make it numeric by making a blank cell instead or replacing the en dash with a hyphen '-'. This is considered a minus when sorting and is a valid character in a number. I first thought you would need something like -<span style="display:none">0</span> to make it -0 without displaying the 0, but apparently '-' works alone. See Help:Sorting#Sort modes. PrimeHunter (talk) 11:06, 26 August 2010 (UTC)
Poor numeric sorting is a long standing bug. What we need to a way for force the numeric sorting of table columns. Interested editors may wish to vote! — Blue-Haired Lawyert11:50, 26 August 2010 (UTC)
The rows below are example from the table here. My question: how can I arrange the cell-texts vertically into "vertical top" (aligning upper characters), from the current "vertical center"? Tried "vertical-align:top" - didn't work.
Emil J removed a space after class="IPA wikitable" in your post and it changed the table formatting. I have restored the space in your post. PrimeHunter (talk) 11:46, 26 August 2010 (UTC)
You did give the solution. I only tried the setting in the top-row (table definition), which failed. btw, already used here. -DePiep (talk) 12:55, 26 August 2010 (UTC)
You can apply the vertical-align to the row, as well as the text-align. Yay CSS (> deprecated HTML attributes)!
(edit conflict)So now it's only a copypaste job. I'll surely try the CSS-by-row first (on "center" too), as suggested. Thanx both three of you (The missing space also prevented triple-rows, evading the problem, but now it's working showingly). -DePiep (talk) 11:56, 26 August 2010 (UTC)
Image links oddness
File:Cinnabar on Dolomite.jpg is used in a number of locations, but only three of them are listed on the image description page, and those three I just added today (so the database rows were recently inserted). Here are some other locations where it's at:
...and I'm sure there are more, such as a Signpost article. Does anyone know if this situation is unique to this image or is it happening to others as well? Thanks. howcheng {chat}16:36, 26 August 2010 (UTC)
Duh! Considering that the FPC page name has it misspelled (which I noticed) I don't know why I didn't think of checking the misspelled filename. Thanks. howcheng {chat}22:25, 26 August 2010 (UTC)
UserAgent pop-up window
I started getting a pop-up window on every click in Wikipedia, and nowhere else. The window heading states:
The page at http://en.wiki.x.io/says:
Then the actual window text is 7 lines long. The first line says:
userAgent:Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US}
I got these as well, briefly. They seem to have disappeared. Perhaps they were added by someone who wished to do some debugging and mistakenly submitted them to the production site? Gary King(talk · scripts)07:22, 26 August 2010 (UTC)
wikEd_dev.js is for developing and debugging, non-developers should not call it and use the wikEd gadget instead as described on the wikEd home page. But for now I have copied the current wikEd version to wikEd_dev.js. Sorry, Cacycle (talk) 07:01, 27 August 2010 (UTC)
I've switched back to the regular wikEd. I was using the beta version for some feature that wasn't in regular wikEd, but I assume that it's probably int here by now. If not, then I'll notice and just switch back. Gary King(talk · scripts)15:31, 27 August 2010 (UTC)
As far as I understand, all wikimedia requests go trough the squid servers. Only there it is determined if something is cached or not. And some API requests are cached (not many though). —TheDJ (talk • contribs) 12:47, 27 August 2010 (UTC)
Possible bug; talk page didn't update
I posted here and the talk page came up on the screen as if I was seeing it for the first time. I could have checked the history, I guess. But I reposted and the topic I had just posted was there, so I deleted the duplicate topic.
If we know a location, is there a tool for longitude and latitude?
There may be something going on that I don't know about. Radio stations have radio-locator.com and mountains have a database of some kind. This time I want to add coordinates for a college. I've looked and looked using Google maps and such, and while I've found an arrow pointing to the school on a map, I haven't found coordinates except for a satellite campus.Vchimpanzee· talk·contributions·16:50, 25 August 2010 (UTC)
On Google Maps, if you right click on the spot you want coordinates for and click What's here?, it'll put the coordinates in the search bar above the map. Killiondude (talk) 16:54, 25 August 2010 (UTC)
Hmm, interesting; I never knew that. You can also use this site, which uses Google's maps; type the address into the search bar, and it takes you to the site you want. Unlike on the main Google maps site, you can drag the little icon wherever you want to center it, or you can drag the screen around to get the little icon to recenter. I find it more convenient than any other service I've found. Nyttend (talk) 17:04, 25 August 2010 (UTC)
Well, this is what I got. The physical address of the school is 2050 Hwy. 501 East, Conway, SC. When I type that in I get something on the other end of town. If I enter 108 James P Blanton Cir, I get the area map I want. Horry-Georgetown Technical College is next to the Coastal Carolina campus. Any ideas on what to do next? Vchimpanzee· talk·contributions·13:51, 26 August 2010 (UTC)
Could it be that Hway-nr and housenr got mixed up? (does it point to house 501 somehow?). I tried "Church Street East, Conway 2050, South Carolina", and I get numbers close to 2050. I tried ACME btw. -DePiep (talk) 18:41, 26 August 2010 (UTC)
Should be even better "Church Street East 2050, Conway, South Carolina" ... Oh these American numberings. -DePiep (talk) 18:43, 26 August 2010 (UTC)
I'm not sure what the policy question might be. Finding the physical address is a simply a matter of finding the right reference source, such as these. Translating the location on a reference map to lat and long is a basic mapreading skill that doesn't come under wp:OR. Several templates such as {{coord}} can be used to place the location in the article and provide links to various online maps. Are you perhaps concerned about where on the campus the map should center? I'd just pick the approximate middle of the campus. LeadSongDogcome howl!21:23, 27 August 2010 (UTC)
I've found the physical address. 2050 Hwy. 501 East Conway, South Carolina. I don't remember now where the 108 James P. Blanton Circle address came from but it gives the right area map, even if James Blanton Circle is across the highway. I'll just pick a random location on the campus and see what that gives me.Vchimpanzee· talk·contributions·21:45, 27 August 2010 (UTC)
Damn right they pushed you off the policy page. What's policy got to do with finding a location? How did you come to write there at all? Anyway, you've not searched enough, Vchimpanzee. I've found the physical address. 2050 Hwy. 501 East Conway, South Carolina. you write -- yeah, that's what you were looking for, innit. So you found your own question? Then, if any question left, please give us information. Like: which websites did you use, what did you enter, what did they say. -DePiep (talk) 22:34, 27 August 2010 (UTC) Way too cynical. I struck. -DePiep (talk) 08:50, 28 August 2010 (UTC)
I followed the advice I was given above, and used this tool. The map for "2050 Hwy. 501 East Conway South Carolina" doesn't even show the school. In fact, what it shows is miles away from the school. On the other hand, "108 James P Blanton Circle Conway South Carolina" shows the school but the marker is in the wrong place. I can put the marker anywhere I want but just want to make sure I'm following the correct policy. On the policy page, by the way, I was told about another place I could ask the question, and I was told I was justified in asking the question there. The problem has been fixed. I won't even ask how the person did it.Vchimpanzee· talk·contributions·17:23, 28 August 2010 (UTC)
Actually, this person gave me this tool, but I haven't figured out how to make it give me latitude and longitude. I figure once the school comes up I click on the graduation cap. How that translates into coordinates I haven't figured out.Vchimpanzee· talk·contributions·17:31, 28 August 2010 (UTC)
This table appears to be randomly sorted, until one clicks on the arrow in the header cell. Is it possible to have this table automatically sorted by default when one arrives at the page? To be exact, the "Wikimedian" column should be alphabetically sorted downwards automatically and by default upon loading the page, without having to click the arrow. I think that table would be much more useful this way. But even so, I would like to know how this is done, if it's possible. Is it just a matter of formatting the table differently? I couldn't locate a precise answer in Help:Sorting. Thanks, Ϫ11:45, 28 August 2010 (UTC)
What would be the most efficient way to compare which articles two or more categories share? (Is there an online database for such a comparison...I thought I saw one a while ago.)Smallman12q (talk) 15:56, 29 August 2010 (UTC)
I wish MediaWiki had some kind of database support that would allow templates to access a database to populate the articles. This will make programming bots so much easier. What I envision is an ability to have a template access a table and display the sql query result in the article. Has this been discussed before? — Ganeshk (talk)01:40, 29 August 2010 (UTC)
Yes, a data wiki that can store and display tabular data is what I had in mind. Wikipedia is mostly text-based, but I was thinking it could be a mix of text and data. When the table values are changed, the articles will reflect the changes. I was thinking some type of wiki-customized database implementation that will diplay the history for each record on the table (who changed, when it was changed and diffs). The table can be a simple page that has comma-seperated data or as complicated as a full implementaton of a RDBMS. The templates must have ability (functions) to query this page and retrieve some record values. For example, taxonomy information for species can go into a table. If the classification of a genus has moved to another family, the table records for all species under the genus can be quickly updated with the new family information. Right now, this process is a manual one; find and replace text on the actual article. Text based content makes it difficult for software (bots) to update the content while a table can be easily traversed and updated. — Ganeshk (talk)15:54, 29 August 2010 (UTC)
It's technical; they would never display on the redirect page. Redirect templates used to display on redirect pages, but that hasn't been true for some years now. Graham8701:57, 31 August 2010 (UTC)
Secure links padlock too close
Consider this link: http://www.wiki.x.io. Notice how the padlock () is clinging to the link, due to padding-right being 13px, which was fine for the original yellow padlock (which was statically linked). The image is set in "index.php" and the padding in "mail-ltr.css", so I do not know where to override it. Perhaps a better solution is to change the image, as it has 3px room to spare to the right. — Edokter • Talk • 17:10, 30 August 2010 (UTC)
I just found it being declared in Vector.css (and Monobook.css). I'm going to try something in my own CSS. — Edokter • Talk • 19:14, 30 August 2010 (UTC)
No. The padlock image has a fixed size (16x13), and the default padding is 13px. (Edit:) Besides: IE8 seems to scale the image and padding according to the browser's font size (zoom) anyway. — Edokter • Talk • 22:02, 30 August 2010 (UTC)
And so does Safari. BTW. it is the yellow padlock that has changed in size (at least for Vector, not sure of other skins). Why the change in size was made, i'm not entirely sure, but probably has something to do with lineheight/font-size. —TheDJ (talk • contribs) 22:50, 30 August 2010 (UTC)
Over here in MonoBook it's a blue one that has a couple more pixels on the right than the left. (And if you're changing the style of the lock, then it would be padding-left you're after). OrangeDog (τ • ε) 23:06, 30 August 2010 (UTC)
The space to the right of the padlock is whitespace in the image itself; not much can bo done about that (unless you start playing with negative margins). Also, padding-right applies to the link span as a whole (not the image) in order to create space for the image, which is actually a right-alligned background image for the link; so it is the correct attribute. — Edokter • Talk • 12:35, 31 August 2010 (UTC)
Here I started a talk about an emoticon {{:)|devil}}, that put the whole article Bracket into the Talkpage, thereby putting a Userpage in Category. It is solved, but not understood yet. -DePiep (talk) 13:56, 31 August 2010 (UTC)
That code says to transclude the ) article (passing the parameter "devil" as 1=, which will be unused by the transcluded object), which redirects to Bracket. There is no such template named ":)", it is a bad title. Template:=) is probably what they were going for. –xenotalk14:03, 31 August 2010 (UTC)
I want to be able to search the revision history of an article for a text string. My objective is to find out when the text string was FIRST introduced into the article. I used the "Revision history search" link on the revision history page, which in turn uses WikiBlame, but, honestly, I simply couldn't understand the results. I even looked at the manual and was still in the dark. I posted a message to User:Flominator's Talk page, but I didn't get a response. Does anyone understand how this works? Can I achieve what I want?--Bbb23 (talk) 16:09, 31 August 2010 (UTC)
Yes, WikiBlame does what you want. You might need to enable the "force searching raw text" option and may need to check more than the default 50 revisions. Rjwilmsi16:32, 31 August 2010 (UTC)
Well, it may do what I want, but I'm unable to use the results. So, I'll break it down. I use the history of the Stephen Breyer article. I click on the search link. I get a WikiBlame form. I use the string "world justice project", which I enter in the Search for text box WITH quotation marks. I change the date to August 1, 2006, because I want to see the first time the string appears. I change the radio button selection to oldest first. There is no "force search raw text" option, only a "force searching for wikitext" option, and I don't know what that means. For the moment, I'm leaving it alone. I've doubled the versions to check from 50 to 100. I click on Start. Here are the results:
The version history of Stephen_Breyer is being searched for "world justice project" as plain text
100 versions found
Comparing differences in 14:08, 11 April 2006 between 50 and 51 while coming from 99:XX [Search from here]
Comparing differences in 02:12, 13 May 2006 between 25 and 26 while coming from 50:XX [Search from here]
Comparing differences in 02:23, 16 June 2006 between 12 and 13 while coming from 25:XX [Search from here]
Comparing differences in 13:59, 21 July 2006 between 5 and 6 while coming from 12:XX [Search from here]
Comparing differences in 14:49, 25 July 2006 between 1 and 2 while coming from 5:XX [Search from here]
The quotation marks are not necessary, and I guess they may be actually harmful (unless they literally appear on the page).
Leave the start date as today's date, you are searching from that date backwards. (If you want to search before another starting date, you have to make sure that the string actually appears in the revision you search from, otherwise it will probably fail.)
The number of revisions of the article (since you want to search the entire history) is much more than 100, so let's raise the number of versions to check to 1000.
You get:
The version history of Stephen_Breyer is being searched for world justice project as plain text
809 versions found
Comparing differences in 18:01, 8 February 2007 between 404 and 405 while coming from 808:XX [Search from here]
Comparing differences in 09:54, 28 June 2008 between 202 and 203 while coming from 404:XX [Search from here]
Comparing differences in 18:22, 11 March 2009 between 101 and 102 while coming from 202:XX [Search from here]
Comparing differences in 13:45, 12 January 2010 between 50 and 51 while coming from 101:XX [Search from here]
Comparing differences in 23:25, 19 June 2010 between 24 and 25 while coming from 50:OO [Search from here]
Comparing differences in 12:54, 25 March 2010 between 37 and 38 while coming from 24:OO [Search from here]
Comparing differences in 15:08, 17 February 2010 between 43 and 44 while coming from 37:XX [Search from here]
Comparing differences in 22:50, 23 February 2010 between 40 and 41 while coming from 43:X 0
Insertion found between 17:16, 18 February 2010 and 22:50, 23 February 2010
And to clarify one thing: the diff found by this method is an edit that introduced the string into the article, but it is not necessarily the first such edit, in principle the string may have been introduced and deleted several times. It is not possible to find the first edit without sequentially checking every single revision of the article until a match is found; you can force this kind of search by checking "Search method: linear" and "Order: oldest first", but you should avoid this unless really necessary as it is a drain on resources and takes a lot of time.—EmilJ.17:17, 31 August 2010 (UTC)
It looks like the MediaWiki parser function #replace function does not work on Wikipedia. Is it not installed? Are there any alternatives? I really need this function to create some templates
All of the StringFunctions (everything on the page you linked) are not installed due to the performance burden they would add. Prodegotalk19:06, 31 August 2010 (UTC)
I think citing performance in this context is a red herring. Though they could have performance implications, the last serious discussion on the topic I am aware of basically boiled down to: "Template syntax sucks, and we don't want to add anything more to that until we have an alternative." Dragons flight (talk) 19:20, 31 August 2010 (UTC)
(edit conflict) According to the developers string functions are not enabled because they would prefer to embed a real programming language, such as Lua. But they are also not interested in working on embedding a real programming language right now, and any embedded programming language must be able to run securely without PHP extensions or external programs so people running MediaWiki on really crappy webhosting can still copy our templates. This old thread has a humorous summary of the years-long discussion regarding StringFunctions. Anomie⚔19:35, 31 August 2010 (UTC)
Where are interface translations set? When language is set to Serbian (and probably others, though problem does not occur for "fr") the "rollback" and "undo" button use the same word sometimes leading to issues (see [19]). –xenotalk15:36, 31 August 2010 (UTC)
Fairly regularly, I find that a page I'm watching has disappeared from my watchlist. When I realize, I have to view the page and add it again, but by then I may have missed talk/article developments I should have been aware of. Is this a known issue? PL290 (talk) 08:31, 1 September 2010 (UTC)
I don't believe it has ever happened to me, but I usually have about 5-700 pages on my watchlist. How large is your watchlist? NW(Talk)11:05, 1 September 2010 (UTC)
Good thought, but no: I have AWB set to "leave watchlist unchanged", and it appears to abide by that—and I can think of at least one page I've never AWB'd that suffered from the problem just recently. PL290 (talk) 17:48, 1 September 2010 (UTC)
Is there a way to search a particular namespace for a specific string?
Is there a way to search a particular namespace for a specific string?
For instance, using the search function to look for two words seems to return results for "word1" OR "word 2". What if you want to search for "word1" AND "word 2" or "word1 word2" ? pablo22:23, 31 August 2010 (UTC)
It does AND by default. You can verify this by looking at the total number of results returned for initial words, and for words together. You can use AND/OR special words with - for NOT. e.g. Special:Search/word1 OR word2 -word3 which will return all documents with word1 or word2, but not word3. --rainman (talk) 12:50, 2 September 2010 (UTC)
Character set used on zh.wikipedia
I know this Village Pump pertains to en.wikipedia, but my Chinese is too nonexistent to ask at the zh.wikipedia equivalent page: can anyone tell me which character set (i.e. traditional, simplified, or something else) is used on the Chinese Wikipedia at zh.wiki.x.io? Many thanks. Gonzonoir (talk) 11:06, 1 September 2010 (UTC)
The Chinese Wikipedia uses a tab at the top-left corner of the page that switches the text between Taiwan Traditional (two forms, apparently), Mainland Simplified, and less well-known varieties. The actual content of the articles is the same in any character set. Intelligentsium00:09, 3 September 2010 (UTC)
Won't Jump
Please look at That '70s Show#Running_gags. Note that there are many links to episodes. Each link jumps to the spot on the page where the episode is EXCEPT Street Fighting Man, which just displays the page normally (starting at the top). I don't see anything I've done wrong in how I link to that episode. I've even tried to compare the underlying table code for the season 7 article (Street Fighting Man) with, for example, season 1 (Stolen Car), and I don't see any obvious difference that might make season 7 misbehave. I'm sure it's something very silly I've missed, but I'm tired of staring at it.--Bbb23 (talk) 23:15, 1 September 2010 (UTC)
Fixed. Some of the episode templates had non-breaking spaces before the pipe characters, so the link identifiers included the non-breaking space after the episode number. I replaced these with ordinary spaces which get stripped when the page is rendered. — Richardguk (talk) 00:09, 2 September 2010 (UTC)
A non-breaking space (or a hard space) is a character that prevents two adjacent words' being separated should they happen to lie at the end of a line. This is useful for keeping quantities and units together but can be a problem in cases like this because they are interpreted differently from normal spaces. WikEd or another tool might highlight these, but often you just have to recognize the 'symptoms' of the problem you are having. Intelligentsium01:28, 2 September 2010 (UTC)
Indeed. I would add that this is not a common problem on Wikipedia, because most non-breaking spaces are encoded in the wikitext as the corresponding HTML character entity reference .
To help you fix the same problem in any related articles, notice the lines in the wikitext which are marked as having changed with my edit. In each case, there is a character that looks like a space immediately before the wikitext |EpisodeNumber2 but previously the character in this position was in fact a non-breaking space.
Ironically, though the character was impossible to see, my web browser seemed to convert it to a space when it displayed the edit box, so I only had to copy and paste the entire text back in and save it again. Or at least, I think that's what fixed it!
Question: What sorts of edits result in such drastic increases in page views?
To understand my question, first look at http://stats.grok.se/en/201005/Sony_Dash (page view statistics). So on May 21st I edited and moved Sony Dash from Dash (personal internet viewer). I found pictures on Flickr and added an infobox. Out of curiosity I checked the pageview stats, and was surprised to see that starting that day the page views went up drastically. The previous month had a total number of 230 views, the following month 2231. Now I understand that adding content and a picture make it show up on more search engines, and I can assume that some of those page views on that day are from me and maybe some Commons people checking out a new image, but what about all the days after? I see my edits to this article as relatively lazy, and the majority of the text I added were bare facts in the infobox. I'm certain that the WP search takes days if not longer to register changes, and I changed nothing on the Sony template which is where most of the inbound links come from, so I'm curious if anyone knows what edits in particular could create almost instant results like this. Where do all these new views come form and how, while still being lazy, can these results be reproduced? ▫ JohnnyMrNinja07:46, 2 September 2010 (UTC)
In general, to increase page views, you can add links to the article on other pages, interwikilinks, redirects, and categories. you can also try to get the page on the front page as a WP:DYK or a WP:FA. Finally, try to get the subject of the page in the news, and your pageviews will increase drastically. Idose, a carbohydrate, went from 40-50 views a day to 100 a day and more when I-doser hit the news, only because of mistyped searches. Raymond Gafner goes from 300 monthly views to 3000 monthly views every two years for one month during the olympics, just for having won the Pierre de Coubertin medal in 1999 (the latter article similarly goes from 1200 or so vies a month, to 30,000 views for one month during the Olympics). Fram (talk) 09:14, 2 September 2010 (UTC)
Style for headings
I'm looking for the CSS which will produce the headings like we use on Wikipedia. I want to hardcode rather than using == or <h2> because I do not want it to appear in the table of contents. For background, this is for the Template:Invitation to edit initiative which will inivite readers to edit certain articles and produce a collapsed mini tutorial. It's these headings which I don't want to interfere with the article itself. Thanks — Martin (MSGJ · talk) 19:46, 2 September 2010 (UTC)
Is there a change I can make in my vector.css that will cause the notice to always appear? Or in a different similar file? Thanks, meshach (talk) 21:46, 1 September 2010 (UTC)
In other words: the OP wants the page size to be listed at the top of every page when editing it, even if smaller than 32kb. –xenotalk17:52, 3 September 2010 (UTC)
Old Wikipedia logo has disappeared in Monobook
Maybe it's just me, but the Wikipedia logo has disappeared from the upper left corner of the page in Monobook. Clicking the empty space still takes me to the Main Page, but the logo is blank. Logged out I can see the new logo in Vector. Any idea what happened? –Grondemar05:20, 3 September 2010 (UTC)
still there for me – not sure if there's a choice of image flavours though, and if so, it's the right one for you! Always use Monobook. Trev M ~ 12:43, 3 September 2010 (UTC)
Anyone else screaming because of unnecessary case sensitivity in templates?
I was just trying to assign an article a wikiproject class. I have done this several times and would have thought it could be a fairly intuitive exercise, having read the criteria for assigning articles. But to get the 4 permutations of upper and lower cases of "Class" and the letter signifying the value right took me... yees, you guesssed it, 4 goes. And then there was getting the order of the quality and importance right to please the templates.
I presume there's a master template for the plethora of minor wikiproject categorization templates (many unfilled, perhaps not helped by this tediousness). Wouldn't some template wiz feel great about disposing of unnecessary case (and order) sensitivities to help move this along a bit smoother? Trev M ~ 12:38, 3 September 2010 (UTC)
I want to use the {{{name}}} in fullurl: like this:
[{{fullurl:template:abc{{{name}}}|action=edit}}] however, as you can see, it does not seem to work. Is there any other way to make this work?--Hengsheng120 (talk) 20:31, 3 September 2010 (UTC)
It does seem to work, as you say but maybe didn't mean. It gives the result I would expect when it's transcluded with a parameter called name. For example, the code {{Wikipedia:Village pump (technical)|name=example}} currently makes your above code produce a link to http://en.wiki.x.io/w/index.php?title=Template:Abcexample&action=edit for me (who is at http://en.wiki.x.io). Did you mean it does not seem to work? In that case, what exactly do you want to happen when exactly which code is used? Or if you don't know which code to use then which effect do you want to achieve? PrimeHunter (talk) 21:04, 3 September 2010 (UTC)
Every link on the french Wikipedia is archived in the Wikiwix archive. By default, we chose to display the archive only in the references section because that's where we need them. It was also done to improve performance. The rest of the links can be displayed using a gadget. Dodoïste (talk) 16:17, 1 September 2010 (UTC)
Jack, I think that's a brilliant idea. t would solve some of the huge problems where pages , especially on government sites, are removed because the webmasters beleive they are no longer of interest. We lose a lot of valuable WP:RS this way, and articles then suffer from dead links. Please consider starting a project page to discuss this. I would imagine it would not be to hard to obtain a consensus.--Kudpung (talk) 03:35, 4 September 2010 (UTC)
Jack, in response to your message on my talk page, I felt I had already offered the best suggestion. However, WT:RFA is not really the best place for such a discussion because it's where we discuss our policy regarding the making or breaking of admins, which of course is an entirely different topic. What I meant was our system of WP:RfC. This is best opened as a sub page of the policy it is expected to change or improve. Your suggestion, in my opinion, would be an RFC opened as a sub page after you have read our en.Wiki policies here, in this order;
I would suggest that perhaps the WP:CITE, or the WP:REF/ES page would be a good place to locate an RfC. I see little advantage in translating the entire French discussion (I have read it, but most people might possibly consider it WP:TLDR, and would want to draw their own conclusions). What I do suggest, however, is translating the proposition of the French debate, and adapting it to our purposes here at en.Wiki. The way to go now is to start everything by drafting in your user space what you would like to do, and then translating it into English. I would chip in and give you a coup de main there. Keep me up to date again on my talk page.--Kudpung (talk) 07:45, 11 September 2010 (UTC)
Do admins have access to 'HideUser'? I'm not sure what that is. If you're referring to the "Delete editor's username/IP" option in RevDel, that still doesn't hide the name from the Special:ListUsers listing, I don't think. -- Ϫ02:16, 4 September 2010 (UTC)
Hi! I've noticed that with some old IP-addresses which edited uninlogged like the same year as Wikipedia was formed, not the whole IP-address (e.g. not all the numbers in the IP-address) are shown. Why is it like that with the oldest edits? Were the actual IP-addresses not identifyable at that time? /HeyMid (contributions) 22:20, 3 September 2010 (UTC)
I think it was just a change of privacy policy. Masking the last 8-bits was probably judged to provide little additional privacy for many anonymous editors, and would have been confusing for editors or readers with a similar address to that of an editor who had received a user talk page message. — Richardguk (talk) 23:35, 3 September 2010 (UTC)
That would be true, except I have never encountered any evidence of someone trying to communicate with an anonymous user in those days. I thinkt the last octet of the IP address was masked with "xxx" to make it slightly more difficult to do WHOIS lookups. IIRC there was a time when admins could see the full IP address while other users could not. Also see Wikipedia:Phase II feature requests/Cookies, logins, and privacy, which was written when Wikipedia used UseModWiki. Graham8707:41, 4 September 2010 (UTC)
Hello! Thank you for asking. Well, I am running the secure version, and I also had downloading issues yesterday. It took like 5–10 seconds to load a page, which is kinda slow, but it is a bit better today. So no, you weren't the only user who were experiencing slow-down issues yesterday, probably all users around the world were affected. /HeyMid (contributions) 07:39, 4 September 2010 (UTC)
allow non-registered users to choose the skin they like
Not everyone likes vector & not everyone wants to become a registered user. Therefore, it should be allowed to set a preferred skin (to be stored in a cookie).
Likewise, not everyone's eyes are as good as they used to be. Therefore users should be allowed to set a preferred font-size as well.
Hpvpp (talk) 06:20, 4 September 2010 (UTC)
Using Ctrl + and Ctrl - should adjust the font size in your browser. Most browsers should automatically remember the setting. This has nothing to do with the skin or registration. Zunaid06:50, 4 September 2010 (UTC)
Adding different versions of pages for anonymous users is a gigantic penalty on the caching capabilities. That is the reason it isn't implemented already. If people want custom versions, they can register an account. —TheDJ (talk • contribs) 10:06, 4 September 2010 (UTC)
Well surely the majority of pages are cached in monobook for registered users? Incidentally do we have any page impression stats by skin for registered users? RichFarmbrough, 17:57, 4 September 2010 (UTC).
Don't logged-in users usually bypass the squid cache, among other reasons because the page has their username at the top? Which would mean the squids don't actually have pages cached in monobook. Since the squids are said to handle about 78% of all requests (almost all of which are from non-logged-in users), just offering monobook versus vector to anons could make a significant enough difference to make the sysadmins unhappy. Feel free to ask them if you want to know for sure. Anomie⚔20:16, 4 September 2010 (UTC)
Correct. Logged in users only benefit from the parsercache. From the point where the html is generated, each user will get his own html (otherwise your username couldn't appear at the top of the page for instance). —TheDJ (talk • contribs) 20:52, 4 September 2010 (UTC)
Well there would be ways and means, but if that's how it works. Any idea on the page impression stats by skin for registered users? RichFarmbrough, 23:34, 4 September 2010 (UTC).
Some clarification might be in order. I don't like vector and have set my preference to monobook. But that is just me. I would have clicked <take me back>, but I feared that would undo all the editing improvements as well. Point is, however, that I am not at all convinced that the world-out-there agrees that vector is better. For example, my mother (86 years) makes very occasional use of Wikipedia, but has difficulty reading the screen and so prefers larger fonts. I told her that Ctrl-+ does the trick, but she forgets. Indeed, she is not really computer literate. Point, however is that she does not appreciate change where she sees no need. She is not sure whether vector is an improvement or not, but says that it worked well the way it was and changing something that does not need change is only confusing. And that puts her off.
Now, I am sure that most people use Wikipedia for what it is intended, namely to get information. And I am also sure that most people don't want to be bothered getting an account when other services on the web offer comparable functionality without that hassle. And whether vector is better than monobook should not be the outcome of a poll amongst registered users, but the result of a proper scientific study. And as long as there is no scientific proof that vector is actually better it should not be the default. Wikipedia is there for users and not just for registered users. Thus, my request to let the choice of skin be stored in a cookie. Hpvpp (talk) 01:05, 5 September 2010 (UTC)
Ahem, I believe you are wrong. The usability initiative goal was "to measurably increase the usability of Wikipedia for new contributors" and the studies were aimed "towards greater ease of use for novice users". Note that this means in both cases contributors. And contributors are likely to have a different outlook than most people who just go to Wikipedia to get some information. Hpvpp (talk) 03:36, 5 September 2010 (UTC)
My main point was it was not a poll of users as you said. One of the main goals of it was to reduce the divide between "readers" and "editors" by making it easier to learn to edit. The majority of the study participants had never edited before. I would say that most, if not all editors originally came to Wikipedia to get information, so I don't think their outlook is that different. In any case, the initiative also focused on site navigation, which is why the search box was moved and the sidebar links (most of which are unused by readers) were collapsed by default. Though those were really the only changes for readers. The actual layout of articles didn't change. Mr.Z-man04:29, 5 September 2010 (UTC)
I concede your point, though my argument still stands: the majority of user are not contributors, but will still have preferences which should be honored. —Preceding unsigned comment added by Hpvpp (talk • contribs) 05:13, 5 September 2010 (UTC)
"intitle" and "lookfrom" Wikipedia search options for finding section titles ?
I find the "intitle" and "lookfrom" Wikipedia search options (as in intitle:"search phrase" and lookfrom:"search phrase") to be very useful to find articles relating to a subject of interest.
However it would also be useful if I could look for section titles within articles using "intitle" and "lookfrom" rather than just article titles. I bet there are plenty of sections tucked away which would be of interest to me and i might llke to link to.
If it isn't possible with standard Wiki software, perhaps somebody could do a couple of searches for me on my behalf using specialist Wiki software.
I suppose one possibility would be the ability to search on the raw HTML version of Wikipedia and make use of the fact that section titles always start with a "=".
--Penbat (talk) 10:25, 4 September 2010 (UTC)
I'd prefer to say "raw wikicode" but if you have a couple specific searches that don't require a bang up-to-date version of Wikipedia, drop me a note on my talk page. RichFarmbrough, 17:54, 4 September 2010 (UTC).
Since the search results often point out headers in pages that include your search term, I'm guessing it might be that the headers for pages are stored somewhere in the searchindex. Wether or not they can be individually targeted with the current software, I'm not sure, but if it is in the db already, then perhaps this could be implemented if not yet possible. Try bugzilla. —TheDJ (talk • contribs) 20:58, 4 September 2010 (UTC)
Odd edit summary
For me with Firefox this displays a smudge, it crashes IE [21]. The entry is the one for Zalgo 02:48, 24 September 2009. Probably needs a bugzilla raising. RichFarmbrough, 17:51, 4 September 2010 (UTC).
Yes, it does look rather odd in FF, but I guess it displays properly - not that I know what it's supposed to look like - with Safari and Chrome. If IE can't display it, then Microsoft has another bug to fix. :D —DoRD (talk) 20:03, 4 September 2010 (UTC)
I don't see the advantage to allowing stuff as complex as this text in edit summaries, although I can see the advantage of allowing "normal" unicode. Certainly it seems a little unwise without going into BEANS. RichFarmbrough, 23:30, 4 September 2010 (UTC).
Notifying Devs about a Bugzilla request
I understand that the devs are busy. But I was wondering how does one bring a bugzilla posting to their attention?
I posted one (11499) several years ago, and it's still listed as "new". With no comments that I see.
It's not a big deal, but I would find it incredibly useful.
Is there a way of validating the text specified for external website addresses in links? I know there's the help section which includes information on external link syntax with "http://", but is there a way of asking the editor to confirm if [www.website.com] should be [24], even if it is just a prompt "did you mean http://www"?
I think I'm right in saying that the software only knows it's supposed to be a link because of the existence of a protocol prefix. That said, www. probably could trigger something and not interfere with this. The only way of implementing it quickly I can think of would be through the edit filter, but the performance overheads might be too great. - Jarry1250[Humorous? Discuss.]11:58, 5 September 2010 (UTC)
Confusion with new articles which were redirects
Wikipedia seems to have problems detecting that a new article is a new article if it was converted from a redirect. They dont appear on NEW PAGES and last time i checked any new articles i converted from a redirect werent recorded against my user name in my user stats as being new articles created by me. Also is {{NUMBEROFARTICLES}} (currently 6,873,118) accurate ? Does it allow for conversions from redirects to new articles ?
As they dont seem to appear on NEW PAGES, they arnt subject to the scrutiny of new page patrol that new articles do. It seems like a good wheeze to set up a redirect then later replace it with some sort of garbage article (such as a hoax or spam or BLP violation) and there is a fair chance it wont get detected.
I'm currently looking for beta testers for the Wikipedia:Script Installer. In brief, what the tool does is it allows you to install any script with just one click, rather than need to go to your skin.js page and add the script in there manually while following a set of poorly worded instructions. If you have the time, please install this tool, play around with it, and if you find any bugs or problems, then post then on the talk page. Feature suggestions are welcome, but I've got a list of almost 100 items on my todo list, so there's a good chance that what you want is already in the works. To give you a taste of how the tool looks like, see the following gallery:
Script documentation page
Script Library
Script page
Skin page
Toolbox
Probably the feature that's missing the most is the ability to uninstall scripts, which is coming. However, I figured that it's more crucial to get the installing part working, first. Gary King(talk · scripts)21:03, 5 September 2010 (UTC)
I didn't mean for the source code to be examined, but in any case, I was planning on uploading the code soon. The code's not obfuscated; you can easily decode it using base62, which I'm using so as to decrease the script's size by about 70%. At the moment, the code needs to be cleaned up a bit first and documented, but it should be available within 24 hours. Gary King(talk · scripts)23:47, 5 September 2010 (UTC)
If anyone's interested in the tool's development, then please watchlist the page to watch for any discussions regarding ideas about the tool, since most things are still merely ideas for now, and could use some useful input from others. Gary King(talk · scripts)04:21, 6 September 2010 (UTC)
Are you a developer that has written a script or two?
For developers who have written their own scripts, there are also ways to make them work with the Script Installer. Right now, when a user visits a .js page, the Script Installer will tell the user whether or not it is verified, and create a link for the user to click to install the script, if they want. However, when a user visits a documentation page, the Script Installer can't tell if the page is a documentation page or not, so the {{Script data}} template needs to be added to the bottom of the page to indicate this. This will then show the "Install this script?" box at the top of the documentation page. Some sample code:
{{script data | page = User:Gary King/script installer.js }}
It's because of the unclosed <span style="font-size: smaller;" class="autosigned"> at the end of the previous section, which was added in this edit. Anomie⚔02:05, 6 September 2010 (UTC)
I'd never noticed until now that when I click on a pronunciation link I ma brought to a new page to play the sound and then have to click the back button in order to continue reading. I'm using the latest version of Firefox but was wondering if this is intended as default behaviour. Try for yourself:
Just out of curiosity, what is the purpose of the 'contentSub' div that appears on pages just below the 'From Wikipedia, the free encyclopedia' line? as far as I can tell, it does nothing except add a line of (potentially unnecessary) whitespace. at least, I've never seen a case where it actually has content. is it purely decorative? --Ludwigs202:21, 6 September 2010 (UTC)
It might have the "(difference between revisions)" if you are on a diff page, "View logs for this page" on a history page, the "redirected from" text if you went to a redirect, the "Redirect page" text if you are viewing a redirect page, the parent pages link list if you are on a subpage, and so on. Anomie⚔03:14, 6 September 2010 (UTC)
In the preferences, as a "Lab feature" under Editing, I've found "Enable preview dialog". After looking at the source code I think that some button for a preview in a popup window should appear, but I can't find anything. I asked about this in the German Wikipedia, but nobody could help. Do you know what this option does? --Schnark (talk) 09:47, 7 September 2010 (UTC)
Tourist Place Nrusinghnath being removed !
Hi,
I have been creating a page on a famous tourist and religious place of Western Orissa, which is being removed by the administrator.
This is to clarify that the article does not contain any advertising material, and sole purpose of the same is to give genuine information regarding the place.
I am a localite from that place, and know it from my childhood.
Hope it will be restored back. If any writing that seems as advertisement of any sort, then please intimate me, so that I can write it up properly.
Please sign your post by typing four tildes (~~~~) or clicking the signature button above the edit box which looks like this: . Do NOT sign in articles.
I have just created Template:EstcatCountry based on Template:Estcat. Now I realized that I need to have a sort key routine implemented for categorizing such countries as the Netherlands, the United Kingdom and the United States.
I need to have this statement modified: [[Category:{{{1}}}{{{2}}} establishments by country|{{{5}}}]] so that if a parameter (or is it variable?) {{{6}}} exists (which would be the country name without "the", e.g. Netherlands) then that should be the sort key instead of {{{5}}}. See for instance Category:1956 establishments in the United Kingdom. __meco (talk) 10:10, 7 September 2010 (UTC)
Look at the category shown on the bottom of the page the categore is placed on; categories take time to populate. (Although I just realised you can't see the sort key on the page.) Just wait for the category to update. — Edokter • Talk • 13:24, 7 September 2010 (UTC)
I'm renewing my call for help on this issue. The categories still have not updated. Could someone please look into this issue again? __meco (talk) 07:08, 8 September 2010 (UTC)
I've been editing templates with such code from time to time. I have even created templates using existing, similar templates as "templates", making minor modifications to make the new code work. However, when the code becomes even a little bit advanced, including such elements as #ifeq: I usually glaze out, and no documentation page has yet been lucid enough for me to be able to learn these things. Therefore I figured I would try another approach now that I'm again stuck in the editing of a template. I'm looking to create a new template based on Template:EstablishmentsInDecade, or if that proves more feasible, modify this template to also do what I am looking for it to do. Now, what I'm initially requesting is for someone to spell out the meaning of the following code for me:
Organizations, places or other things founded or established in the '''[[{{{1}}}{{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}}{{{2}}}s]]'''.
When applied, the input to the template would typically look as follows: {{EstablishmentsInDecade|19|70|20th}} __meco (talk) 16:50, 7 September 2010 (UTC)
Let me call the value of first parameter CC and the value of the second parameter DD. If DD equals (numerically) 0, then the inner subexpression {{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}} expands to 00–CC09|CC, otherwise it disappears. Thus the whole expression expands to ...[[CC00–CC09|CC00s]] if DD is 00, and it expands to ...[[CCDDs]] if it is nonzero.—EmilJ.17:19, 7 September 2010 (UTC)
It could be written in a less obfuscated way, IMHO: [[{{#ifeq:{{{2}}} | 0 | {{{1}}}00–{{{1}}}09{{!}}}}{{{1}}}{{{2}}}s]].—EmilJ.17:28, 7 September 2010 (UTC)
Wow! I'm floundering, basically. I must revisit this explanation more times to see if perhaps I shall be able to comprehend what you just wrote. __meco (talk) 17:56, 7 September 2010 (UTC)
Could you make an attempt with this staement of you own:
If DD equals (numerically) 0, then the inner subexpression {{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}} expands to 00–CC09|CC, otherwise it disappears.
Could you try and explain only that bit? Please start with the segment before the minus sign and how to read it! __meco (talk) 18:30, 7 September 2010 (UTC)
(It's not a minus sign but an en-dash, but that does not matter here.) {{#ifeq:A|B|C}} means "if A equals B, then expand to C, otherwise do nothing". Here, A is {{{2}}}, B is 0, and C is 00–{{{1}}}09{{!}}{{{1}}}. Thus, if {{{2}}} (i.e., DD) is not zero, then the subexpression has empty expansion. If DDis zero, the #ifeq-function expands to C, that is, 00–{{{1}}}09{{!}}{{{1}}}. Here, in turn, {{{1}}} expands to what I denoted above as CC, and {{!}} expands to | (it's encoded here like this so that it does not conflict with the |'s used in the #ifeq-expression syntax), giving 00–CC09|CC.—EmilJ.18:45, 7 September 2010 (UTC)
In the example given above for the use of this template, {{EstablishmentsInDecade|19|70|20th}} we would see {{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}} become {{#ifeq:70 | 0 | 00–{1909{{!}}19} which would according to what you have already explained in detail become 70–{1909{{!}}19}? If I'm right so far, how should this then be read and expanded? __meco (talk) 19:05, 7 September 2010 (UTC)
I think it actually expands in a different order, but anyway: {{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}} becomes {{#ifeq:70 | 0 | 00–1909{{!}}19}}, which turns to 00–1909{{!}}19, which in turn makes 00–1909|19. Remember that templates and parser functions are delimited by double braces, but parameters are delimited by triple braces.—EmilJ.19:17, 7 September 2010 (UTC)
The color coding helps seeing the expression more clearly, but it isn't enough to make me understand everything. What about the single bracket around {1909{{!}}19}? Also, having apparently fully expanded this expression, the result makes no sense. [[{{{1}}}{{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}}{{{2}}}s]] becomes [[19{{70–1909{{!}}1970s]]. What do I do now? __meco (talk) 19:34, 7 September 2010 (UTC)
Oops... I saw a mistake there. The last expression would become [[19{{70–1909{{!}}19}}70s]]. Still, I don't know how to go further with that. I'll make a try though. How about [[1970–1909|1970s]]. That doesn't make sense though, as the correct expansion obviously should be [[1970–1979|1970s]] __meco (talk) 19:43, 7 September 2010 (UTC)
Assming a call {{EstablishmentsInDecade|19|70}}:
[[{{{1}}}{{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}}{{{2}}}s]] → substituting {{{1}}} with 19 and {{{2}}} with 70 →
[[19{{#ifeq:70 | 0 | 00–1909{{!}}19}}70s]] → 70 ≠ 0, so remove the whole #ifeq →
[[1970s]]
If {{EstablishmentsInDecade|19|00}} was used instead:
[[{{{1}}}{{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}}{{{2}}}s]] → substituting {{{1}}} with 19 and {{{2}}} with 00
[[19{{#ifeq:00 | 0 | 00–1909{{!}}19}}00s]] → 00 = 0, so leave the inside of the #ifeq →
[[1900–1909{{!}}1900s]] → {{!}} is a template that evaluates to | →
Oh good! Now I understand it. I was in error in my last post above when I wrote [[19{{70–1909{{!}}19}}70s]]. When I evaluated the #ifeq: expression of [[{{{1}}}{{#ifeq:{{{2}}} | 0 | 00–{{{1}}}09{{!}}{{{1}}}}}{{{2}}}s]] I perceived #ifeq:{{{2}}} | 0 | 00 as the expression, discounting the rest. So I read "if 70 equals 0 then substitute with 00, otherwise leave as 70". I then took 70 and expanded that with the rest of the expression: [[19{{70–{{{1}}}09{{!}}{{{1}}}}}{{{2}}}s]] to become [[19{{70–1909{{!}}19}}70s]] and finally [[1970–1909|1970s]]. I'm not sure you follow my warped expansion, but at least I see clearly what I did wrong. __meco (talk) 21:08, 7 September 2010 (UTC)
Emil, as you can tell if are able or bother to go through the details of my misunderstandings and sorting these out above, you will realize that despite the added visual aid of the color coding I still missed realizing that the third part of the #ifeq: expression wasn't 00 but everything that followed until the closing double bracket (which you had made dark blue to match the opening double bracket). __meco (talk) 21:16, 7 September 2010 (UTC)
A minor suggestion
On the "1909" problem in [[1970–1909|1970s]]: if you can produce "1970" (somehow), you can calculate 1979: using {{#expr: expression }}: {{#expr: 1970 + 9 }} --> 1979 —Preceding unsigned comment added by DePiep (talk • contribs) 20:11, 7 September 2010 (UTC)
I'm afraid I don't understand that. What does {{#expr: 1970 + 9 }} mean? Please compare with how the function of #expr: was explained to me above by Emil J. __meco (talk) 20:31, 7 September 2010 (UTC)
#expr wasn't explained to you above. What {{#expr: 1970 + 9 }} does is evaluates the mathematical expression "1970 + 9", giving output of "1979". Anomie⚔20:37, 7 September 2010 (UTC)
(only adding, after Anomie) Just as {{#ifeq:... starts a comparision, {{#expr:... starts a straight calculation, using numbers & maths. That is how it calculates end of the decade by adding 9 (years). There are no more parameters.
Now to your question: calling by {{EstablishmentsInDecade|19|70|20th}}, you can create the "1979"-part as follows: {{#expr: {{{1}}}{{{2}}} + 9 }}. Because first it does {{{1}}}{{{2}}} -> 1970, then it does the math: 1970+9=1979. -DePiep (talk) 20:47, 7 September 2010 (UTC)
Although I now appreciate how the #expr: works, I also realize why it isn't being used in the template. The result of the expression is going to turn out a decade, e.g. 1970s. However, that is not sufficient for 1900s (and you will understand why if you click the wikilink. For the first decade of any century the outpust has to be of the form [[1900–1909|1900s]] (see 1900–1909). __meco (talk) 06:44, 8 September 2010 (UTC)
Well, if the template is to be used on Decade-pages only, and if it is even named "...Decade", it should not end up at the 1900s disambiguation page at all. The Decade-thing already says it should be linked to 1900s (decade), skipping the dab-page. Or, altetrnatively, if you want that option (20th century), rename the template and add one more parameter (say 3=d or c for decade/century, default=d) -DePiep (talk) 07:35, 8 September 2010 (UTC)
Ah, your question was about 'explaining', not creating or improving. I must have distracted you with unsollicited help... I will make good another time. -DePiep (talk) 08:30, 8 September 2010 (UTC)