This page contains discussions that have been archived from Village pump (technical).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 think it is related to a known problem with the image server. Twinkle kinda works for me, but most of my other tools that in any way depend on images are dead. MBisanztalk22:10, 12 July 2009 (UTC)
Is it something to do with the last edit at 22:36 to Common.js ? My Firefox is showing a message in the error console abour area undefined in Common.js/shuffle.js -- WOSlinker (talk) 22:19, 12 July 2009 (UTC)
and the line before is probably not kosher as well. Also, iterating arrays with for (foo in bar) will break if arrays are extended with functions. Amalthea22:28, 12 July 2009 (UTC)
It wasn't working because that whole statement is misuse (probably from misunderstanding) the getElementsByClassName, which is not a built-in function; it's defined in wikibits.js. It's just a function, in the form getElementsByTagName(document, 'tag name', 'class name'); - so it wasn't working. People need to test code before just activating it in common... Ale_Jrbtalk22:33, 12 July 2009 (UTC)
(ec) Ah, I see it was supposedly guarded in Common.js. If that would have been if (document.getElementsByClassName( 'shuffle' ).length) it wouldn't have had issues. Amalthea22:35, 12 July 2009 (UTC)
Unfortunately, as far as I know, the native (document.) version of that function is only supported in Firefox, Safari and Opera (and probably Chrome) - that is, not in IE. So it would still have caused issues for a lot of people. That's the whole point of the wikibits function, which uses the native implementation if possible, but works around it if not. Ale_Jrbtalk22:39, 12 July 2009 (UTC)
<trouts self/>. Sorry guys; I tested it on FF3.5 both with and without the elements on the page, but I think I only tested in IE and Safari erratically. And then by the time I'd finished it turns out it's not even wanted... :( And I should know better than to live-hack JS files... <goes to cry in corner/> Happy‑melon12:48, 13 July 2009 (UTC)
The error was in Common.js, as in the common script file for all skins, so people using other skins are of course affected by errors in it. Amalthea19:48, 13 July 2009 (UTC)
Ranking by number of edits is dumb. Some people make lots of edits when they don't have to (I won't name any names) and it certainly would skew such a count. Plus, it shouldn't matter anyway. ♫ Melodia Chaconne ♫ (talk) 23:53, 12 July 2009 (UTC)
I'd like to see who the most prolific editors are, so I can then explore what their areas of interest/activity/expertise are, so I know who to contact for what. I don't want to contact editors who have 2 edits per month. So yes, it does matter. The Transhumanist19:45, 13 July 2009 (UTC)
Hello. Just reporting. I use Opera 10.00 Beta and Friendly, but when I use the welcome or warning options of Friendly, I am unable to scroll down in the pop up (to select another template below of which I see.) Is this a Friendly known bug? Or should I reconfigure Opera? So far it is the only issue I've had with the browser. RUL3R*flaming | *vandalism17:35, 13 July 2009 (UTC)
You set the width of the second column to the same value in each row, but the value you chose (11.25em) wasn't large enough to fit the text in, so the cells expanded appropriately. I've set the widths to 15em, which is enough (with my settings and browser, anyway). Algebraist03:24, 13 July 2009 (UTC)
Wanted to know if any change has gone in to the login page. I'm unable to login using Opera since the last 2 weeks. Clicking on the "Log in" button does nothing, also the same with hitting the Enter key. I had been using this browser for Wikipedia without any problems for years. My opera version is 8. Jay (talk) 10:20, 8 July 2009 (UTC)
The only thing that I am aware of is that our servers newly serve jQuery with every page (which apparently causes alerts about ActiveX on IE6 with strange "security" settings). Maybe that causes problems on Opera 8, too. Opera 8 is a browser not supported by jQuery, according to their Website, they limit themselves to IE 6.0+, FF 2+, Safari 3.0+, Opera 9.0+, Chrome. (I wonder about other older browsers. IE5.5, IE5, Konqueror? We recently had someone using IE5.01 commenting here...) Using jQuery is all fine and dandy, but if this breaks the site for users using older browsers, that's not good at all. (Yeah, I know... "update your browser"... that's maybe not an option for some people?) Lupo11:55, 8 July 2009 (UTC)
I can confirm this problem. It doesn't seem to be jQuery though. There are no errors on my console at least (perhaps Opera 8 is confused because of my Opera 9 installation). Weird. —TheDJ (talk • contribs) 12:50, 8 July 2009 (UTC)
jQuery was just disabled because it was causing a few more troubles, and wasn't being cached. It might return later, so the bugreport is NOT closed. —TheDJ (talk • contribs) 13:49, 8 July 2009 (UTC)
I don't think it'll work at all. It loads a Script from the Polish WP using importScriptURI. The loaded script then tries to load other scripts at the Polish WP using importScript. That won't work because importScript will look for those scripts here at the English Wikipedia, not at the Polish WP. Lupo10:11, 14 July 2009 (UTC)
English/German Inconsistencies.
User:FHessel has voiced some concerns here regarding the old revision page and the language inconsistencies between the English and German language versions. I'm posting this on his behalf :). Cheers, — DeontalkI'm BACK!11:00, 14 July 2009 (UTC)
Can Google Chrome cope with image maps? With Firefox, Template:Talyllyn map has elements that are clickable (sorry if that's not the right technical terminology) to take you to the corresponding article. But Google Chrome doesn't seem to allow this.--94.196.254.170 (talk) 20:34, 12 July 2009 (UTC)
Works for me using the latest beta version of Chrome. It doesn't show the href at the bottom of the page, though. That makes it a bit difficult to use. --- RockMFR21:50, 12 July 2009 (UTC)
I'm using chrome 3.0.193 and the image map is working properly, we might just have to wait for it to be released as a stable version Redekopmark (talk) 18:41, 14 July 2009 (UTC)
The page [[Folio (text)]] is a redirect to Folio (printing). When you click on "What links here" on the [[Folio (text)]] page, you get a list of articles, many of which involve early printers of Shakespeare, such as Edward Blount, Thomas Creede or Thomas Cotes . But when you go to those pages, the link [[Folio (text)]] does not appear. It previously did appear in the Template:Earlybard which is at the bottom of many of these pages, but I replaced that in the template with Folio (printing) in order to eliminate the redirect.
My question is why does the [[Folio (text)]] page still show pages linking to it that do not contain that link? Is it a question of waiting until the Template:Earlybard somehow is reindexed? Thanks. Ecphora (talk) 01:34, 15 July 2009 (UTC)
They'd have to write their comments as before, but include them before the '</div>'. That said, I don't think the div is a good idea? It looks too blocky and the colour by any regard is too dark. I'd advise lightening the box or removing it altogether. But, in answer to your question, just reply on a newline before '</div>'. Greg Tyler(t • c)23:06, 18 July 2009 (UTC)
Yes, but I am still experimenting with the template. I am just trying to see what is possible, and what would be not practical. Thanks Greg! Kind regards, LouriePieterse10:02, 19 July 2009 (UTC)
Stub link formatting threshold is not saved in preferences
Hi, I noticed recently that the stub link formatting threshold option in the preferences no longer seems to work. No matter what I change the value to, it always resets to 0 bytes when I hit save. For the record, I'm using Firefox 3.5. – TMF06:40, 15 July 2009 (UTC)
I'm trying to delete some images and keep getting database timeouts and other errors. Also, many images aren't loading for me. Anyone else having this problem? Stifle (talk) 13:22, 15 July 2009 (UTC)
Due to an operating system bug in the Solaris servers (and ZFS filesystem) we're using for image storage, we're having continual problems with our image servers. In essence, when the disk gets close to full, I/O performance is significantly degraded, resulting in load problems. We're trying to mitigate this by moving some of the data onto other servers (we were intending to simply start filling other servers instead) and we've contacted Sun to try and get things sorted out. In the meantime, images will be slow in high load periods (generally when the Europe and American peaks coincide, which is around midday to 9pm UTC). — Werdna • talk13:33, 15 July 2009 (UTC)
Looks like ms1 is still ill. That graph doesn't look good (system CPU constantly rising, and network throughput dropping). Though it's a different symptom than before (previously, both went up when there were problems).Lupo13:48, 15 July 2009 (UTC)
View changes since last visit
Is it possible to view only the history of articles for changes since you've last viewed the page? Thanks. SharkD (talk) 03:35, 15 July 2009 (UTC)
I don't think you can. I know you can view *before* a certain date using the &offset=20071226164055 parameter which includes a timestamp in what I think is YYYYMMDDHHMMSS. Not sure if that helps, but maybe have a bit of a play until a smarted editor than I comes along :). Cheers, — DeontalkI'm BACK!05:06, 15 July 2009 (UTC)
I would've recommended that, but the user wants to see the edits since they last viewed the page, not edited it. Slightly more difficult eh? — DeontalkI'm BACK!08:16, 15 July 2009 (UTC)
To do that, you'd need to store the time of the visit to every page on the Wiki, and if a script, do that for lots of people, which would be totally impractical. So to answer your question, no, not really. Ale_Jrbtalk08:46, 15 July 2009 (UTC)
This might be related to what you are trying to do: If you are looking at a diff, Twinkle adds a "current" tab that will show the diff from the shown diff to the current revision. This is useful if you try to keep up with a fast moving discussion since you can move in larger steps than one diff at a time, but you need to keep the page open of course, nobody is remembering for you which revision you last looked at. One might be able to build some script that stores the last revision in a client-side cookie. Amalthea10:56, 15 July 2009 (UTC)
In smaller WMF projects such as Wikimedia Commons, it's possible to view a diff of changes since your last visit to the page. This feature is probably disabled here because of server load. Graham8716:56, 15 July 2009 (UTC)
extreme slowness?
We're encountering exciting new issues with the upload fileserver which cause intermittent slowdowns on the site; see tech blog notes.
Pretty much every page is taking 1-2 minutes to load for me today instead of the usual <5 seconds. Anyone else experiencing problems and/or know what is up? --ThaddeusB (talk) 15:06, 11 July 2009 (UTC)
The IRC welcome message says: "Wikipedia is currently experiencing some server errors, and is slow or down for most people. This is being worked on and will be fixed as soon as possible. We apologise for the inconvenience this is causing and thank you for your patience." I can confirm this. HansAdler15:52, 11 July 2009 (UTC)
I was getting the same thing (lag). Seems much improved now, although (possible unrelated) the main picture isn't showing up. - Kingpin13 (talk) 16:03, 11 July 2009 (UTC)
I've been going on other web sites too during the past hour, but they've all been slow. I got a message that I needed to download new software. So ever since I clicked on that, everything is like dial-up access must be. Sites are slower when I go to them for the first time.
Although the file image seems to display lately, the Wikipedia logo globe is still disabled; at times I see an error message which I finally managed to capture. Here it is:
Proxy Error
The proxy server received an invalid response from an upstream server.
The proxy server could not handle the request GET /wikipedia/en/wiki/Wikipedia:Village_pump_(technical).
Yes I've experienced it too. The last few hours have been dreadul I'll click en.wiki and my modem will go crazy and nothing will happen,. ERvery other website is fine. I think the size and numbe rof users is placing serious strains on the server. I think an upgrad eis needed (not to mention a front page resign), its not 1996 after all. Dr. BlofeldWhite cat16:21, 11 July 2009 (UTC)
I for one have to agree. The slowdowns have been getting worse the past 6 months or so. For a top 10 website it's almost inexcusable. Not only does it prevent people from reading it also makes editors give up in disgust. --Brad (talk) 18:59, 11 July 2009 (UTC)
You're welcome to either live with it, or donate us some money. While you were complaining, I've fixed it by disabling CentralNotice for now. — Werdna • talk19:19, 11 July 2009 (UTC)
Thanks, does seem much better know. I agree - if people complain they should be donating $ or £ to help improve the hardware. Nothing comes for free. Smartse (talk) 19:31, 11 July 2009 (UTC)
It's still noticably slower today. Don't give me that BS about donating - I donate by improving the site's content. If anything, you should be paying me! Lugnuts (talk) 09:02, 12 July 2009 (UTC)
The speed still isn't improving in English Wikipedia, where other Wikimedia site works fine. The Wikipedia logo on the corner and some images are loading very slowly. Can anyone explain the exact cause of the slowness here? The second time I encounter a problem in short period of time (see this page for another error I encountered before), looks like I got to donate to Wikipedia. --98.154.26.247 (talk) 20:04, 11 July 2009 (UTC)
It got better around 16:00 GMT time but I totally agree with Brad, the server problem is getting worse and worse over the last few months. A lot of the time I blame my own computer but it seems that it affects others too so has to be the server. It is not good enough I agree.Dr. BlofeldWhite cat20:14, 11 July 2009 (UTC)
From the devs: "ms1 started showing signs of overload at the daily peak; nobody cares, two weeks elapses, overload becomes critical, site goes down". In english: bad bada boom. There is a partial spotty echo of the log at twitter and identi.ca. --Splarka (rant) 07:46, 12 July 2009 (UTC) -- P.S., server admin log working again. 08:12, 12 July 2009 (UTC)
Maybe we should start asking where exactly is the donation money being spent on. The slowdowns may be an indication that something is going wrong in this respect. Server operations should be priority nr. 1. Offliner (talk) 18:36, 12 July 2009 (UTC)
"Donate more money or live with it" is a really shitty response to legitimate technical problems and I hope not to see it posted here again. --MZMcBride (talk) 20:41, 12 July 2009 (UTC)
The sysadmins (specifically domas) think they might have found the cause for the problem, and it is possibly the same problem as this one (quite technical). A few snapshots have been dropped and it seems that performance is improving again, though that is a temporary solution. Likely problems will persist, but with a bit of luck it should be considerably less than in the past two days. —TheDJ (talk • contribs) 20:52, 12 July 2009 (UTC)
There are indications that the problem is already returning. Most sysadmins are asleep, so it might be a while again before performance will return. —TheDJ (talk • contribs) 23:32, 12 July 2009 (UTC)
All this back and forth about donations implies that this is merely a chronic high usage problem, but what's happening today is unusual. It's really slow today in a way that it wasn't slow before. I can't see the logo either, and I don't remember that happening before. Server load isn't higher than usual, is it? Fewer people use Wikipedia on the weekends, if I'm not mistaken (or is Sunday night different?), and Michael Jackson's death caused a big spike in usage but didn't cause problems like these. 146.151.21.117 (talk) 00:43, 13 July 2009 (UTC)
We're encountering exciting new issues with the upload fileserver which cause intermittent slowdowns on the site; see tech blog notes.
I just posted a comment on this blog entry, and I will echo my request here. For those with the rights to do so, please consider posting a "We are aware of this problem and are working on it" type message in a prominent location in all wikis affected by the slowness. Whether this is a watchlist message for logged-in users, or something at the top of every page, I think it would go a long way towards helping with user frustration. I think a lot of people are confused by the symptoms and aren't sure whether it's something with their connection/proxy or on the wikimedia end. More information is always helpful. Thanks --Dfred (talk) 15:37, 13 July 2009 (UTC)
As to the difference between contributions and donations, one of the biggest criticisms of highly compensated individuals who operate non-profits is when they fail to contribute to their own non-profit vs. the typical religious figure who assiduously tithes even though all the money is just being recycled - they pay him/her X, he/she contributes X/10, lives on 0.9X, as an example of do what I do, and what I say, or do what I say, not what I do or some variation thereof. It is often noted that if everyone chipped in a € or a ¥ there would be plenty of money to buy new servers, and if everyone waits for someone else to chip in there will be none. I would highly recommend both contributing and donating, and not pretending that your contributions in the form of edits are enough to make up for a lack of donations. I suspect that small donations of a penny would work too, except that Paypal will just say thank you very much for anything under about 30 cents, and the foundation will get nothing, so $1, 1€, etc. is probably the smallest practical donation: "If you haven't got a penny, a hay penny will do, and if you haven't got a hay penny, God bless you." Apteva (talk) 15:36, 13 July 2009 (UTC)
Actually, everyone has to determine their own budget, that's all. No one is asking anyone to pay for everything, only the items that are important to them. Without anyone supporting them they would disappear. For example, you could pick the ten causes most important to you and donate $1 to each once a year. For most people that wouldn't be a stretch. Each of us get to choose which ones survive. Fortunately Wikipedia, so far, at least, is one of those, although, of course since everything is CC/GFDL, if no one did support Wikipedia, the content would still survive. Apteva (talk) 22:49, 13 July 2009 (UTC)
I'm going to copy/paste MZMcBride's response, used above:
"Donate more money or live with it" is a really shitty response to legitimate technical problems and I hope not to see it posted here again.
The "missing logo" is a serious issue and should not be a problem relating to server loading -- if it is, it's a bug. The way I check my watchlist means I open multiple tabs in quick succession, and none of them are finishing loading -- I have a screenful of rotating 'downloading' indicators and a cursor which is 'waiting' 90% of the time. Editing pages is much harder as a result. Affects Firefox 3.0.11, but NOT IE7 (which I have opened using IEtab in a FF tab). How about disabling the icon/thing for Firefox users until the bug is fixed? -- EdJogg (talk) 21:48, 14 July 2009 (UTC)
I am using Firefox 3.5 and have encountered the same errors, but the logo is working for the moment. It's anyone's guess as to when that will stop. —Animum (talk) 02:38, 15 July 2009 (UTC)
I have Adblocked all three logos and now am running nice and smooth. Someone please poke me when its fixed? –xenotalk 03:19, 15 July 2009 (UTC) I lied. still hanging issues. –xenotalk03:47, 15 July 2009 (UTC)
The failure to load the full page is a major problem as the pop-ups are not active until the page is fully loaded. It is often quicker, at the moment, to open the required page rather than wait for the pop-ups to work. I also have a problem of the editing tool bar not appearing until after the logo is loaded and often I can complete the required edits before the page is fully loaded. It was OK earlier today but now it is taking several minutes to fully load a page. Keith D (talk) 19:47, 15 July 2009 (UTC)
Page number citation template
Can anyone tell me what the page number citation template is? I recall there being something like "Jane saw spot run.[3][p. 41]". I think I remember there being a template like that, but I can't find it. Thanks, Oreo Priesttalk19:36, 11 July 2009 (UTC)
This was a side effect of our temporary offline-images situation (see section above). Should be ok for now, I've temporarily disabled the fancy images on the captcha checks. --brion (talk) 20:43, 15 July 2009 (UTC)
I am sure it is an add-in; any way to look and see what add-in the file is using? Also, I tested and it does the same thing in IE 8. UnitedStatesian (talk) 18:24, 15 July 2009 (UTC)
The webkit developers say that this is an especially bad case because: "I bet if we looked at the GIF we'd find that every single frame is "complete", So you have to decode all the pixels in it instead of only a small set that overlays the previous frame. Combine that with big GIF + short frame times and you get heavy CPU usage. I'd be surprised if any GIF decoder can do this properly." So the advice is to get that thing into a saner format (where not every frame is entirely decoded from scratch), or at least make it smaller. —TheDJ (talk • contribs) 18:32, 15 July 2009 (UTC)
Actually, the webkit devs also said the both the Mac and the Chromium GIF decoder could probably significantly be improved, but that they find it more likely that they will implement an APNG decoder instead. —TheDJ (talk • contribs) 22:12, 15 July 2009 (UTC)
Just curious, how could this be coded so the text is in front the image without doing <tr><td> tables? Or if you do use <tr><td> can you do so without set height and width parameters?-- penubag (talk) 07:24, 19 July 2009 (UTC)
Did it for you. You close your divs properly and add another with style="position:relative;", but please avoid using this hack in articles. Also, it probably wont work IE6. — Dispenser19:38, 19 July 2009 (UTC)
After discussion at Wikipedia:Help desk#Images, I was directed here. On the Meridian, Mississippi article, File:Bonita Lakes 1.JPG's thumbnail will not show up at its current size. If I change the size of the thumbnail (by forcing a size - my default is 250px), the image shows up. I've tried purging the page and bypassing my cache already, but the image still won't display. Any thoughts as to what could be causing this problem? I know thumb generation was down earlier, and I was editing the page during that time (the edits included changing the size of the thumb I think), so maybe something went wrong with that one specific thumbnail size. --Dudemanfellabra (talk) 08:02, 16 July 2009 (UTC)
I presume you know what autoconfirmed is? Well confirmed has just come in, and can "jump start" this, as autoconfirmed is auto granted after 10 edits and four days, whereas confirmed is the same, but is manually granted, so that the "extra" stuff which autoconfirmed gives (e.g. Moving pages, uploading images, editing semi-protected pages) can be given without the user having to wait. - Kingpin13 (talk) 12:03, 16 July 2009 (UTC)
I know it's not standard practice on Wikipedia but is it possible to make an image not clickable? so when I click on an image, it doesn't follow up to an image summary or full screen image. Something like [[file:imagename.jpg|link=none]]-- penubag (talk) 10:08, 16 July 2009 (UTC)
[[Image:Smiley.svg|30px|link=]] :
This should only be used for public domain and similar licenses where one is not required to provide attribution or license information, since removing the link generally prevents those functions. Dragons flight (talk) 10:19, 16 July 2009 (UTC)
I have made some changes to how the Search engine selector is added to the search page. If people make sure they have refreshed their browser cache, they may see this change. The selector is now only visible in the Advanced mode of the Search page. I figured this might be a good idea in light of the usability redesign. I realize some might think this is a bad idea, and I have no objections if we change it back, but I would like it if people report what they think of it. I think it's a better design especially for some of the newbies.
Do you think it should always be visible ?
Do you think that if it is not always visible, this position is better ?
PS note that the searchengine selector isn't working properly atm due to bugzilla:19747
Advanced search options should always be visible. Anyone doing a simple search will leave the settings as they are. Anyone who wants to do a more advanced search doesn't want to have to go to another page or make more clicks. I think you should revert this change until you have more input. Delicious carbuncle (talk) 18:39, 15 July 2009 (UTC)
For clarity, anyone who is logged in will always be in the "Advanced search" mode. Only IP users default to "Content pages", so for most people nothing is changing. —TheDJ (talk • contribs) 18:48, 15 July 2009 (UTC)
That's not happening for me. I have the option of going to the advanced search by clicking on a link, but I just tried it a moment ago and what you describe isn't the case for this logged in user. Delicious carbuncle (talk) 19:50, 15 July 2009 (UTC)
I'm using the search box. Can this change be reverted while you sort out what is going wrong and gather some feedback on what users might or might not want? Thanks. Delicious carbuncle (talk) 21:34, 15 July 2009 (UTC)
And I just ran this by Trevor Parscal of the Usability team, and he was very happy with this change. It was one of the points they still meant to address in the future. —TheDJ (talk • contribs) 22:07, 15 July 2009 (UTC)
DJ, you implemented a change that no one appears to have asked for and then asked for feedback. My feedback is that your change doesn't work as you thought it did (which suggests you haven't tested it properly) and reduces the functionality of search without further action from the user. I don't mean to be rude, but I don't think that unilaterally running a one-day "experiment" and posting about it here after you've made the change is a responsible way to develop one of the world's most used website. Your comment about the "usability team" approval further implies that you aren't concerned about actually finding out what users want or actually use the search box to do. Can you give me Mediawiki access so that I can start hacking away, too? Delicious carbuncle (talk) 22:41, 15 July 2009 (UTC)
The search engine selector was created because MediaWiki's search engine used to, well, suck. We now have a much better search engine and not only is the selector no longer needed, it overcomplicates the user interface.
Language preference applied to translations on Commons
Call this a pet peeve, but some Commons images are so multi-lingual that they present an awkward wall of translated text. File:Anime_Girl.svg is one example, though far from the worst.
I know that one can use CSS to totally kill the foreign languages (and thus make one's individual life easier), but that's an awkward general solution since different people may be looking for different languages.
I'm wondering if someone could create a bit of javascript that would reorder such multi-lingual blocks so that a user's interface language preference was positioned top most and other languages were placed in a collapsed box (or something similar). In general, I looking for a way to highlight the text a person is most likely to be looking for without totally removing access to the rest. Dragons flight (talk) 15:46, 16 July 2009 (UTC)
No, the last instance of anything is used; it's a cascading effect. If there were two title fields, the second would be used. EVula// talk // ☯ //21:30, 20 July 2009 (UTC)
I see two images above. Actually, there should be three- the left hand image is [[Image:Geastrum saccatum.jpg|324px]], the right hand is [[Image:Geastrum saccatum.jpg|326px]], but, in the middle, there is [[Image:Geastrum saccatum.jpg|325px]]- showing as nothing. Why is this happening? I need the image to show at that size for Portal:Fungi/Selected picture/21- all of the other portal pictures are that size, as that is the size the template uses. Why is this happening? J Milburn (talk) 22:45, 20 July 2009 (UTC)
Now that IE 6 and 7 often render HTML, XHTML tags and CSS syntax incorrectly, and formal version of IE 8 was released this year, we'd better not consider page rendering in IE 6 and 7, just consider rendering on IE 8, Mozilla Firefox 3.x, Opera 9 & 10, Safari 4 and Google Chrome 2 & 3, thus we can only consider W3C standards compliance and implement newer webpage technology. For IE 7 or earlier users, we can use message or redirect to prompt them to upgrade to IE 8 or switch to other browsers, just like Youtube, Facebook or Twitter.--RekishiEJ (talk) 10:38, 16 July 2009 (UTC)
Mediawiki strives to be accessible in IE7 and 8 and to a large part still IE6. I think the general trend has always been to support them until it becomes cumbersome to do so. As long as IE6 still shows _more or less_ ok, there's no reason to throw warnings at people telling them to upgrade. Now, will we go out of our way to be backward compatible with old browsers? No. That's why we don't support IE5 and below, it's simply too much effort for not enough return. All this goes for non-IE browsers as well. ^demon[omg plz]12:06, 15 July 2009 (UTC)
We cannot force users to upgrade, and many are locked into their current browser. Many libraries, schools, universities and companies use PCs that are locked down. ---— Gadget850 (Ed)talk13:16, 15 July 2009 (UTC)
Well, I do not want Wikipedia to be best seen in Firefox, neither. I just mean that we can drop IE 6 and 7 support to ease maintainment (no longer requires CSS hack). Some notable commercial and individual sites have dropped support for older IE for their incorrect CSS rendering. For libraries, schools, universities and companies, we can ask them to upgrade to the newest version. --RekishiEJ (talk) 23:52, 15 July 2009 (UTC)
Digg is thinking about dropping support for IE6, but I think we should continue to support it if we can. IE6 has more users than any of the browsers besides the top 2 or 3, I think. - Peregrine Fisher (talk) (contribs) 23:59, 15 July 2009 (UTC)
IE 6 still has about a 15% traffic share, and IE 7 about 20%. While it is okay if they aren't pixel perfect, I'm sure we will continue to make sure they render consistently and well, even if that means including hacks and what-not. Dragons flight (talk) 09:43, 16 July 2009 (UTC)
We should try to use warning message to discourage the use of older browsers, hence each user updates their browsers to the newest, enjoys the new features and avoids security concerns since browser vendors usually don't release patches for older versions. Microsoft should also think why some net surfers degrades their IEs to 6 or 7, since IE 8 is slower and less stable than IE 6 and 7......--RekishiEJ (talk) 10:38, 16 July 2009 (UTC)
It's not our job to tell people what browsers they ought to be using. Many readers are in corporate or school environments where they don't have a choice anyway. Rather we should pay attention to what browsers have a significant traffic share and make sure we support those well. Dragons flight (talk) 12:46, 16 July 2009 (UTC)
Also, we can use the time and effort doing CSS hacks to debug and add new funtions to current web applications.--RekishiEJ (talk) 11:31, 16 July 2009 (UTC)
It's up to the user to investigate if scripts they install have any malicious code. Of course practically no one does; it's only a matter of time before someone installs a seemingly harmless script and has their account compromised. That said the risk is pretty low IMO, just keep an eye out for it when installing new scripts. --Chris09:31, 16 July 2009 (UTC)
Without going into specifics, is it actually possible to get your password with a malicious script? Oh, and I always grab scripts that I see other people using, which provides, I would hope, a little security in these matters. The scripts listed on the gadgets section of your preferences are rigorously tested, but nothing stops admins adding them before this testing has taken place. - Jarry1250[ humorous – discuss ]10:29, 16 July 2009 (UTC)Bold text
Any script in MediaWiki namespace (including gadgets) can only be changed by administrators. So any malicious code would have to be added by an admin and the addition would have to be unnoticed by other admins. Then this admin needs to entice other users to install this code. All in all, it's not too likely. Scripts in the User namespace can be edited by admins AND the user who has the code in his namespace. So this user can be a freshly registered user, but again, he would need to entice other users into running his javascript. Quite hard to do, without going unnoticed by other users. And then there is step of doing something useful after you have installed the malicious javascript code. Doing real damage, often depends on the user running a compromisable browser. So all in all everything is possible, but damage would likely be limited to a few users at most. And the blocks would be swift and indefinite. —TheDJ (talk • contribs) 11:58, 16 July 2009 (UTC)
It's always safer to use a gadget or make your own than to trust some code someone's sent you. Scripts can be malicious and its up to you to check them. If you're not sure about one, it's best not to bother and continue manually. As for twinkle itself, I've never come across malicious claims about it and have made good, safe use of it for a fair while now. Greg Tyler(t • c)16:15, 16 July 2009 (UTC)
To answer your question, Jarry, user/site JS is disabled on Special:UserLogin and Special:ChangePassword, so it shouldn't be possible for scripts to steal your account password. However, it would be trivial for any script to steal your session cookies, which would allow an attacker to gain control of your account for as long as you remain logged-in. Happy‑melon17:21, 16 July 2009 (UTC)
But with malicious javascript, you could probably just change all links to those pages to slightly misspelled versions of the same page, and before load send an ajax request to the actual page, and delay page loading until the ajax request returned. On return, you'd then replace the contents of the false Special: page with the remote form and appropriate <title> (etc), practically identical to the regular Special page except with the victim's user JS still loading. When they submit the form, you could then send an off-server request (like via <img>) containing the username and new password, and still allow the form to submit as normal, giving them little reason to suspect anything amiss. Probably several administrators log in to various alternate accounts while still logged in as administrator. WP:BEANS yes, but people need to look out for this sort of thing, not be kept too ignorant. I could probably stick something into User:Splarka/sysopdectector.js or User:Splarka/dabfinder.js without too much notice, if it was obfuscated enough (I wouldn't though!). --Splarka (rant) 07:48, 17 July 2009 (UTC)
@ TheDJ, it would be trivial for an admin to compromise another users account (I'm pretty sure there are other messages that still use raw HTML as well). As for a new user tricking people into installing scripts it wouldn't be that hard for a person who's had some practice at social engineering. Although personally I think most people have better things to do than hack Wikipedia. --Chris06:58, 17 July 2009 (UTC)
mediawiki API & encoding
Hi all
I hope it is the right place to ask this:
I'm using the mediawiki API to update some pages with an experimental java program. This robot uses the java apache http-client library to update the pages.
(...)
PostMethod postMethod = new PostMethod("http://mymediawikiinstallation/w/api.php");
postMethod.addParameter("action","edit");
postMethod.addParameter("title",page.replace(' ', '_'));
postMethod.addParameter("summary","trying to fix this accent problem");
postMethod.addParameter("text",content);
postMethod.addParameter("basetimestamp",basetimestamp);
postMethod.addParameter("starttimestamp",starttimestamp);
postMethod.addParameter("token",token);
postMethod.addParameter("notminor","");
postMethod.addParameter("format","xml");
int status = httpClient.executeMethod(postMethod);
(...)
however the 'content' String contains some accents. System.out.prinln(content) looks ok, but the accentuated characters in the wiki look bad. E.g. 'Val�rie' instead of 'Valérie'.
Make sure your HTTP client library is using UTF-8 encoding; it sounds like it's trying to encode to 8-bit Latin-1 (ISO 8859-1 or Windows-1252). --brion (talk) 21:04, 16 July 2009 (UTC)
Okay, not really getting the hang of it. I wonder, if I can request you or someone else to implement the following: I think we should have a dropdown like MediaWiki:Deletereason-dropdown on the nuke page as well but I cannot figure out how to put one there. It'd be great if someone could take a look and add it. Regards SoWhy12:48, 16 July 2009 (UTC)
For the record, Amalthea, even setting the limit in the new paginated version to 99999, and thereby still getting the complete list, is a significantly smaller pagesize than the current version, so it's still an improvement. Plus it doesn't look totally horrible, and doesn't use snail-pace javascript to fake-filter the results... :DHappy‑melon14:50, 16 July 2009 (UTC)
Yeah, I figured that getting the complete list somehow would still be possible. And I didn't look at the specifics of the change, so if it generally reduces the page size of that moloch then that's a Very Good Thing. :) Amalthea12:09, 17 July 2009 (UTC)
Edit button position
There's a preference to enable an [edit] button to appear to enable editing of the lead only. There's also a preference to enable the [edit] button to appear immedately to the right of a section heading, instead of at the right side of the page.
Can we get these two together so that the lead edit button also appears immediately to the right of the header text instead of the far right of the page please? Mjroots (talk) 12:07, 17 July 2009 (UTC)
Hmm... If you go to WP:UPLOAD (not Special:Upload), then the first option is "Entirely my own work - I created it, own all the rights to it, and have not used anyone else's work in making it", this takes you to Special:Upload, with the preload "~~~" for author, which gives the user's signature, without date. This indicates the user as the author of that image. Is that what you wanted? - Kingpin13 (talk) 12:44, 17 July 2009 (UTC)
I just added the signature to the source as well (which I think is what Sfan wanted). But perhaps a subst:REVISIONUSER would be better here. — Martin (MSGJ · talk) 12:50, 17 July 2009 (UTC)
Good idea, we don't need people's superfluous signatures in this case. {{subst:REVISIONUSER}} would seem to be a better approach - Kingpin13 (talk) 12:55, 17 July 2009 (UTC)
The reason for this request, as stated was that 'self-made','own-work' etc get confusing hen stuff get's moved to commons.
A while back I raised the question of whether there was a way to have footnotes positioned directly under the relevant paragraph, whilst maintaining the ongoing numbering system throughout the article to avoid duplication.I understand that having reference duplication is currently picked up as an error automatically so it would need to get round that. The advantages are 1) it means you can meaningfully and usefully include short quotes in the references and 2) sometimes references are the surest indication of the general approach within an article. —Preceding unsigned comment added by Dungur (talk • contribs) 16:54, 17 July 2009 (UTC)
Sure you can put quotes in a footnote at the bottom of the page but then there's a greater visual and functional disconnection between what's being said in the quote and what's being said in the original text. What I'm getting it with (2) is that not all reference sources are of the same quality: furthermore, they may imply a specific approach to a subject, and knowing what that approach is may be an important bit of information for the reader who's hip to the references. Dungur (talk) 19:36, 17 July 2009 (UTC)
Are you are asking for a way to have a references section under each paragraph? This can be done, but there are issues that I will discuss if this is the correct assumption. ---— Gadget850 (Ed)talk12:11, 18 July 2009 (UTC)
Problem in scaled-down SVG images
The picture in Quark#Classification suddenly stopped showing, but if the |300px parameter is removed it does show. What on Earth is going on?
Error creating thumbnail: Error saving to file: /mnt/upload5/wikipedia/commons/thumb/0/00/Standard_Model_of_Elementary_Particles.svg/300px-Standard_Model_of_Elementary_Particles.svg.png Most probably the same issue as mentioned in #extreme slowness? and other sections above, I'm afraid. Amalthea09:58, 18 July 2009 (UTC)
Please report these in the irc channel #wikimedia-tech and a developer will check into it (perhaps with some delay). We just fixed a set of permissions issues; hopefully no new ones will crop up. -- ArielGlenn (talk) 15:55, 18 July 2009 (UTC)
CAPTCHA curiousity
When I made this change to News presenter (and retested here), I noticed something that might be an unintended feature of the arithmetic-based challenge-response authentication I've recently been encountering. If the sum/difference I'm asked to calculate is of the form , I can reply with and it will consider the challenge to have been met successfully (i.e., I can reply that instead of and my edit is still accepted). I don't know what the odds are that users will be asked to subtract zero from a number, but I would think it's a loophole that automated spammers could exploit occasionally. 68.167.254.171 (talk) 04:37, 18 July 2009 (UTC)
The image captcha was better. With these add/subtract ones they can just write a script to add (or subtract) the two numbers and input the result into the field. 59.95.105.187 (talk) 04:43, 18 July 2009 (UTC)
Just wondering, would 5 - 1 = 5 - 1 work too? If so, that would really be a crappy captcha. MER-C03:41, 19 July 2009 (UTC)
No. The issue is that "(int)$variable" in PHP converts the string in $variable into an integer by removing leading whitespace and then using all the digits up to the first non-digit character. So for "6 - 0" it sees the space after the 6 and stops there; for "5 - 1" it would return the integer "5", which doesn't match the expected answer. The bug here exists for equations of the form "x - 0" and "x + 0"; there are also two related bugs:
Anything not beginning with a digit will be accepted for "0 - 0" and "0 + 0".
Not serious, but for example "4", "4?", "4 + 1", "4 million", and so on are all accepted as answers to "5 - 1".
I would recommend everyone logging out and putting an external link in a sandbox just to see the new captcha. In my opinion "57 + 8 =" is hardly a trivial calculation and I suspect would stump a third of the population. I would recommend going back to a more traditional captcha or participating in the project to decipher scanned books. I'm sure someone knows what I am talking about. 199.125.109.81 (talk) 14:48, 19 July 2009 (UTC)
Hi. I'm left handed and I've only noticed that cologne blue and classic or one other has the option to change the quickbar from left to right and fixed or moveable. Given that I mostly use modern and occasionally mono, would it be possible to be able to add the quickbar option for modern so the navigation bar can be moved to the right hand side. I think you should give people this option for all skins givne that some people are left handed and might find it more natural to have it on the otherside. I also think there should be a shrinkable option on the nav bars like we have with big templates for instance as if we could conveniently shrink the task while reading or at least have the option to this would ease useability I think. Also for the task bar on modern does it have to be so wide? There is at least a centimetre gap, it could be trimmed easily by 1-2 centimetres.Dr. BlofeldWhite cat13:41, 19 July 2009 (UTC)
http://bugzilla.wikimedia.org for bugreports and feature requests for the software, incl. skins. And knowing how stuff like this usually goes, patches are welcome :D 15:37, 19 July 2009 (UTC)
OK, I and apergos were kinda stumped as to why this was broken, but with a bit more brainstorming, it suddenly dawned on me.... GIF thumbing is disabled atm due to issues with animated gifs. Probably someone, or somehow the old thumb for this gif image was purged, and now the image is not regenerated.... This is rather a major pain, because those symbols are in many CSS files.. —TheDJ (talk • contribs) 15:54, 19 July 2009 (UTC)
I forgot about the gif thumbnailing, that's certainly the issue; probably the old cached thumbnail was removed during the image server fixing. Is there any reason not to convert these to png? It would need a Commons admin to protect the image there, and an enwiki admin to change our css files referencing that image. I'm neither, so I can't do it. Anomie⚔16:30, 19 July 2009 (UTC)
I was delighted recently to see that Wikipedia, when viewed on mobile browsers like my iPhone, redirected to a much more tractable mobile version, http://en.m.wikipedia.org/wiki::Home . I wasn't so delighted this morning to find that apparently all pages requested from http://en.m.wikipedia.org/ give a 404 Not Found from "nginx/0.6.37". The problem is that explicitly trying to load http://en.wikipedia.org redirects to this mobile site, so as it stands, Wikipedia is pretty much entirely unusable from my iPhone. What happened, and can we expect a fix soon? SeanWillard (talk) 15:00, 19 July 2009 (UTC)
Brion created a "wmf-deployment" branch to publish the exact code currently running on Wikimedia wikis (including live hacks and the like). But that means version numbers in Special:Version no longer match those on trunk. Anomie⚔02:43, 20 July 2009 (UTC)
Yes, this is a feature of the sysadmin global group: editinterface on all wikis. As opposed to old edits like this: nowhere is it recorded how Domas got his sysop bit for that edit, becuase he wrote it directly into the database tables and then removed it after making the edit. Which isn't very 'wiki'. Happy‑melon11:07, 19 July 2009 (UTC)
However, as one person noticed when testing it, if the user works quickly enough. Some anti-vandalism channels get spammed with image uploads, something that's not all that desirable. Aside from things like bot accounts, which aren't really necessary (in terms of technical limitations) and not desirable for such a tool, what can be done?
The mobile skin does not work properly on my iPod Touch. It has iPhone OS version 1.15, and I cannot expand the headers as this requires javascript which is not supported by the OS. I don't know if it works in later OS versions (like 2.0 or 3.0) but I know it does not work with 1.0 and I don't want to pay $US 9.99 to upgrade my OS. While I don't use the browser that much on my iPod Touch anyway, I just thought I'd let you guys know.--Richard (Talk - Contribs) 18:32, 20 July 2009 (UTC)
I have looked into this, and with Safari 2 I was able to find a problem in the Javascript code that is used when using this older browser version. I have created a fix for this and submitted it to the maintainer of the mobile website. Hopefully it will be deployed soon. —TheDJ (talk • contribs) 19:48, 20 July 2009 (UTC)
Any idea why this edit to Jackson County, Iowa set off the abuse filter as "Possible Michael Jackson vandalism"? Obviously it's Jackson County, and the change involves adding a link to http://www.jacksoncountyiowa.com, but I can't see why adding a short string of text with a reference could be construed as vandalism, unless the abuse filter isn't programmed the best. Nyttend (talk) 03:50, 21 July 2009 (UTC)
Yes, it's a spacing issue. The initial letter of an article page is always capitalized, but thru some template magic (the {{lowercase}} template), it displays as lower case in the article title. However, wikilinks don't care about initial capitalization; a wikilink to myNetworkTV is identical in every way to MyNetworkTV. The page myNetwork TV just redirects to MyNetworkTV (with no space). --Floquenbeam (talk) 19:08, 20 July 2009 (UTC)
Hmm? That isn't usually very hard. You just change the redirects so they point to the right place. There's no need to edit the articles that use the redirects at all. --Trovatore (talk) 21:07, 20 July 2009 (UTC)
Yeah, it's not a particularly bad thing if there are redirects being used. What's more important to check is that there are no redirects going to a redirect; double-redirects break. EVula// talk // ☯ //21:29, 20 July 2009 (UTC)
X<span style="text-decoration:underline"> foo </span>X → X foo X
If the spaces are underlined, Tidy is off, if they aren't, Tidy is on and has pulled the spaces outside the span. And yes, I consider this a bug, Tidy has no business making visible changes if the source code wasn't invalid. That's at least true for the Tidy settings used here on Wikipedia. Amalthea15:56, 21 July 2009 (UTC)
Recently I have been using Google Chrome 2.0.172.37 to read English Wikipedia articles, however I found out that search box can not be used while using Modern skin (searcj box can be used while using other skins). Also the text in the edit box can be in drag-and-drop; Ctrl+drag-and-drop results like solely drag-and-drop, though. Why does this bug occur? The operating system I'm using is Windows XP Home Edition SP2.--RekishiEJ (talk) 17:51, 21 July 2009 (UTC)
This is a fairly minor glitch, but I've noticed on my watchlist that after clicking "Hide my edits", the name will change to "Show my edits" without actually changing the parameters in the link to showOwn instead of hideOwn. A quick refresh of the list solves the problem, but I though I'd bring it up. — Trust not the Penguin (T | C) 22:50, 21 July 2009 (UTC)
Do we have a list to see which redirects or search terms that don't already have an article get the most views? I might like to create an article for some of these, as a service to our readers. - Peregrine Fisher (talk) (contribs) 04:18, 22 July 2009 (UTC)
To take the second question first, it's hard to distinguish between a search that results in a list (to be selected from) because an article doesn't exist, versus a list that results from a misspelling or an unusual combination of words (for example, Sarah Palin scandal or, for example, Barack Obama birth certificate) And even if there is no misspelling, the problem may well be lack of a redirect, not lack of an article. (Having said that, I know of no list relevant to the question.)
The first question is easier: Category:Redirects with possibilities provides a good starting point. Someone would need to take pages in that category and compile a table, using the page view statistics available from the history tab. That's not particularly hard, I think - see, for example, the third column in the section here: Wikipedia:Academy/NIH 2009/Welcome#WikiProjects. Of course, all that presumes that folks creating redirects do a reasonably good job of categorizing them.
This edit[3] makes me wonder, does DISPLAYTITLE actually work? If it does, what was wrong with the edit? In my own experiments in a sandbox I was unable to make DISPLAYTITLE work even if I copied it literally from another page. — EmilJ.12:34, 22 July 2009 (UTC)
DISPLAYTITLE only works if the title produced could be copied into a page, surrounded with square brackets, and form a wikilnk which would still lead to the original page. So DISPLAYTITLEs almost certainly won't work when you copy them to another page. (also)Happy‑melon12:52, 22 July 2009 (UTC)
(e/c) This is not what happens, from what I can see. On the one hand, [[Apple IIGS]] does not make a valid link to Apple IIGS, yet the title change works there. On the other hand, ε-inductionis a valid link to epsilon-induction, nevertheless Carl's attempt to make it a DISPLAYTITLE failed. So what is the actual rule? — EmilJ.13:01, 22 July 2009 (UTC)
It's not the contents of the DISPLAYTITLE tag that has to be a valid link, it's the final output of the title on the page. After the conversion has taken place, the title displays as "Apple IIGS". If you copy-and-paste that text (from the rendered view, not from the edit window), you get a valid link to the original page. (also)Happy‑melon13:10, 22 July 2009 (UTC)
(ec) [[Apple {{sc|II|GS}}]] doesn't link properly because it expands to "[[Apple <span style="font-variant:small-caps;">II</span><span style="font-size:80%;" >GS</span>]]", and that string has characters that aren't valid in links. However, the text representation of that output (i.e. all tags removed) is still "Apple IIGS"; you can still triple-click the title, copy&paste it, and build a wikilink to the page from it. Amalthea13:13, 22 July 2009 (UTC)
Hmm. The funny thing is, the system seems happy to accept {{DISPLAYTITLE:<span style='content:"completely different title"'>original title</span>}}, which circumvents the restriction on some browsers (Opera, Konqueror). The same goes for {{DISPLAYTITLE:<span style="display:none">original</span> title}} (this one does have universal browser support). — EmilJ.13:38, 22 July 2009 (UTC)
Well, I'm not actually saying that this is the correct title, if anything it should be ∈-induction, and people have been known to have problems with displaying this character in crappy browsers. I was asking about the general principle. — EmilJ.13:05, 22 July 2009 (UTC)
Although from EmilJ's span examples, it looks like that restriction can be got round:) The question is whether we want to allow ourselves to do that (and whether the devs would stop it working if they found out...)--Kotniski (talk) 14:01, 22 July 2009 (UTC)
The problem is, the current displaytitle system is using a blacklist of tags, which is absolutely not the way to do it. You can still do shit like this and this and even this. I've asked Skizzerz to make it less crazy (or revert it to a pre-weakened state). --Splarka (rant) 07:29, 23 July 2009 (UTC)
That is absolutely not to use in main space and is a horrible hack that has accessibility problems, breaks in other skins, with other font sizes, with customized CSS, with sitenotices, ... Amalthea07:54, 23 July 2009 (UTC)
Fixed, you just forgot to add in the [[File: and ]]. Added now, take a look at what I did if you like. You may also want to resize the image by changing the px (pixels). Cheers :) - Kingpin13 (talk) 19:14, 22 July 2009 (UTC)
This seems like a good place to advertise {{Image}}, which can be used in templates to accept several ways to declare an image. Amalthea19:40, 22 July 2009 (UTC)
The special characters box
Unless I'm mistaken, once upon a time the special characters box was indeed a single box—no need to select from any drop-down menus to get to the characters you want. Is there any way to restore this functionality? A preference I'm missing? Custom CSS? Strad (talk) 20:18, 22 July 2009 (UTC)
The page directs me here if I have any questions - at least it is a polite page.
My guess is that I did something very small wrong, like leave out a bracket, or "this is the end."
If so, a little help for a non-techy would be appreciated. On the other hand if there is a manual to read, I might just forget about it. Thanks for any help. Smallbones (talk) 20:57, 22 July 2009 (UTC)
If you're talking about the boxy red warning, it says that on any .css or .js subpages. Can be safely ignored. Your code looks fine. –xenotalk20:59, 22 July 2009 (UTC)
Thanks, that's good to know, but the problem is that it doesn't work. When I go to Wikipedia:featured articles it is supposed to show some articles in bold, but it doesn't. I've done the obvious little things that sometimes work, cleared the cache, exited the browser (Google Chrome), and turned the computer on and off, but no luck. Smallbones (talk) 22:17, 22 July 2009 (UTC)
I'm comparing your monobook page to Raul654's (the user who suggested that addition), and there is an obvious discrepancy. I have a feeling that the variable has_been_on_main_page is not defined, but instead needs to be called as part of featured_article_metadata. Perhaps you should try the code on Raul654's monobook.css page? TheEarwig(Talk | Contribs)22:28, 22 July 2009 (UTC)
Raul's monobook is not correct. The code you want is exactly as he posted on WT:FA:
We'll none of the 3 versions I've tried have worked so far. I'll try all of them again, and then ask Raul. Perhaps there is something really simple wrong here, like maybe a box to check in my Preferences? Smallbones (talk) 13:31, 23 July 2009 (UTC)
Automatic underlining in wikilinks to Chinese characters
In this edit to the page Fuck an editor wikilinked from the Chinese character 干 (Unicode U+5E72, pinyin gān) ("shield", and other meanings as described in http://en.wikipedia.org/wiki/%E5%B9%B2 ), which was not linked from before. Wikilinking on this Wikipedia automatically underlines, making this character as wikilinked look like the character 王 (wang, meaning "king"). Many other Chinese characters get a completely different meaning if an underline is added. It would be useful if wikilinking from Chinese characters and some other foreign or scientific symbols did not cause automatic underlining. Anthony Appleyard (talk) 09:13, 23 July 2009 (UTC)
This is a pretty large change, but I just wanted to know how feasible it would be. My problem is that, in places like WP:VP, I have certain sections that interest me. For example, this section I'm now writing is interesting to me, and I'd like to follow it up. But I'm not so excited about the discussion about underlining Chinese characters links. How possible would it be to add a feature where, instead of "watching this page", you could "watch this section", and the only changes that would show up on your watchlist would be to the sections that interest you? Could Wikipedia technically have it? Or is it really not worth the bother? Greg Tyler(t • c)17:14, 23 July 2009 (UTC)
I considered JavaScript, and I'll see what I can do with it if nothing else comes. But ultimately I was wondering about something a bit more robust... Thanks for the link though! Greg Tyler(t • c)17:22, 23 July 2009 (UTC)
Not as elegant, but when I want to ignore the threads I haven't participated in, I use the browser's Find feature (command-F on a Mac, dunno the Windows key command) and look for my name. (alternatively, I use my contrib page and just click on the section links). EVula// talk // ☯ //17:23, 23 July 2009 (UTC)
This will become feasible when LiquidThreads is implemented. I assume that you're primarily talking about watching particular sections of discussion pages, which is a feature of LQT (The user can watch individual threads...) haz(talk)17:30, 23 July 2009 (UTC)
I was going to jokily reply with some text representing me holding my breath, but I can't work out how to convert sharp inhalation into something readable. Greg Tyler(t • c)20:20, 23 July 2009 (UTC)
Template:32FC is supposed some sort of placeholder template for characters in 1632 series but I cannot figure out how it is reportedly in all these templates, including a number of major ones that are protected. Does anyone know what's going on? The template itself just has one line which wouldn't seem to be such an important function. -- Ricky81682 (talk) 23:27, 23 July 2009 (UTC)
Those templates are exceedingly nested. Many are templates to link to subsections of the characters article (see the last two days at TfD for a picture). It's been some work. -- Ricky81682 (talk) 05:25, 24 July 2009 (UTC)
This thing is a Bad Idea™ — please start pruning this thing back. I expect it got to be overly used by having been left underfoot for too long. Cheers, Jack Merridew04:50, 24 July 2009 (UTC)
That good and all, but can someone with some coding skills tell me what in the world is going on? This template cited everywhere has just five characters and a noinclude to its documentation. What in the world is so valuable about that? -- Ricky81682 (talk) 05:25, 24 July 2009 (UTC)
The thing just puts a non-breaking space into where ever (for no good reason and it could have weird effects). It's being used to park stuff in the template arguments, which are not used, so they are, in effect, comments. Junk; prune it out of usage, and delete it. Most of the stuff being hidden should probably be deleted; old shite is not useful. Cheers, Jack Merridew12:09, 24 July 2009 (UTC)
That'll only get you the oldid. I know that you can get the page id (=curid) through an API query, e.g. <//en.wikipedia.org/w/api.php?action=query&indexpageids&titles=Henry+VIII+of+England>, but no on-wiki way to get to it. FWIW, your link will get even shorter if you write http://en.wikipedia.org?curid=14187. Amalthea13:24, 22 July 2009 (UTC)
Well, except from the page source :\ Try pasting javascript:void(prompt('','http://en.wikipedia.org?curid='+wgArticleId)) into the address bar of your browser on some page. With most browsers, you could store that javascript link in a bookmark, so you'll just have to click on it. Or you could add some script to your Monobook.js that displays the id on every page. Amalthea13:33, 22 July 2009 (UTC)
Neat! Thanks a bunch, Amalthea. Btw, TheDJ, I think I remember that the earlier version had the curid too somewhere, not sure it's in the Permanent Link though. I think it's in the revision history or something Bennylin (talk) 07:16, 23 July 2009 (UTC)
There's some advantage having a reliable short URL, such as something started with wikiped.ia/123, rather than bit.ly/123 or tinyurl.com/123. I don't put much faith on those links, they're like cats in a bag. (sorry if I got the expression wrong).
I was wondering if it would be possible to count how many users are editing Wikipedia in a given timeperiod (let's say, an hour). Technically, the data is there - on Special:RecentChanges; the problem is that counting that manually is a daunting task. I was wondering if we already have a tool to compile such statistics, and if not, could such a tool be written? It would provide a very interesting statistic, an equivalent to "x users active at the same time" in MMORPGs (if we compare Wikipedia to Second Life, we have roughly the same amount of registered accounts, but it's much more difficult to compare user activity; such stat would help a lot).
I have created a tab based page similar to Wikipedia:Tutorial on my personal Wikipedia. I can see that the titles in the tab are scrolling to two lines and doesn't appear neat. How can I increase the width of each tab. I have copied the same code which is present in the above mentioned page.
Thanks -Amol.Gaitonde (talk) 13:41, 24 July 2009 (UTC)
Wikipedia is a free encyclopedia on the Internet. The top of this page says the page "... is used to discuss technical issues aboutWikipedia". Your page is on another wiki which probably uses the same MediaWiki software but is unrelated to Wikipedia. This page is not the place to discuss your company's internal wiki and besides, we cannot say what you did wrong without seeing your page source. Just a small note: When you edit Wikipedia:Tutorial, the bottom of the page lists six pages starting with "Wikipedia:Tutorial/Tab". They are all part of making the tabs. PrimeHunter (talk) 11:19, 26 July 2009 (UTC)
"Kullanıcı mesaj" is apparently Turkish for "User talk"; they must have "E-mail me when my user talk page is changed" enabled by default there. What most likely happened is that you followed a link to the Turkish Wikipedia for the first time so SUL automatically created your linked account there, and someone watching the "new user" log gave you the Turkish equivalent of {{welcome}}. Anomie⚔14:24, 24 July 2009 (UTC)
I followed the link above, and a couple of hours later I received my own Turkish welcome message in my email. — Martin (MSGJ · talk) 16:05, 24 July 2009 (UTC)
I got a Ukrainian one today. I would have preferred Turkish, at least I would have been able to read the letters. This kind of thing is a bit annoying. HansAdler16:59, 24 July 2009 (UTC)
The logs show your Turkish account was automatically created [4] 5 minutes before the welcome [5] on your user talk page, and I guess the email came after that. You probably followed a link to some page at http://tr.wikipedia.org without noticing it, and SUL recognized you. Maybe your browser history shows which pages at tr.wikipedia.org you visited at the time. PrimeHunter (talk) 11:07, 26 July 2009 (UTC)
Not that it's really important, but amusing: I got a Romanian one a couple days ago--with a link to an English translation which nicely explained to me that I obviously received it by error--since I had clicked on the English translation link, the bot or whatever it was was nice enough to deduce that I didn't speak Romanian. Peace and Passion (talk) 05:47, 27 July 2009 (UTC)
The "new section" link exists, but is no more a tab. I think it's a wise choice, since it is only used by advanced users. The "action" tab will be more usable with the second release, Babaco. Dodoïste (talk) 22:25, 26 July 2009 (UTC)
Google translate on toolbar
Somebody helped me a while back add a A-Z link in my tool bar. As I use google translate a lot to extract basic facts I was wondering if somebody could give me the syntax needed in my monobook to enable me to also have a direct link to google translate in the top panel here. If I could recall who it was who helped last time I'd have asked him. Dr. BlofeldWhite cat19:30, 26 July 2009 (UTC)
addOnloadHook(function(){addPortletLink('p-cactions','http://translate.google.co.uk/translate_t','Google translate','ca-googletranslate');addPortletLink('p-cactions','http://translate.google.co.uk/translate?hl=en&sl=en&tl=de&u='+encodeURIComponent(document.location.href),'de','ca-googletranslate_de','Translate this page to German');addPortletLink('p-cactions','http://translate.google.co.uk/translate?hl=en&sl=en&tl=fr&u='+encodeURIComponent(document.location.href),'fr','ca-googletranslate_fr','Translate this page to French');});
here. Since the protection is marked as 16:57 (UTC), shouldn't it take effect immediately? The IP probably edited a few seconds later, but I've never seen this happen before. Enigmamsg17:05, 22 July 2009 (UTC)
Anonymous edit tokens are (currently anyways) always just <input type='hidden' value="+\" name="wpEditToken" /> and so never change or expire. The +\ is corrupted by certain badly configured open proxies (so automagically blocks edits from such, as they corrupt quote marks on all pages users edit with them). --Splarka (rant) 08:40, 24 July 2009 (UTC)
Simple idea: When wikipedia links are pasted all over the web, the websites algorithms usually leave off the close bracket on links such as: en.wikipedia.org/wiki/Stalker_(film) hence breaking the link. In such situations why doesn't wikipedia notice the obvious mistake and redirect the user? 122.108.142.120 (talk) 01:44, 27 July 2009 (UTC)
This would be simpler to do by creating lots of redirects than by modifying the software. Perhaps a bot? Algebraist02:22, 27 July 2009 (UTC)
That would be about 325,000 redirects. Personally, I see no real strong reason to do either. Parentheses are valid URL characters. Blog/forum developers should fix their code to work properly. Mr.Z-man03:20, 27 July 2009 (UTC)
The problem is that users also put URLs in parentheses, and in those cases the link breaks if the close-paren is included. Ideally, "link bare URLs in text" wouldn't be needed, but that's not the case. Anomie⚔03:37, 27 July 2009 (UTC)
MediaWiki handles it correctly. They should still fix their software. If your wonder how media wiki does it, it only matches ')' if there's '(' in the link. It also avoids matching any of the trailing punctuation character. I should write something up one day on "how to fix your broken internet software"... — Dispenser16:05, 27 July 2009 (UTC)
If a page name is not recognized, and if it contains an unmatched ( , tell the software to append a ) and try again. Easy to program in. Anthony Appleyard (talk) 09:53, 27 July 2009 (UTC)
Wiktionary has some auto-redirecting thing that has redirected me from redlinks to the proper place. Perhaps this same functionality can be co-opted here? –xenotalk17:58, 27 July 2009 (UTC)
Third day of crippled access
Many users in Russia for a third day have a crippled or no access to Wikipedia. From my computer, pages take up to a minute to load now, pictures do not load at all. Here's the discussion in Russian. Here's tracert to en.wikipedia.org from my pc right now:
3 2 ms 2 ms 9 ms c1-sh112-v701.telenet.ru [87.224.130.73]
4 10 ms 14 ms 36 ms core.telenet.ru [87.224.130.13]
5 3 ms 2 ms 2 ms 87.226.229.249
6 131 ms 124 ms 133 ms xe-1-0-0.frkt-ar2.intl.ip.rostelecom.ru [87.226.133.110]
7 119 ms 104 ms 102 ms de-cix.he.net [80.81.192.172]
8 109 ms 134 ms 140 ms 10gigabitethernet1-4.core1.ams1.he.net [72.52.92.94]
9 110 ms 111 ms 111 ms ams-ix.tc2-ams.nl.p80.net [195.69.145.52]
10 * * * [time interval exceeded] .
11 103 ms 102 ms 102 ms te-8-2.csw1-esams.wikimedia.org [91.198.174.254]
12 103 ms 101 ms 125 ms rr.esams.wikimedia.org [91.198.174.2]
It works fine from my place (and I connect to the same rr.esams.wikimedia.org). Your traceroute shows that the bulk of the ~100ms delay already appears on the way to xe-1-0-0.frkt-ar2.intl.ip.rostelecom.ru. This seems to indicate that the problem with the net is somewhere in Russia, not anywhere near the Wikimedia servers. — EmilJ.13:26, 27 July 2009 (UTC)
But then again, I may have completely misinterpreted it. I observe that this "ams-ix.tc2-ams.nl.p80.net [195.69.145.52]" appears to be down:
PING 195.69.145.52 (195.69.145.52) 56(84) bytes of data.
From 195.113.235.65 icmp_seq=13 Destination Net Unreachable
From 195.113.235.65 icmp_seq=19 Destination Net Unreachable
From 195.113.235.65 icmp_seq=47 Destination Net Unreachable
--- 195.69.145.52 ping statistics ---
48 packets transmitted, 0 received, +3 errors, 100% packet loss, time 46988ms
Thanx for the answer! I'm not too savvy in these matters. The connection is blissfully up and running smooth right now, pictures are showing, but I'd like to know a more precise place to write in case of similar problems.. --CopperKettle17:30, 27 July 2009 (UTC)
From the sysadmin log: "Shutdown BGP peer 8342 (RTComm) to see if it resolves the reachability issues from Russia". So the connection with RTComm was severed in hopes that a path AROUND that peer does work. The problem is most likely on the Russian side however. This is a workaround for this problem from the wikimedia side. —TheDJ (talk • contribs) 17:39, 27 July 2009 (UTC)
Several years ago when the International Herald Tribune mirrored an article published on the NYTimes I would reference the IHT version because, at that time, the NYTimes required a micropayment or a subscription to see most content that was more than a few weeks old. The NYT ended that two year
experiment a year and a half ago or more.
So -- I'd like to discuss whether it is possible to write a bot that goes to every reference to an IHT article, and figures out the URL of the corresponding NYTimes article.
If it is possible I'd like to discuss whether it is practical.
If it is practical I'd like to discuss whether it is worthwhile.
Seems both practical and worthwhile, rather than having deadlinks that will just get removed. I'd suggest moving this over to WP:BOTREQ. –xenotalk17:54, 27 July 2009 (UTC)
Did the NYT already fix this? I checked the first three sample entries in the above list of broken links and the old links to entries in iht.com all delivered me to the appropriate new entry on the NYT website! If this is already solved, there is no need to write a bot. EdJohnston (talk) 22:03, 27 July 2009 (UTC)
For image files, problem with "What links here"
I'm not sure if this is a "problem", but to me it is undesirable.
The file Ilse Stanley.jpg - which announces that it's a file from Commons, lists 4 pages that link to it (#What links here); but if you go to the corresponding Commons page, it appears that there is only one page that links to it - a page not really related to the others. Seems to me that, being the same file, the "what links here" should come up the same in both places!
I hope this is the right place to ask this question—would appreciate it if you could put an alert on my "Talk" page to any answer. — Martha (talk) 03:06, 28 July 2009 (UTC)
"What links here" only covers that particular project; on Commons, it won't show pages that link to the images on other projects, just on Commons. You can use the "check usage" tab on Commons images to find where all the image is used (though it'll be actual uses of the image, not just links to it). EVula// talk // ☯ //03:29, 28 July 2009 (UTC)
Has anyone else noticed that JPEG thumbnails on Wikipedia seem significantly brighter and warmer than the original images? Check out this thumbnail compared to the actual image as an example. What are we using to create our thumbnails? Imagemagick? GD? I wonder if it is respecting embedded color profiles or not. Kaldari (talk) 16:59, 27 July 2009 (UTC)
For jpeg, it will be imagemagick's convert. And it's not impossible that it is ignoring those, because I believe the jpeg's are stripped of EXIF before they are resized.... —TheDJ (talk • contribs) 17:32, 27 July 2009 (UTC)
Quite apart from anything else, the thumbnailed image has had its focus significantly improved. I can't say that's a Bad Thing... Happy‑melon17:40, 27 July 2009 (UTC)
Ok confirmed, Wikipedia uses convert with the -thumbnail option, which also triggers -strip. This latter option strips all metadata and profile data from the generated images. Because "you don't need them" in the web thumbs. Of course our definition of thumbs is somewhat stranger then for most hosts. I'm struggling to find a way to strip meta, but NOT profile information however... It sure is interesting, if anyone has ImageMagick tips, I welcome them. —TheDJ (talk • contribs) 18:21, 27 July 2009 (UTC)
Maybe we should use -resize instead of -thumbnail then (since our thumbnails aren't really used as thumbnails). Is the only difference between these two flags that -thumbnail strips the profiles? Kaldari (talk) 18:32, 27 July 2009 (UTC)
Ticket created bugzilla:19960. -strip removes metadata AND profile data. Though it is possible to remove JUST profile data, it does not seem to be possible to remove ONLY metadata (which can be a LOT of bytes on a small 240x160 thumb). I'm asking around a bit, but haven't been able to find if it is possible at all to only remove metadata. —TheDJ (talk • contribs) 18:35, 27 July 2009 (UTC)
If someone can comment on how this effects these thumbs, and what you think is a bad idea etc, please speak your mind. —TheDJ (talk • contribs) 00:18, 28 July 2009 (UTC)
It looks like the strip issue will soon be moot according to the Imagemagick developers. (They are going to change -thumbnail to leave the color profile intact.) Regarding the other examples, I think the thumbnails with sharpening look better. The quality change is very hard to perceive. I imagine this is because 80 is very close to the default value of 85. In this particular example, I think I slightly prefer the one without the quality change as the fine details are slightly more defined, although the difference is nearly imperceptible. Kaldari (talk) 18:41, 28 July 2009 (UTC)
Discrepancy between a diff and its final revision
I posted the below at the help desk but didn't get an explanation. I've since made a null edit which seemed to fix the problem, but I'd still be interested in knowing what was going on. Anyone have an idea?
Here is a version of the Watchmen article. Here is a second. For me, the former contains a section called 'Story' and the latter does not. But looking at the diff between them, I see a) no change in the article and more importantly b) a story section in the latter version, not visible when viewing that revision directly.
And, although I can't think why it would affect a permanent link like that, I had tried purging my cache to no avail. Olaf Davis (talk) 16:37, 28 July 2009 (UTC)
Yes, now. The indented part of my post was what I originally wrote at the help desk, and as I say above I subsequently fixed it with a null edit. Perhaps I should have rewritten the whole thing but I was feeling lazy. Olaf Davis (talk) 19:12, 28 July 2009 (UTC)
Contributions prior to a block in block logs
Similarly to the article history prior to a protection linked in protection logs, it could be useful to have a link to the user contributions prior to a block in block logs. This would ease the search for informations on the circumstances and reasons for the block, though this may be controversial, as most block-related matters. What do you think ? Cenarium (talk) 20:08, 28 July 2009 (UTC)
This was supposed to be posted at the village pump (proposal), although this is also relevant to this forum, it has been re-posted there. Cenarium (talk) 21:09, 28 July 2009 (UTC)
Extract the specific revisions the user created in user:contributions
What do you want exactly from that data? All articles edited by that user? Or just ones originally created (first edit), or last edited (top edit), or just the specific revisions the user created? Sounds like an odd way to get a list of articles for export in any case.
If however you just want an XML of a user's contributions (not related to what Special:Export does), try the api? --Splarka (rant) 07:22, 21 July 2009 (UTC)
Thank you splarka for your time. I looked at the link provided, and I also installed and used the usercontributions.py bot. Both only list the edit summary, but do not actually list the specific edit differences. I am looking for "just the specific revisions the user created". Is there a tool to do this? Rumpsenate (talk) 16:52, 24 July 2009 (UTC)
Well you can get the diffs by using the revision IDs (acquired previously) and popping it into a link like "/w/index.php?&oldid=####&diff=prev". If you don't want/need the diff, you can pop the revision IDs into the api for contents, eg: revids=1|2|3|4|5 . --Splarka (rant) 07:15, 26 July 2009 (UTC)
Wow, that is incredible, I see the example you give is from this edit.[6]
I notice the example you provide only takes the first three numbers, 1|2|3.
Does each number represent a different revision id?
looking at the sourcecode for my revids=304364959 there is no difference in the source code between my new edit and older ones.
Is it possible to get just "the specific revisions the user created"? I notice in your example it lists the whole page also. Rumpsenate (talk) 21:42, 26 July 2009 (UTC)
Each number is a revision ID yes, so 3 recent edits of yours would be: 304370201|304370077|304369694. This doesn't show the difference, this shows the full text of the article at the revision you had created (as mediawiki stores edits as the full text (and this is what I meant by the "specific revisions" you had created: the full text of each article after your edits)).
If you want a diff for each, you can either get the text of the previous revision as well and use an external diff tool, or use the UI diff method (there is an API diff method, but it relies on the same diff cache as the UI, and looks pretty awful in the API). Bit messy. --Splarka (rant) 07:34, 27 July 2009 (UTC)
"you can either get the text of the previous revision as well and use an external diff tool"
Can you point me to the external diff tool?
I looked up "external diff tool" mediawiki and got little results. Searching other combinations of this is even less helpful.
Searching mediawiki "external diff" was much more promising.
"use the UI diff method"
User interface = UI? can you point me to the script/instruction page?
"API diff method"
Really? I will take any code no matter how ugly at this point.
(undent). By external diff tool, I mean you can use any tool that compares two pieces of text (nothing to do with MediaWiki). The UI method is simply sending &oldid and &diff parameters to index.php. See mw:Manual:Parameters to index.php. The API diff method does the same thing, but cannot be relied on because it doesn't always generate the diff on the query (or last time I tried), because it uses the UI diff cache (since diffs are expensive). Don't bother with API diffs at this time. There is no tool to do what you want that I know of, you'll have to write it. I can't really help you with anymore than this. Sorry. --Splarka (rant) 07:52, 29 July 2009 (UTC)
Calculating the distance between two coordinates
Hello everybody! I'm an user mainly on eu:wp. There I have developed a template for rivers with two sets of coordinates, one for the source and the other one for the final point of the river. The result can be seen for example in eu:Urola. The question is: would it be possibe to calculate in the same template the lineal distance from one point to the other using this coordinates? -Theklan (talk) 18:09, 22 July 2009 (UTC)
I wasn't trying to put the programming tasks on your back, by the way, just pointing the way to an experienced programmer who could translate the script on that page into a wiki useful template function, through [7]. I don't think I could do that as easily as some of the other participants here, or know if it is even possible. Sswonk (talk) 18:28, 22 July 2009 (UTC)
It is pretty complex, I'm hoping someone else here will try to give you an answer and possibly assistance in writing the function. Give it a few hours for everyone to respond. Sswonk (talk) 18:33, 22 July 2009 (UTC)
A calculation that complicated will take a lot of nested templates. I don't think the iterative part can be fully implemented, but may be approximated by defining a fixed number of iterations. (The Vincenty documentation indicates that only 3 or 4 iterations are generally required.) I think it's doable, but will take some work and may be bound by template complexity limitations. As a proof of concept, I created a template that assumes a spherical earth, which appears to be accurate to within 0.5%. Sample results follow:
Paris to Berlin 875.9 km (Vincenty 878594.288 m)
Los Angeles to Perth 15015 km (Vincenty 15023747.718 m)
London to Rome 9664.5 km (Vincenty 9635442.275 m)
Linear distance along the Urola/eu:Urola river: 35.5 km (note this is not typically how river lengths are measured.)
Is this roughly what you are looking for? At this point it only accepts numeric lat/long values, but several options could be supported with wrapper templates. It may be possible to improve accuracy with a few tricks, but I don't expect I can get it better than 0.1% accurate without implementing the full Vincenty algorithm. -- Tcncv (talk) 02:00, 23 July 2009 (UTC)
Well, it would be something like that, in essence. I know that that's not the way for calculating the lenght of a river, but maybe it wuold be useful for lote of thinks, like linear distances between places, roads... etc. Maybe it's only useful as a mathematic artifact. I like the resolution of this template, in fact. -Theklan (talk) 10:57, 23 July 2009 (UTC)
I'm not sure I see how I could use the Spherical law of cosines, but the Haversine formula is a define simplification of the vector calculation I originally used. Both assume a sphere and they return the same results. I have updated the {{User:Tcncv/Geodist}} template to implement the simplified calculation, and eliminated a nested template in the process. Thank you.
Is there any current or potential interest in this calculation beyond that stated above by Sswonk? If so, I can continue to refine it, possibly looking for a compromise algorithm that provides at least sub-kilometer accuracy. (I think millimeter accuracy is more than we need for Wikipedia.) If not, I can clean up what I have, wrap it in a friendly interface (with the usual D/M/S parameters), and put it out for general use. -- Tcncv (talk) 00:14, 25 July 2009 (UTC)
Thanks for taking the ball on this. I'm guessing you checked the linked page at Haversine formula, here, which talks about error. My initial thought is that 30m should be the maximum error allowed for Wikipedia, and am prepared to take my medicine from others who think that is to large. Thanks again. Sswonk (talk) 00:50, 25 July 2009 (UTC)
To give you a status update, I took a stab at implementing the Vincenty algorithm in {{User:Tcncv/Geodist2}}. I don't know if there is a better way to do it, but the approach I took required me to implement templates nested almost twenty deep. Even with that, I was only able to provide two iterations of the λ calculation, but enough to provide accuracy to within about 100 meters, and usually much better. Take a look a let me know if you really want the full implementation. For relatively short distances of a few hundred kilometers or less, it may be sufficient to use the much simpler version that assumes a spherical Earth and uss the Haversine formula. -- Tcncv (talk) 01:39, 29 July 2009 (UTC)
They stopped working during the server crashes but its been awhile since. When will they be live again on all wikis?-- penubag (talk) 08:04, 29 July 2009 (UTC)
Some editor made a javascript dealy that takes you to a random link, based on the links in the page your on. If you can find a page that has a giant list of dab pages, it could take you to a random one of those. I can't remember who made it, but it's in my monobook.js history somewhere. - Peregrine Fisher (talk) (contribs) 00:41, 28 July 2009 (UTC)
The problem is, the technical definition of a disambiguation is a bit convoluted. If a page transcludes a template that is linked to from the MediaWiki:Disambiguationspage list, it is a disambiguation. So to get a perfectly random disambiguation, you'd need to build a list of all transclusions of all these templates (keeping in mind the list itself will change) and then pick one. There is no index or table for any such action, so the resulting query might take several seconds to several minutes on a large wiki, probably as a toolserver tool.
There was a similar conundrum at en.wikibooks, where they wanted a way to acquire a random book (rather than a random page of a random book, or a non-book article). The plan was to take a massive number (37 I think) of categories, build a whole combined page list, and then choose a random item from the list. However, as about 10% of all mainspace articles on the wiki were books, the solution I suggested was to simply choose random pages until one was found in one of the correct categories: b:en:MediaWiki:Common.js/RandomBook.js. This sounds crazy, but on average it takes only about 1 or 2 hits to the API, getting 10 sequential random (yes, it's weird, don't think too hard) pages each time, before one is found. Depending on the percentage of disambiguations here, possibly something similar could be done as user JS or a toolserver tool. Maybe. --Splarka (rant) 07:42, 28 July 2009 (UTC)
Well, the good news with that is (if one can trust that category, heh), is that up to 5.4% of content pages (non redirects in content namespaces) could be disambiguations (though if there are a lot of non-main disambiguations, there are several in ns-4 for example, it could be less). But ~5% is worth the slightly-crazy-method of going random until you find a disambiguation. This query seems to bring one up after 3-4 tries, just needs a wrapper like the random book thing. --Splarka (rant) 07:39, 29 July 2009 (UTC)
I've found the category to be accurate. There are apparently 3,202 non-article disambiguation pages, BTW. Anomie⚔11:08, 29 July 2009 (UTC)
I've just edited the article on Coal. I did not write anything in the edit summary but the revision history page of the article shows the edit summary "Possible Scibaby sockpuppet" beside the log of my revision. I've noticed this also happened to other users who edited the same article days ago. Ope23 (talk) 11:05, 29 July 2009 (UTC)
The Wikimedia Board of Trustees election has started. Please vote.
Not sure if this is the right place to mention this, but it sure is annoying to get asked to vote, and then be denied when you click the link. WP should not ask users to vote if they are ineligible; is this a policy decision, a bug, or feature that has not yet been implemented? Webbbbbbber (talk) 21:49, 29 July 2009 (UTC)
Why no secure login (only)?
I'm puzzled about why WP doesn't use HTTPS for logins. I know that I can choose HTTPS for the entire session, but that's way overkill. OTOH, secure login is fundamental. I think that ALL logins should be secure. Passwords shouldn't be traveling the internet as open text.
I once asked this, and was basically told that it is a problem of human resources at this point. The developers can't shake their normal operations long enough to do the sysadmin work required to get this rolling (in a controlled and checked way, that doesn't bring the entire site down). Basically, make it a priority for the Foundation, and it might tell the sysadmins to focus their time there. —TheDJ (talk • contribs) 09:38, 27 July 2009 (UTC)
Also (correct me if I'm wrong here), because you would be viewing some pages without HTTPS cookies would still be vulnerable to packet sniffing, meaning although a man in the middle couldn't steal your password they could still take over your account. --Chris13:24, 30 July 2009 (UTC)
Playing ogg files
I have associated ogg files with Windows Media Player, but when I click the play button (either in an article or on the file page), it tries to play it in QuickTime. And it says I need to install XiphQT. I have installed XiphQT (and rebooted), but I still get that message. It seems what I have to do to get WMP is go to file page and click on the the file link.
Is there a known solution to my XiphQT problem.
Is there a way to associate play buttons with WMP? I don't see anything in preferences.
Used to be that when a template changed caused pages to be sorted in a new category, the categories would get updated in a few days. At most, a week.
Lately (the last few months) I've been seeing stragglers show up over a month after the original edit. Did somebody change the job queue processing mechanism?
Last winter, we would regularly see the job queue get up into the hundreds of thousands, even a million or two, and yet all the categories would still pick up in a couple days. Recently, the job queue seems to over an order of magnitude smaller (rarely over 10k), and yet pages don't show up in the category for over a month?
What's the size of Wikipedia in Gigabytes (GB) with everything included?
What's the size of Wikipedia in Gigabytes if you include everything from talk page archives, to User pages, to Afd debates and the historical versions of every single article?
According to Wikipedia:Database download, in January 2008 it was 2.8 terabytes. That's compressed, and doesn't include images or other media files. The full history dump hasn't worked since then, so it's harder to get figures. Algebraist19:18, 28 July 2009 (UTC)
2.8 TB! Thanks for the quick response. That's cheap, that's less than $500 in hard disk space after 8 yers. I thought Wikipedia occupied more space something like 10-15 TB.
If one considers the fact that Wikipedia's total expenses were over $3 million in just one year, the hard disk cost becomes even more negligible.
I wouldn't worry about that... but that figure was from January 2008, without images. So it's a lot bigger than that. –Drilnoth (T • C • L) 02:32, 29 July 2009 (UTC)
I recall that images were something like 20 TB as of some time ago. I forget if that counts only the full sized version or also the space taken up by thumbnails. Dragons flight (talk) 02:39, 29 July 2009 (UTC)
If I'm reading it correctly, "All pages, current versions only" takes up 9.4 GB, as of 2009-07-08 .[8] That probably doens't include files hosted on Wikicommons. Will Bebacktalk21:01, 30 July 2009 (UTC)
re-direction
Recently, when I Google a topic, and then click on the resultant link,
I get redirected to such sites as Toseeka.com, findadditonal.info, findlifespan.info, and on and on.
How do these links get into Wikipedia? It happens so often now that I have to copy and paste the actual link from Google, rather than clicking on the header. —Preceding unsigned comment added by Iintgrty (talk • contribs) 00:58, 30 July 2009 (UTC)
I don't think Wikipedia is responsible for these links. If your searches are made at google.com and the displayed url says wikipedia.org but the clickable link doesn't go to wikipedia.org then it sounds like your computer has been infected with malware designed to drive traffic to certain sites. Try running an updated anti-virus/spyware/malware program. Maybe List of antivirus software is of help. PrimeHunter (talk) 01:29, 30 July 2009 (UTC)
There's some more info at this site, though I personally don't advise downloading that particular program (SpyNoMore) unless you're sure it's safe. Interesting information though. There's more information if you search on the internet, but make sure you use an anti-spyware program you trust, rather than something that's advertised on a whim. Look around for something you feel secure with, or ask tech-savvy friends for advice. Greg Tyler(t • c)16:23, 30 July 2009 (UTC)
Yes, I've heard of this virus before; it exists. However, I strongly advise against going to that website: Web of Trust blocks it, so the program is potentially spyware in itself (e.g. SpySheriff). I can't be sure, but you really do need to be wary with those kinds of "anti-spyware" sites. TheEarwig(Talk | Contribs)17:07, 30 July 2009 (UTC)
Seeing as Web of Trust blocks it, I've removed the link and added a vaguer comment. You've got to be careful with these things and I don't want to give the wrong advice. Greg Tyler(t • c)20:44, 30 July 2009 (UTC)
Because node.children was introduced in Firefox 3.5. Earlier versions used node.childNodes instead, which lead to the (correct) code a few lines lower being used. Algebraist21:57, 30 July 2009 (UTC)
Sorry I don't understand technical matters well but how would adapting to html5 be beneficial? Would load times become shorter in newer browsers? -- penubag (talk) 08:14, 29 July 2009 (UTC)
Eventually, yes. Moving to HTML5 will also have benefits for our video and audio support (new HTML tags <video>...</video> and <audio>...</audio>, which are becoming natively-supported in almost all browsers). The schema is also more durable: it's harder to produce malformed markup, which causes tools like Twinkle to choke and die. And all it requires is for us to get rid of a few attributes that have been deprecated for years anyway. Happy‑melon11:37, 29 July 2009 (UTC)
I love the idea that they could suddenly decide to change the whole specification, but it ain't gonna happen. It's not really how the W3C works. So yeah, we're safe on that one. - Jarry1250[ In the UK? Sign the petition! ]16:24, 31 July 2009 (UTC)
Exactly. The odds of them up and changing the spec significantly (especially in a breaking way) is very very unlikely. Furthermore, the changes I made have been standardized for quite some time (not unique to HTML5), we're just catching up :) ^demon[omg plz]17:24, 31 July 2009 (UTC)
This is an SVG file. Normally, wikipedia will serve you a PNG thumb that is generated, because many browsers don't support SVG unfortunately. HOWEVER, it is known that the SVG renderer that wikipedia is using has some problems esp when it comes to rendering text in the SVG. There is some information about all this on commons. —TheDJ (talk • contribs) 12:42, 30 July 2009 (UTC)
What <text> tag? The full coding to produce the image and text is this:-
{| cellspacing="0" style="width: 238px; background: #F8F8FF;"
| style="width: 45px; height: 45px; background: yellow; text-align: center; font-size: 14pt;" | [[Image:ISO 639 Icon new.svg|42px]]
| style="font-size: 8pt; padding: 4pt; line-height: 1.25em;" | As of 1 July 2009, this user has [[Help:Starting a new page|started]] '''[http://tools.wikimedia.de/~escaladix/larticles/larticles.php?user=Mjroots&lang=en 333 new articles and 150 new lists], ''' on the English Wikipedia and '''[http://tools.wikimedia.de/~escaladix/larticles/larticles.php?user=Mjroots&lang=de 8 new lists]''' on the German Wikipedia.
</div>
|}
This is the file on its own. It is currently not displaying correctly here, but if you click on the image it displays correctly see here]. Mjroots (talk) 20:16, 30 July 2009 (UTC)
Well, the actual image is stored here. You can see it in Firefox, but I don't know about other browsers. Anyhow, the code actually used to generate the image (in SVG format) lies thus (notice the <text> tag):
The problem, as said before, lies in the software Wikipedia uses. Since not all browsers can show SVG files, Wikipedia converts it to a PNG thumbnail. This is what is ultimately displayed abovew and also what you see here (both thumbnails at different sizes). The issue is that the PNG converter isn't very good, and leaves some images looking really ugly as thumbnails, such as this one. The solution? Try and hurry along browsers which can't do what they're supposed to. Until that happens, you'll have to use a different file or fix the current one. Greg Tyler(t • c)20:56, 30 July 2009 (UTC)
Is that handling of SVG files something that was changed recently? If so, that change might explain why it used to work for Mjroots. —teb728tc21:59, 30 July 2009 (UTC)
The general method hasn't, but I don't track development enough to know specifics. The MediaWiki software may have changed and so might its converter (ImageMagick, for example, is in continual development). So I'm personally not sure, but it's certainly possible some code has changed recently. Greg Tyler(t • c)22:57, 30 July 2009 (UTC)
Wikimedia actually uses rsvg (due to mw:SVG benchmarks probably). And while the rsvg settings/fonts/software don't change that much, a bunch of pre-generated thumbnails were deleted during the recent server probs to make space, so thumbnails generated from long ago (I think the last rsvg change was October 2008, so before that) may suddenly start looking different now. --Splarka (rant) 07:54, 31 July 2009 (UTC)
It doesn't show up in Firefox 3.0.12 on Linux here, but it does if I edit the file to remove Helvetica from the list of fonts requested in the file; in this case, it probably depends on which fonts you have installed. A list of fonts available on Wikipedia's servers is at meta:SVG fonts. Anomie⚔23:32, 30 July 2009 (UTC)
Lately I have noticed that the article alert bot's reports are not hiding properly anymore, and I have traced the error to the following line of code in the individual peer reviews (see here for details):
Oh how wonderful, somebody decide not to implement a header right. Unfortunately, I can't just simply fix it for all the new PR because we weren't removing them when archiving. So, I've implemented a workaround fix. — Dispenser17:32, 31 July 2009 (UTC)
Problem using the NoTitle Extension with version 1.15.1
I get the following error spammed at the top of every request:
Notice: Use of undefined constant MAG_NOTITLE - assumed 'MAG_NOTITLE' in /usr/local/httpd-2.2.12/htdocs/mwiki/extensions/notitle.php on line 28
Notice: Undefined variable: action in /usr/local/httpd-2.2.12/htdocs/mwiki/extensions/notitle.php on line 36
Notice: Undefined variable: action in /usr/local/httpd-2.2.12/htdocs/mwiki/extensions/notitle.php on line 36
I've triple checked and the code is code as well as the require statement needed in LocalSettings.php.
Ideally, what I'd like is this: when I look at my watchlist, every time someone pastes this template or any of several others on someone's user talk page, I'd like it to appear in chronological order on my watchlist just as if someone had edited an article on my watchlist.
Obviously I could do that if it were about editing a template, simply by watchlisting the template, but this is about using the template.
Almost as good would be that this would not appear in my watchlist, but in another "special" page. I'd click on that page, and I'd see all of today's uses of that template on people's user talk pages, and if I scroll down, it would not just be today's, but yesterday's, etc.
Sort of this: Special:RecentChangesLinked/Category:User_talk_pages_with_Uw-coi_notices. Anticipating "That isn't exactly what I wanted" -> but it is possibly closest without watching every change to User/talk namespaces; and "Is there any way to filter out changes that aren't adding or removing the template?" -> not on-wiki, you'd have to make a bot or toolserver tool or something. From here on out you get diminishing returns to the effort put in, sadly. --Splarka (rant) 07:18, 1 August 2009 (UTC)
Thank you. Maybe we need to improve our technology to make it possible to do exactly what I requested. But what you suggest is definitely useful. Michael Hardy (talk) 20:44, 1 August 2009 (UTC)
If I add a redlinked page to my watchlist, and then move an unwatched page to the former name, it gets removed from my watchlist. Presumably this is a bug in the way the "Watch this page" checkbox on Special:MovePage works. --NE212:36, 2 August 2009 (UTC)
Template problem
{{resolved}}
Look at Template:Wilmington Radio. Now look at any radio station linked to by that template, and see what you get when you click on "e". Based on what's there when you edit the radio station article, you should be editing "Template:Wilmington Radio" (even though you get redirected). But you're not.Vchimpanzee· talk·contributions·20:01, 5 August 2009 (UTC)
Someone broke the template a while ago diff. But I fixed it now. There is a name descrepency between template title and template name there though. That should be sorted. Either the template should be moved (and the name updated), or the 'jacksonville' part should be removed all around. —TheDJ (talk • contribs) 20:16, 5 August 2009 (UTC)
See the documentation for {{Navbox}}. The name parameter has to be the page name of the template. The name parameter is used to create the v(iew) d(iscuss) e(dit) links. The diff in the post by TheDJ broke that so the v d e links got a wrong target. You copied the template source to that target but it was the wrong way to fix the problem and your copy was deleted. TheDJ made the right fix in [9]. PrimeHunter (talk) 12:58, 6 August 2009 (UTC)
Go to Category:Scouting. Click the arrow beside Category:People associated with Scouting. Mouse over People associated with Scouting and its subcats. Notice the links to the subcats are different (if you click they are broken). Mouse over other cats and subcats of Scouting. They all work except for People associated with Scouting. But if you actually click on People associated with Scouting and mouse over the subcats, the links are correct. Can you fix this? — Rlevse • Talk • 20:16, 31 July 2009 (UTC)
I did. It does it in FF and IE. And it only does this on the secure server (do you use that) and only on this one subcat. When you mouse over the links in that cat/subcat on the secure server, watch the path that comes up in the lower left of the browser, you'll see the problem. Very odd. — Rlevse • Talk • 22:02, 31 July 2009 (UTC)
{{resolved}}
I'm sure that for a long time, the "Go" button was case-insensitive - if it failed to match a page title exactly, it would take you to one that matched it up to case. (I remember experimenting with this and concluding that was the case, and updating some help pages on that basis.) However I see that at the moment this is not always true (it seems to work sometimes, but not always). Has something changed? Does anyone know what algorithm is currently applied?--Kotniski (talk) 09:17, 6 August 2009 (UTC)
If finding the exact title entered fails, the software tries all lowercase, each word (separated by spaces only) capitalized, all uppercase, and each word (separated by spaces, ASCII hyphen-dashes, parentheses, braces, periods, commas, question marks, or exclamation points) capitalized. It's also possible for extensions to try additional possibilities if all of the above fail.
What is stats.grok.se? Is it run by 'pedians? Anyways, I wonder if they could set it to display the current month, and not put the focus in the text area, making it hard to go back by pressing backspace. - Peregrine Fisher (talk) (contribs) 05:07, 31 July 2009 (UTC)
My ego will probably make me regret doing this but...
Why do I get rollback? I am not in the rollbackers group, never requested rollback permissions, and was never told I got them, YET I CAN USE ROLLBACK. I use Opera 10.00 and Twinkle. I mean, having rollback is somehow cool an' stuffs, I read WP:ROLLBACK and I understand it and I believe I only used it once outside of the sandbox anyway. I just want to know why I got it, and if it is a mistake, well, so long for rollback. =P -- RUL3R*flaming | *vandalism00:23, 3 August 2009 (UTC)
It is not a mistake. One of Twinkle's features is its own version of rollback. It is slower than 'normal' rollback, and provides a different type of edit summary, but it is, for all intents and purposes, rollback. TheEarwig(Talk | Contribs)00:27, 3 August 2009 (UTC)
That's Finnish. You probably have some online translation program running (such as Google Translate), look around for a notice saying that the page has been translated. Or, try shutting down your web browser, and then starting again. - Kingpin13 (talk) 23:23, 3 August 2009 (UTC)
I get the following error when I try to put an image in one of my articles. Please Help what is to be done in this context —Preceding unsigned comment added by 59.180.28.197 (talk) 17:22, 4 August 2009 (UTC)
Nice, PHP in some or other massive loop. Assuming you're talking about your own wiki, you'll want to google for the MediaWiki support desk and ask them. - Jarry1250[ In the UK? Sign the petition! ]18:05, 4 August 2009 (UTC)
Thanks, I did wonder about that. I thought for certain that not including the quotation marks would make the link not function since the title included the italics. *Slaps self with trout* for not trying to remove italics. TStein (talk) 19:13, 4 August 2009 (UTC)
Problem with some png equations being cut off at bottom using Firefox 3.5
In firefox 3.5 the first div=0 is cutoff at the bottom by 1 or two pixels while the second is not. I checked the png file through the view source and it looked good. In fact the two pngs are the exact same file. The problem seems to be miscalculating the size of the first png.
TStein (talk) 20:48, 4 August 2009 (UTC)
I'm posting this in Firefox 3.5.1, and both images render identically to me. Also, I checked, and Mediawiki is serving the same image in both cases, just as you say - so if there is an issue, it's with Firefox. Are you using the original 3.5, or 3.5.1? I noticed some rendering issues with 3.5 myself, but that release has other problems, so if upgrading fixes the issue, there's really nothing we need to do. — Gavia immer (talk)20:59, 4 August 2009 (UTC)
I had 3.5.1 before and I just applied an upgrade to 3.5.2 and it is still there; although it might be smaller. I never noticed this problem until today. When I started to see it all over wikipedia. I have not checked other sites too much, though or to make a test case. Whatever it is it is irritating and may be very hard to chase down. TStein (talk) 21:21, 4 August 2009 (UTC)
I realize that there's been a lot of issues with the image server recently, so if this is just a byproduct of that, please let me know and I'll just wait this out. Otherwise, I've got a problem with the way two of my newly uploaded images are rendering on the core (manufacturing) article: file:chaplets.svg and file:casting_cheek.svg. They render just fine in the commons preview, but don't render properly in the article. There's also a third image that I recently uploaded in that article that's just fine, so I'm thoroughly confused. Thanks in advance! Wizard191 (talk) 22:07, 4 August 2009 (UTC)
They seem to render just fine above 320px width, but a fill pattern or something is getting messed up at smaller sizes. Interestingly, I can replicate a similar problem viewing the original SVG in Firefox -- zoom out a few times and the shapes suddenly turn from patterned fill with stroke outline to solid black (vs solid white in the librsvg output we render here). There may be something funky about your file that causes it to behave poorly at small render sizes. --brion (talk) 22:19, 4 August 2009 (UTC)
I was able to replicate the solid black phenomenon as well. I'm no expert with SVG, and to be honest have had nothing but problems with getting them to properly render on Wikipedia/Commons. I'm using the latest stable version of inkscape and just uploaded a "plain" version of file:casting_cheek.svg, but that didn't resolve anything. I guess I'll need an expert SVG user to come along and "fix" my SVG files. Wizard191 (talk) 22:31, 4 August 2009 (UTC)
The raw SVG files work in firefox as long as you aren't zoomed out too far. Did you check that in Safari as well? If that's the case wouldn't that mean its how the SVG files are parsed by firefox and the Wikipedia PNG converter? Wizard191 (talk) 01:50, 5 August 2009 (UTC)
Since Wikipedia:WikiProject Database analysis seems to be dead, Wikipedia:Most wanted articles has been outdated for 2 years now. It's featured prominently on many project pages and should allow newcomers to find articles they could create. As such, it's in need to be updated. Question is, who will do it? Can this be done by a bot or does anyone know how to do it? Regards SoWhy13:19, 5 August 2009 (UTC)
This was due to the forced clear used in that template. Older Opera hates floating clears on tables and divs. I've fixed it now i think. —TheDJ (talk • contribs) 00:35, 1 August 2009 (UTC)
No, that's apparently not it since it sadly still doesn't render right in Opera. Anybody other ideas on how to fix this? If it helps any, I'll note that Template:FixBunching will result in similar rendering problems if {{FixBunching|mid}} is omitted within a stack of images, but will work correctly when {{FixBunching|mid}} is present (there is some old discussion of this over on Template talk:FixBunching). —Lowellian (reply) 01:43, 3 August 2009 (UTC)
In my personal opinion, you should just update your browser. BTW. a more exact version might be handy. There have been many 9.0 builds. —TheDJ (talk • contribs) 12:32, 3 August 2009 (UTC)
[Technical] – How can I improve Google PageRank for articles with title in English, but with original title in a foreign language?
Consider the article Institute for Federal Real Estate (Germany). If I google "Institute for Federal Real Estate" I get a PageRank of 1, but if I google the proper name "Bundesanstalt für Immobilienaufgaben", I get a PageRank of ≈120. I'm guessing this is caused by Google giving high priority to an article with the same title. Is it possible to somehow put a larger emphasis on the foreign translation?
Any input would be appreciated, and if informative, may be put into an wiki-essay on the subject. --Marcus13:25, 4 August 2009 (UTC)
I assume you mean ranking number rather than pagerank (which is a scale from 0 to 10). IMHO trying to mess with this is probably a bad idea, but I would guess that the standard rules apply, which you can (ironically) google for. - Jarry1250[ In the UK? Sign the petition! ]14:47, 4 August 2009 (UTC)
Thank you for your input. Yes I meant ranking number instead of PageRank. Pardon my blunder.
Wikipedia articles are often found using an external search engine, and it is hence IMO of encyclopedic value that the appropriate article is found. According to the Wikipedia naming convention for foreign organizations, the English translation should be the article title, rather than the foreign name. However, people may search the web for the organization using its foreign name, and because it is not part of the title of the page, it does not receive the high ranking it deserves. --Marcus05:39, 5 August 2009 (UTC)
I think it would be best if we found a way to focus the search box without interfering with basic navigation. This is I think possible, just need someone to poke more at it. See bug 1864 for more discussion. --rainman (talk) 09:54, 5 August 2009 (UTC)
If and when English Wikipedia admins choose to redesign the main page to be oriented around search -- like http://www.wikipedia.org/ portal is -- then it would make sense to put the cursor in the search box. Otherwise, doing so simply interferes with keyboard navigation of the information-rich front page (eg, arrow keys, space, and pageup/pagedown do not function for scrolling). --brion (talk) 22:38, 5 August 2009 (UTC)
This is true, however people keep requesting it, and if there is a way to do both: have it focused and have the navigation working then we should do it. --rainman (talk) 23:18, 5 August 2009 (UTC)
When would tooserver errors be fixed?
If this question is just adding to a mountain of same or similar complaints, I apologize for that, but seriously, what is the problem and why the fixing (supposedly by technicians) taking so long? We can not use any tool servers that I and many have used for gauging various affairs. If here is not a right venue for the issue, please direct me to a right place. Thanks.--Caspian blue01:39, 4 August 2009 (UTC)
It is a common mistake to think that 'Clerk' was Maxwell's middle name. In fact, Clerk-Maxwell was his double barreled surname. Unfortunately the wikipedia link to the article James Clerk Maxwell only works if we separate his double barreled surname. Is it possible that this technical error can be fixed? David Tombe (talk) 18:40, 5 August 2009 (UTC)
Unless I'm missing something, the OP is suggesting that the article name should be James Clerk-Maxwell, with James Clerk Maxwell simply a redirect to the correct portrayal—the current situation is t'other way 'round. (Having said that, my cursory Google search has failed to find support for the claim)--SPhilbrickT01:22, 6 August 2009 (UTC)
If that's what he means, this is the wrong forum to bring it up. David, if that's what you mean, you should raise the issue at talk:James Clerk Maxwell. But I doubt you'll get very far, as WP prefers the most common variant of the name of anything, even when someone feels that a different variat is "more correct". --Trovatore (talk) 01:27, 6 August 2009 (UTC)
Yes, I realise now that the mistake is in the article itself. I used James Clerk-Maxwell in an edit, and somebody changed it to James Clerk Maxwell. I clicked on their alteration and discovered that the link worked, and so I wrongly assumed that the alteration had been necessary. I realize now that both links work, and that the matter needs to be raised on the talk page. Thanks for your help in pointing this out. David Tombe (talk) 07:46, 6 August 2009 (UTC)
previously titled
<span id="full"/><!--former tag allows section referencing while accommodating section name changes. DO NOT REMOVE IT.-->
Could someone make a template for this, perhaps {prev-title|title1|title2...}? People less often remove templates, and that name should make the purpose clear. M08:59, 6 August 2009 (UTC)
Hi. Do talk page "Edit notices" such this not work in project space? If they do, have I done something wrong? (I based it on my user talk edit notice.) If they don't, I'll G8 it (unless somebody else deletes it first). --Moonriddengirl(talk)12:39, 6 August 2009 (UTC)
I cannot rollback this change, though I am able to rollback other changes on other pages. While I get the screen that informs me the rollback was completed, it does not show up in the article history or my contribution history. The only aspect that is weird relative to some test rollback that I just did is that no delta comparison shows up when I try it. Yes, the info to rollback is easily fixed in normal editing but I dunno if this is a bug or not with rollback function. --MASEM (t) 15:37, 6 August 2009 (UTC)
The same user made a change and then reverted it, so the rollback is just a null edit, and hence not recorded. Algebraist15:43, 6 August 2009 (UTC)
Pareserfunctions in local installation of Mediawiki
I have a local Mediawiki installation, and am testing a few templates. Each uses basic parserfunction logic, but the parser does not seem to render it correctly in the articles in which the templates are used.
For example, one of my templates simply creates a formatted link if an article with the supplied name exists (essentially a simple version of {{seealso}}). It looks like this:
{{ #if: {{{1|}}}
:''See: [[{{{1}}}]]''
}}
Within an article with {{{1}}} set to hello, it renders as:
{{ #if: hello
See: hello
}}
(Note: the link to hello is rendered properly, but the pre tag I've used to contain the code above suppresses it, so I've simply removed it. My concern is that the parserfunction itself shows up in the output.)
Oops - I should have known to post there. Anyway, the problem was that I had installed the extension to a different local Mediawiki instance, and mistaken it for this one. Doh! Mindmatrix17:31, 6 August 2009 (UTC)
Image captions can be too wide
I realized today that, if an image caption is wider than the image it follows, the caption is simply cut off. Examples: text: [11], math: [12].
I realize that it is very difficult in CSS to make a floated div have the width of the content inside it. Is there any way we can work around this issue, though? — Carl (CBM · talk) 13:37, 30 July 2009 (UTC)
I've never seen that before. I guess I've never seen one word that was longer/wider than the image size. Would you suggest making the thumbail box wider and have the image itself stay small? If the box becomes wider, the image might as well be so, as well. As for a fix in the meantime, I'd suggest hyphenating a long word if the thumb must remain that small. hmwithτ16:05, 30 July 2009 (UTC)
The first is because it's a single word; it's effectively a non-issue in areas that actually matter, like the actual article namespace, since there's practically no conceivable reason to have a massive single-word caption for a very narrow image. As for the math, well, the argument could be made that equations like that shouldn't be relegated to the captions, but instead put in the body of the article proper. EVula// talk // ☯ //17:23, 30 July 2009 (UTC)
I only included the text example to point out it is not a problem unique to math. But the math formulas are the more problematic case, because there is essentially no other way to say "tableau for ". Making the image bigger would solve the problem, of course, but I was hoping for a more elegant fix. — Carl (CBM · talk) 21:40, 30 July 2009 (UTC)
It is due to the width style on the thumbinner div of the thumbs. This is done to force the width of the thumbnail surrounding. We will have to experiment a bit to see how much this is actually a requirement.... —TheDJ (talk • contribs) 14:46, 31 July 2009 (UTC)
Test of new template html with fixed width and style "overflow:visible;". UUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU
OK, clearly this forced width is a requirement, or your captions won't stay within the thumbsize width of course... I have figured out a way to perhaps slightly improve this. We could set "overflow: visible" on the thumbs, like I have done on the image to the right. It's hardly ideal, but at least the content is not "cut off", it just causes collisions. —TheDJ (talk • contribs) 19:33, 31 July 2009 (UTC)
Just a comment from somebody who is pretty intechnical.
We need to make this accessible not just to people with an average-width screen, but people with small screens, too. I contribute to Wikipedia with a 13 inch screened laptop, and this will affect me personally. I dream of horses (T) @ 01:13, 5 August 2009 (UTC)
[Justin Long article - tech problems?]
I was in the article about actor John Hodgman, and clicked through to an article about Justin Long. When I click on this link, or when I paste the url into my browser, I get a blank white page that says "Override this function." When I type "Justin Long" into the search bar, however, I go to the article. Is this something strange on my end? Do other people have this problem? 128.223.163.109 (talk) 23:29, 5 August 2009 (UTC)
Right now if I attempt to go to Igor Judge, Baron Judge I get a blank page (no sidebar) with just the words "Override this function." on it. However if I go to Lord Judge, which is a redirect, I see the article. I don't believe this situation can have been intended. Sam Blacketer (talk) 23:45, 5 August 2009 (UTC)
For the last minute, when I visit WP:RDC and click the "Humanities" link at the top of the page, I get a page that's blank except for the text, in Times Roman, "Override this function.". (Going to WP:RDH loads the Humanities refdesk normally.) What does this mean? Tempshill (talk) 04:12, 6 August 2009 (UTC)
This is the sixth thread on this topic on this page today, and three of the others are still here. Please check before you post. Algebraist04:16, 6 August 2009 (UTC)
I think we've got these worked out now -- a couple machines ended up not getting fully updated and were still showing the error intermittently. :( --brion (talk) 22:25, 6 August 2009 (UTC)
Good GOD, it looks complicated! (But thanks for the info., though! Now I've just got to figure out how AWB works!) P.S. Yes, it is just a cut-&-paste message. --NBahn (talk) 22:13, 6 August 2009 (UTC)
I don't see anything wrong. The text seen at the top of Template:AFDWarning is part of the source code and is not displayed when the template is used. Such things are common for templates. The template should always be substituted and only has 7 transclusions [13] currently. That's low risk. PrimeHunter (talk) 12:00, 7 August 2009 (UTC)
I'm having a problem with my edit count. 'My preferences' say that I have 425 edits but my edit count here say that I only have 339 edits. Plus, my edit information in the tool server is wrong. Can that be corrected as soon as possible? Thanks! BejinhanTalk13:26, 6 August 2009 (UTC)
Different edit counters give different counts of a number of reasons, including treatment of deleted edits and how many edits are counted for a pagemove. Algebraist14:48, 6 August 2009 (UTC)
Interestingly, my edit count (live edits) with that tool is 225 edits higher than the number in my preferences. I thought the preferences number included deleted edits, but I suppose not, considering it should be about 3k higher... EVula// talk // ☯ //16:26, 6 August 2009 (UTC)
My edit count in the preferences is about 2500 edits higher than in Soxred93's tool. I presume that's due to deleted edits. Will Bebacktalk18:16, 6 August 2009 (UTC)
With only 425 edits, there's not much to review... I'm not trying to discourage you, I'm just saying that an ER might be premature. EVula// talk // ☯ //14:09, 7 August 2009 (UTC)
There is a known fault in the Special:Preferences edit count. The count via SoxRed's tool (when {{Toolserver}}'s S1 replay is at 0) is the most accurate edit count for the time being. There is a bugzilla somewhere asking for the maintenance script to be run to re-sync the Preference's count. MBisanztalk03:17, 7 August 2009 (UTC)
Opera 8
We had a problem with Opera 8 a month ago - see this and circa July 2. The symptoms for me are:
the "Save page" button is dead (I dare not log out because I probably won't be able to log in again!)
weirdly, the "leave a redirect behind" option on moves has disappeared.
The usability initiative code was updated. That probably means that the jquery code is used if you are in the beta. There is a bugzilla:19586 that documents this bug. And then of course; UPDATE your browser, because in all honesty, no one should be using a browser that is as old as that. —TheDJ (talk • contribs) 00:56, 7 August 2009 (UTC)
P.S. from what I was told, jquery should be active ONLY on edit pages atm, Not on login pages, so you should be safe to logout and login. It is active on edit pages, wether you are in the beta or not. I'll leave a note about this in the bugticket, because i can imagine it is annoying. —TheDJ (talk • contribs) 01:04, 7 August 2009 (UTC)
Well, I am a techie so I suppose I could find my way to the code and work out what on earth jquery is all about. Problem seems same under beta or old version. But what are you doing to mess up perfectly good HTML? A submit button is a submit button surely - nowhere else I have I found submit buttons not work under Opera 8. Incidentally I use Opera 8 because it handles my RSS new pages feed just the way I like it - and not the way that later versions of Opera do it. (I also use Windows 98 - what does that make me?) — RHaworth (Talk | contribs) 12:36, 7 August 2009 (UTC)
JQuery is a Javascript API, that simplifies javascript development and will become a core part of all future javascript development on Wikipedia. Opera 8 is not so well supported in JQuery, but you can rest assured that this is ONE bug that the sysadmins will surely investigate, because it is simply not acceptable. Unfortunately the debugging facilities for Javascript were still dismal in Opera 8. I have already tried to find the problem myself, but it's hard to track down without good development tools. I'm asking about the redirect thing. No idea why that happened. —TheDJ (talk • contribs) 14:04, 7 August 2009 (UTC)
I have found the cause of the Opera 8 problem, and hopefully a fix can be added soon.
For some reason it seems sysops no longer have the "suppressredirect" permission, which has caused this redirect option to disappear. When brion joins IRC, i will tell him. There is also bugzilla:20110 for that. —TheDJ (talk • contribs) 15:41, 7 August 2009 (UTC)
This script has stopped working in beta version of Wikipedia. I am a novice javescript editor, can anyone solve this problem? Thanks. Patchy1Talk To Me!03:54, 7 August 2009 (UTC)
The problem is that the new Vector skin in the beta Wikipedia has changed IDs on a lot of divs. I think enabling the compatibility gadget ("Library and compatibility gadgets" in Preferences) should fix the problem. Calvin 1998(t·c)03:57, 7 August 2009 (UTC)
Why no project links in the other-language article link area?
I've noticed that other-language projects can have links to Commons, Wikiquote, etc. in the area just above the interwiki language links. Is this not enabled on enWP? I would like to see this in use, because it provides a consistent place to note related project content. Currently we're at the mercy of looking toward the bottom of the article for a WikiCommons box {{commons}} or inline-text template {{Commons-inline}}. See for example, the French Van Gogh article under "autres projets". Outriggr (talk) 10:46, 7 August 2009 (UTC)
There are more options available at Category:Interwiki link templates, but, as you said, they all go at the bottom of the article. However, I think people (average readers, at least) are more likely to look there than in the sidebar. hmwitht14:52, 7 August 2009 (UTC)
I'd support this, too. Having the sisterlink boxes can kind of clutter up the page, but having them in the sidebar would probably make more sense. Of course, the only problem is if readers will see them as much. –Drilnoth (T • C • L) 14:57, 7 August 2009 (UTC)
Can anyone point to the technical feature that allows this? Is it something that could already be done using markup, or does a wiki setting need to change? It's not just for readers: when you browse articles on art, as an editor, it is nice to be able to tell quickly whether the article has the same subject (linked to) on Commons. (It's too bad the proper one-to-many relationship didn't exist for the same subject, across all wikis, so every page didn't have to keep its own record of what other wikis have equivalent articles. Expect that would be a little hard to implement now!) Outriggr (talk) 04:42, 8 August 2009 (UTC)
I would like to get an idea of the relative importance of deleting articles as compared to bot activity. I very much get the impression that if not the rules of WP so much, then most definitely editors' application of the rules, in places like AfD and Speedy, are driven at least in part by the desire to keep clutter out of WP to save space.
Yet, here is a bot history with five thousand edits in four days. That's not an increase in bytes in terms of the pages, but I am a little concerned about the room that the page history takes up. I am quite sure it is a very small amount each time, but five thousand? How many bytes does it take to store the page history? Even if it was one byte, that is max article size (100k per WP:SPLIT) in twenty days; 18.25 articles per year. Anarchangel (talk) 13:37, 7 August 2009 (UTC)
If I am not mistaken, you are confusing "don't worry about performance" with "worry about performance." :P All kidding aside, there's no point in fretting about how much hard drive space discussion takes up. Most AfDs are, as you say, to keep clutter out of Wikipedia. But this is for editorial reasons, not because of hard drive space; we delete the bad pages because they don't live up to our quality levels in some fashion (chief reason being that the subject isn't notable enough to warrant inclusion), not because we have a finite amount of space and we need to keep a few extra megabytes clear for better articles. Deleted pages still exist within the system, so deleting pages doesn't free up space at all. We don't need to limit bots because they make five thousand edits in four days, either. If those five thousand edits help to improve the encyclopedia, so be it; that's the whole point of the project. EVula// talk // ☯ //14:06, 7 August 2009 (UTC)
The media files are a much larger drain on Hard drive space regardless. I think media takes about 100 times more of our HD space, then all the revisions. —TheDJ (talk • contribs) 14:10, 7 August 2009 (UTC)
A simple null edit takes ~500 bytes in revision table, ~500 bytes in cu_changes, 250 bytes in page, 700 bytes in recentchanges, 100 bytes in 'text', may grow external storage blob by few bytes or do a compressed text copy (depends). Multiply that by amount of replicas (nearly 10 for enwiki), and you get 20k per edit. A bot which does 5000 edits per day consumes at least 10MB of diskspace a day. Of course, thats just null edits - anything involving changes in links tables or bigger text differences do cost more. If anyone wants to see our core database dataset distribution, a somewhat old snapshot is at http://spreadsheets.google.com/ccc?key=pfjIQrTbpVkaIStok1hWAdg&hl=en - it doesn't show external-store use though. Cheers! Domas Mituzas (talk) 08:06, 8 August 2009 (UTC)
Searching MediaWiki content without opening your browser
Is it possible for MediaWiki users that can program very well to create desktop search applications for different platforms so that you don't have to open your browser and go to every single one of the MediaWiki wikis' search boxes? I had the idea that such an application would have the following: the appropriate title bar for the applications operating system, a toolbar with Back, Forward, Decrease Text Size, Increase Text Size, and History buttons, as well as a search box with a button next to it containing a hierarchical list that allows the user to choose which section of which version of which MediaWiki to search in. The main section of the window would display search results. There could be versions for every OS starting with Windows, Mac OS X, and Linux. Preferences would allow users to add other non-MediaWiki wikis. Do you think that this a good idea?
What's the benefit of using a separate application other than the browser? For example, there are several iPhone apps that access and reformat Wikipedia content; the benefit is that they strip out the content irrelevant to iPhone usage (which is limited to just reading, given the limited screen space and the fact that wiki-editing on the iPhone is a pain). EVula// talk // ☯ //22:18, 8 August 2009 (UTC)
Distorted image
The article and section Ancient Egyptian religion#Polytheism uses a Commons image (La tombe de Horemheb cropped.jpg). I recently went into Commons and changed the image, cropping off the uppermost portion. Since then the image has appeared distorted in the article. First the new version of the image was displayed with the taller proportions of the old (thus looking vertically stretched), and then the old version was displayed with the shorter proportions of the new (thus looking vertically squashed). I've tried resizing the image (albeit only in previews), and the same thing happens. Does anyone know what might be causing this? A. Parrot (talk) 19:20, 8 August 2009 (UTC)
The easiest way (regardless of browser) to purge the cache is to enable the UTC Clock gadget (second item in the User interface section); then all you have to do is click on the clock. EVula// talk // ☯ //22:14, 8 August 2009 (UTC)
I don't know that I have a user interface section (I'm on a Mac), but I've tried all the options for clearing and disabling the cache as WP:BYPASS describes, and it doesn't seem to make a difference. To see if this is just a problem with my browser, could someone else click the link above and compare the image thumbnail with the full-size version? A. Parrot (talk) 22:20, 8 August 2009 (UTC)
Bypassing your browser cache and purging the server cache are totally different things. WP:BYPASS describes the former, while the clock gadget provides a link to the latter. Algebraist22:30, 8 August 2009 (UTC)
Well, I have the clock now (I hardly ever look at my preferences, so I didn't know what you were talking about), but I've tried purging the cache several different ways, including following the specific instructions for images at WP:PURGE, and nothing seems to make a difference. A. Parrot (talk) 22:52, 8 August 2009 (UTC)
It seems that templates can't be passed parameters that include XML tags such as <font> (although <u>, at least, works fine). Is there a way to get around this? Specifically, I am trying to do something like:
Oh, I see.... so the problem is that in the version I did, the template thinks |bla <font color= is a parameter name, and red>bla</font> bla is the content? rʨanaɢtalk/contribs03:13, 9 August 2009 (UTC)
The simplest solution is to use CSS to remove that line entirely, since it serves no purpose for experienced users. Algebraist20:23, 7 August 2009 (UTC)
Yes, it fixed it, but I think this is a bug because, you know, surely "justify paragraphs" should only apply to content and should not justify the last (or in this case, first and only) line? x42bn6TalkMess12:48, 9 August 2009 (UTC)
I agree, I just took the path of least resistance. Currently "justify paragraphs" option works by setting the CSS rule
Is there a way to make it so that "preview" is the default button highlighted when editing, so that if I hit "return", I get a preview of the article instead of saving my edits? I've saved in mid-edit more than once by accidentally hitting "return" when I was reaching for a backslash or a quote.—Kww(talk) 19:32, 8 August 2009 (UTC)
I tried that. Doesn't work. It gives you the preview when you start, but if you hit "return" after entering the summary, it saves the file. I want "Preview" to be the active button, so that the only way an article is saved is if I click on the "save" button.—Kww(talk) 20:36, 8 August 2009 (UTC)
There are other projects that force this with Javascript. The german wikipedia of the top of my head. Should be easy to extract the JS from them. —TheDJ (talk • contribs) 22:03, 8 August 2009 (UTC)
What browser/OS are you? Unless I hit return/enter when I'm in the edit summary, it won't automatically save what I'm working on. EVula// talk // ☯ //22:12, 8 August 2009 (UTC)
It's limited to while I'm typing in the edit summary, too. This isn't a huge problem, it just happens often enough to be irritating.—Kww(talk) 00:31, 9 August 2009 (UTC)
Is it possible to use HTTPS to log in securely, but then have that page use non-secure cookies,
so that I can then edit over a standard connection.
Or is the HTTPS overheads not very significant for wikimedia? CS Miller (talk) 21:28, 10 August 2009 (UTC)
Chinese names in the Chinese name template, vs. in the WPBiography listas parameter.
Chinese names present problems on Wikipedia. Sometimes they are handled correctly, other times not, especially with the talk page listas parameter.
Example illustrates the problem
As an example illustrating the problem, consider the football (soccer) player Sun Xiang article.
The article starts out with a very pertinent warning.
{{Chinese name|[[Sun (surname)|Sun]]}} which says "This is a Chinese name; the family name is Sun."
The infobox gives his name correctly.
"playername = Sun Xiang"
Sun, and his twin brother Sun Ji both had trials with Dutch team…
Incorrect handling
The talk page is where this Chinese name is handled incorrectly.
{{WPBiography
|living=yes
|class=Stub
|priority=
|auto=yes
|sports-work-group=yes
|listas=Xiang, Sun
}}
This means the family name is Xiang, and the individual name is Sun, the opposite of the actual name.
Perhaps a bot could be written to find each Chinese name template, and then check the listas parameter on the talk page. I have not corrected this problem as it appears in the article and talk page, allowing editors here to look for themselves at the live situation. --DThomsen8 (talk) 00:17, 11 August 2009 (UTC)
Looking for advanced table/CSS help
Forest-WidgetsCorp
Tree-Eastern
Domain-Boston
Domain-NewYork
Domain-Philly
Tree-Southern
Domain-Atlanta
Domain-Dallas
Domain-Dallas
OU-Marketing
Donn
Mark
Steve
OU-Sales
Bill
Ralph
Example of the geographical organizing of zones of interest within trees and domains.
It appears I am need of a CSS expert to deal with an experiment I've waded into. I am trying to create a hierarchical example of an active directory forest structure for the article, using tables within tables and colspans. There is no "editing help" village pump page as far as I can determine, so I'm afraid this is landing here since it is likely to be seen by someone who knows enough to help.
While this sort of looks okay in Firefox and Internet Explorer, it has weird nonuniform cell proportions between the icons and text in the right-side sub-table that don't coincide with the left-side sub-table, and which I seem unable to fix. Also if your IE/Firefox browser font size is set larger than "medium" then the formatting of the whole thing screws up and turns into a tall/narrow squashed mess that doesn't look at all correct.
I am of the opinion that the structure should be able to proportionally resize larger if the font is larger but I am not a CSS expert and the CSS article isn't really at all helpful. The wikipedia help on tables doesn't seem to cover anything like this.
Example of the geographical organizing of zones of interest within trees and domains.
If no width is specified for the upper table, then the upper table stretches out to put the description text all on one line. It apparently won't wrap the cell text at the minimum-fit size for the tables within.
Also note the cells in the subtables enlarge to fill the space. I want the icon cells as small as possible and for the text cells to be as large as possible, so the text is tightly left-justified alongside the icons.
Is there a point where it is time to forget about tables for special formating, and instead just redo things as object-oriented SVGs?
For anyone else who wants to put those chunky tabs back to normal height, get them back on the left where they belong, and remove that silly menu, copy this into your vector.css:
div.vectorTabs {background-image: none;}
div.vectorTabs ul li, div.vectorMenu ul li {background-image: none; border: 1px solid rgb(167, 215, 249); margin-left: 8px;}
div.vectorTabs ul li a, div.vectorMenu ul li a {background-image: none; line-height: 1.5em; height: auto; padding: 0 .5em;}
div.vectorTabs ul li a span {padding: 0;}
#head h5 {display: none;}
div.vectorMenu div.menu {display: block;}
div.vectorMenu div.menu ul {background: none; border: none; position: static;}
div.vectorMenu div.menu ul li {float: left; display: inline-block; background: rgb(243, 243, 243);}
div#left-navigation {top: 30px;}
div#right-navigation {float: none; left: 344px; position: absolute; top: 30px; margin-top: 0; right: 0;}
#p-cactions {background: none;}
#p-search {float: right; position: relative; top: -14px;}
div#head-base {height: 50px;}
div.vectorTabs ul li.selected, div.vectorMenu ul li.selected {border-bottom: 1px solid white;}
If anyone could find a way to make #right-navigation not position absolutely (so that it goes in the same float as #left-navigation), that would be great, too - I couldn't find a way to group them back together without having to put a <div> around the two of them. ~ Keiji (iNVERTED) (Talk)08:39, 7 August 2009 (UTC)
I would like it better if the tabs were smooth like vector original instead of boxy. Actually, I like the tab format, except that you cannot force tabs to be on the top.ManishEarthTalk • Stalk15:43, 11 August 2009 (UTC)
This caching system is ridiculous
I just made 16 edits to my user CSS (the latter of which being extremely frustrating) - one would have sufficed - before finally realizing that I needed to log out and in again to get the thing to update. Strangely, yesterday morning both of two edits were accepted immediately. Had I not already experienced similar frustration with coding an optional word censor on my own installation of MediaWiki I probably would not have even guessed that I had to log out.
Yes. I initially typed that here, but I redid my entire post after actually getting it to work and forgot to put that part back. I had used &action=purge and pressed F5 and Ctrl+F5 (which AFAIK do the exact same thing in Chrome anyway) an uncountable number of times before finally logging out. ~ Keiji (iNVERTED) (Talk)10:12, 9 August 2009 (UTC)
Perhaps, instead of pressing F5, it may be preferable to follow these instructions. Certainly on Firefox, the difference between F5 and Ctrl+Shift+R is the difference between refreshing the content of the page and refreshing every page linked to it (such as CSS files). Greg Tyler(t • c)14:11, 9 August 2009 (UTC)
The caching is in your browser, not Wikimedia's software. It would be possible for us to disable it, if we were to double or triple the number of caching servers at a cost in the region of US$500,000. I'm sure you can live with clearing your cache. — Werdna • talk11:31, 11 August 2009 (UTC)
What's the easiest way to do this...
After a wee bit of discussion, I've renamed Basel-City to Basel-Stadt, since nobody ever actually calls it Basel-City. No big deal, except there are a few categories using the name Basel-City, and a couple hundred other references. I just don't know the easiest way to rectify all the -City references to -Stadt. I imagine there's a tool for this? --jpgordon::==( o )22:08, 9 August 2009 (UTC)
Hi all -- I have been unable to edit for a couple hours while logged in. On computer #1 (work, IE) I get a 500 error (internal server error); at home, laptop, Vista, Firefox, I get blank pages. However, I can edit if I log out, and also if I change my skin to Monobook (as I have now). Is this a new bug? Does this have anything to do with the "Try Beta" link I see? Anyone else having this problem, or know a fix? I usually edit with the Classic skin. Thanks for any help, Antandrus (talk)00:59, 11 August 2009 (UTC)
It's definitely the skin. I've used the "classic" skin for years now, & at 23:21 10 August 2009 (UTC) whenever I tried to edit I encountered a blank screen. This is the case on my computer at work (Firefox on Windows XP), & at home (Firefox on Linux Ubuntu). Is there a way to get people to support these alternative skins -- or at least not making changes to the software that break them? -- Llywrch05:47, 11 August 2009 (UTC)—Preceding unsigned comment added by 216.20.149.168 (talk)
They are supported, that's why it was fixed in such timely fashion. They are however not actively developed, and no developer uses this skin, so at times it might break with no one noticing. You cannot test every change in every skin, it's just bound to break. The proper way to start actively improving and supporting these skins is to volunteer to maintain them :D —TheDJ (talk • contribs) 11:19, 11 August 2009 (UTC)
I get the same problem. I normally use Cologne but had to switch to Monobook. IE shows a 500 error and Firefox shows a blank screen.--Atlan (talk) 05:50, 11 August 2009 (UTC)
On a side note (I'm not sure if this is related), The "Log out" button at the top right of the Cologne skin can't be clicked. I'm therefore unable to log out while using this skin. I hardly ever log out so I'm not sure if this has always been the case or not.--Atlan (talk) 10:42, 11 August 2009 (UTC)
I already looked there. I couldn't find anything. I might file a bug report later but just wanted to ask here first.--Atlan (talk) 11:56, 11 August 2009 (UTC)
I've been having the same problems, with the additional problem that Wikipedia pages have been taking forever to load and frequently not loading fully. It seems much better so far today, but the log out link still isn't working. Exploding Boy (talk) 16:03, 11 August 2009 (UTC)
Resolved
Mjroots (talk) 19:07, 17 July 2009 (UTC)
Fixed, you just forgot to add in the [[File: and ]]. Added now, take a look at what I did if you like. You may also want to resize the image by changing the px (pixels). Cheers :) - Kingpin13 (talk) 19:14, 22 July 2009 (UTC)
This is the file on its own. It is currently not displaying correctly here, but if you click on the image it displays correctly see here]. Mjroots (talk) 20:16, 30 July 2009 (UTC)
Fixed
The following link, Magnetic field#Force on a magnet due to a non-uniform ''B'' would not work in the article despite the fact I used copy and paste. (See Magnetic field#''H''-field lines begin and end near magnetic poles.) In fact it is not working here either. Here is a manually typed in version: Magnetic field#Force on a magnet due to a non-uniform ''B''. The only possibility that I can find is that it is somehow over 255 characters.
Can anyone help me with this? TStein (talk) 18:03, 4 August 2009 (UTC)
![[14]](http://img188.imageshack.us/img188/8180/65851010.png){kind=link}