This page contains discussions that have been archived from Village pump (technical).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.
Is anything ever going to be done about the slow speed of page loads?
It's gotten to the point where I can't use IE any more. I've been using Firefox at home, but I can't use it here, and pages are taking two to three minutes to load, depending on their length. It looks like the more inclusions on the page, the longer it takes. This has been going on for dayyyys now. Corvus cornixtalk23:51, 11 March 2008 (UTC)
I have IE7 on DSL and I haven't seen any recent problems. Lion for instance loads in under 15 seconds. Even the Spam page loads in less than 30 seconds, and it's a monster. Got a specific example of a slow page? Franamax (talk) 18:38, 12 March 2008 (UTC)
I would suggest trying to find out if there is a specific element that loads slowly. Also try disabling the PngFix script. Other things that might be worth checking are running traceroute's, to see if perhaps there is some huge delay in a networksegment that you use. And perhaps disabling IPv6 on your machine. I have seen cases where a webserver was supporting IPv6 but somewhere between server and client the IPv6 would fail, and especially on Vista this sometimes results in the most unpredictable results. I think that without more technical information, it will be really hard to solve this issue, since only a remote set of users seem to be experiencing this problem. --TheDJ (talk • contribs) 19:43, 12 March 2008 (UTC)
When you load a page, right-click and view the HTML source of the page. There should be a comment on the bottom of the page like <!-- Served by srv155 in 0.390 secs. -->. If the serve time is much more than 5 seconds for a large page, its some sort of server problem. Otherwise the problem is on your end or somewhere in between. Mr.Z-man19:58, 12 March 2008 (UTC)
I'll check that as you suggested, but I don't see how it can be on my end when it happens on two different computers running on completely different ISPs. Corvus cornixtalk20:12, 12 March 2008 (UTC)
OK, so I sit here and twiddle my thumbs as pages take forever to load, then when they've finally loaded, I look at the source code and see how long it claims to have taken to load, and the numbers are suspicious. I can time it with my watch and see that load times are two to three times as long as the source code claims. What appears to be happening, is that the entire page loads, then sits with the page (and, in fact, the entire browser - I can't switch between tabs while this is happening) locked for many more seconds than the actual load time. And then finally, the page gets released and I can scroll or edit. Corvus cornixtalk18:16, 14 March 2008 (UTC)
For example, the spam page just took 58 seconds to load, and when I look at the source code, I get "Served by srv185 in 0.099 secs." Lion took 18 seconds, and I get "Served by srv156 in 0.105 secs." Corvus cornixtalk18:19, 14 March 2008 (UTC)
I have a similar problem on my laptop. Its something with the way my plugins like Quicktime and Java process conflicting with some of my other software installs and old processor. I've given up on fixing it and am waiting till I get a new comp next year, but yes, its exactly the same. MBisanztalk18:21, 14 March 2008 (UTC)
On Microsoft Windows, you might try right-clicking on an empty spot in the menu bar and select "Task manager". You can get some info of what is happening in your system there. -- SEWilco (talk) 18:29, 14 March 2008 (UTC)
Are you using Microsoft Windows with some sort of virus protection software? Those sometimes take a long time to process a web page's contents. -- SEWilco (talk) 18:27, 14 March 2008 (UTC)
Yep, Zonealarm, with my school-modified version of Mcafee on my laptop, which compared to my home desktop is much slower due to the filters and program interactions. MBisanztalk18:31, 14 March 2008 (UTC)
My results are similar to MBisanz. I took one for the team and dropped my ZoneAlarm and TrendMicro, there is a small difference, maybe 10% (and I only see goatse now - just kidding). The bulk of time does seem to be page-rendering as opposed to network activity, but the spam page has always turned my system into a zombie anyway. Franamax (talk) 23:58, 14 March 2008 (UTC)
Like I said, this started happening about two weeks ago, and it happens on both of the computers I use, which are both using IE7 with different virus protection programs. It's possible that something got changed in the virus protection software on both machines, though that seems unlikely, but it's also possible that some update was made in the past couple of weeks to Microsoft Windows ... Corvus cornixtalk18:35, 14 March 2008 (UTC)
I see an IE7 Security Update, KB944533 on Feb. 13, 2008 and four XP updates on the same date. As I say though, I don't remember any significant change over the last month. Maybe you could rollback one of your machines and test? Too bad about that, seems a tough problem to figure out. (One last - have you tried pages as an IP?) Franamax (talk) 23:58, 14 March 2008 (UTC)
Have you tested your connection using one of the many free bandwidth meters on the net? I doubt you do enough wikipedia editing to get hit with throttling, but there might've been some change in the line between you and the central office that slowed things down. MBisanztalk00:03, 15 March 2008 (UTC)
Override a style from .infobox
I'm stuck with a major problem. We're trying to update {{user infobox}} to use class=infobox, however, one style (infobox.td {vertical-align:top}) needs to be overridden, to prevent all the embedded user language boxes to top-align it's contents. I'm trying to do so by embedding it into it's own table cell with vertical-align=middle. But, the (new) parser is not letting me create a seperate table cell using any notation, wether it is dont with HTML, or {{!}}, the output keeps spewing out wiki code. I am at my wits' end. See my struggle here. Look in the history. Output is here. — Edokter • Talk • 15:43, 12 March 2008 (UTC)
We can't seem to override the allignment anyway... We're now in progress of updating all the babel boxes instead. — Edokter • Talk • 16:11, 12 March 2008 (UTC)
I think I fixed the template inclusion issue you were having. The parser can handle this situation, as long as the tables that you include start on a newline. --TheDJ (talk • contribs) 14:03, 14 March 2008 (UTC)
Ugh, spam filter
So I tried posting a simple question to Talk:Smallpox vaccine by means of the "+" tab at the top of the page, and was informed that I had added a spam link, to whale.to, to the page, and that I couldn't save.
I found that there were about a dozen such links that had been previously posted on the page as citations. I had to tediously go through and de-link each of them, then found that I also had to go through and insert a space in the middle of all the addresses in order to have the lovely spam filter not block me from asking my little question.
Request: Make this spam filter only trigger if what *I* posted has the questionable link. I shouldn't be punished for previous links on the page. (To anticipate and counter the argument that it forces users to clean up the pages, I'd counter that it would be much more efficient for a bot to go through the current database and get rid of all the links that the spam filter would catch.) Tempshill (talk) 16:46, 13 March 2008 (UTC)
At minimum, the error message needs to be clearer - that the new section can't be added because somewhere on the page - not necessarily in the new section - there are unacceptable links. -- John Broughton(♫♫)13:44, 14 March 2008 (UTC)
The problem is that the message says You will need to remove all instances of the blacklisted URL before you can save. It would be better if it said something like You will need to remove all instances of the blacklisted URL before you can save; if you're editing only one section, you may need to edit other sections of this page before you can post to this page. -- John Broughton(♫♫)22:47, 14 March 2008 (UTC)
Change to default image page template ("Licensing" -> "Licensing:")
It would seem to me that recently uploaded images have a section title of "Licensing:" instead of "Licensing" (note the colon in the new version). Was this a deliberate change? It looks like it happened in this change (but I am not positive) about a week ago. Anyone know what's going on? - AWeenieMan (talk) 04:48, 14 March 2008 (UTC)
As you can see on the infobox there are sections (that are seperated by headings with blue background) but I'd like to make a section that is collapsable (like on {{Infobox actor}} (with the Awards won)
Wikipedia:Requested templates is usually a good place. I have implemented a collapsible section on your table. Hope it helps you. If you have any other more specific questions, just ask me. BTW. just because you could have a collapsible section in an infobox, is no reason to include it in the infobox. There is much information that is often better left for the prose sections, and this applies especially when you talk about collapsible sections in infoboxes. Perhaps a seperate template that can be used in the prose might be more appropriate ? Just trying to make sure you explore all your options. --TheDJ (talk • contribs) 13:40, 14 March 2008 (UTC)
Anyone having problems with the Wikipedia e-mail system? A few days ago someone said they sent me an e-mail, but it never arrived. An e-mail from someone else did arrive since. I've tested my account by sending myself an e-mail (not using the Wikipedia system) and it worked. I then sent myself one using the Wikipedia system (and ticked 'send me a copy'), but only received one e-mail. I'm now going to try without ticking 'send me a copy'. Could someone also send me an e-mail, and try some tests themselves? Carcharoth (talk) 12:17, 14 March 2008 (UTC)
I have this thing in my monobook which used to take me to amazon.co.uk when I clicked on an ISBN number - now it doesn't work. How do I fix it please? (I know next to nothing about how these things work, so if you can keep explanations simple, that would be highly appreciated!) DuncanHill (talk) 01:31, 15 March 2008 (UTC)
You spot the capitals way faster than I do. The version in my monobook.js right now (copied from you) goes to amazon.co.uk just fine. Are you OK now or still going to .com? Watch out you're not getting redirected (like google.com takes me to google.ca), what is the link shown when you hover over the ISBN? Franamax (talk) 03:07, 15 March 2008 (UTC)
Yes, those capitals were tricky! It now seems to work - and I haven't changed anything since the last time that it didn't work! DuncanHill (talk) 03:12, 15 March 2008 (UTC)
Broken table
List of Welcome to Paradox episodes[1] no longer works, but I'm positive that it did work just fine when I had edited it last on Jan 30, 2008 (although a reversion does not correct the issue, [2]). I even tried changing some spacing in my sandbox and can't seem to get it to work properly. I found that the problem comes from the bullets in the cast section. Like I said, I'm certain this worked at the time I edited the article. In fact, part of {{Episode list}} has a specific line break to make sure that bullets work. Did anything change in the software in this time? -- Ned Scott04:21, 15 March 2008 (UTC)
What's odd is that the following rows are not completely broken, they're still in table form. It appears that they are rendering as separate tables without lines (separate because the table cells do not align). -- Ned Scott04:24, 15 March 2008 (UTC)
Ah, at least there is a fix. This is slightly better, since it just makes a blank space and no actual text. The funny thing is that the template itself inserts a blank space just like this as a cheap hack to keep the table cell open when it's not filled out. I guess it has to work after a line break for this trick to work, though. Still makes me wonder what might have changed to cause this. -- Ned Scott05:21, 15 March 2008 (UTC)
Automated process to update images across all Wikimedia projects?
We are having a bit of a debate on the AN hereand it's clear that there is some confusion as to how the autoblocker deals with logged in users. I thought that it no longer blocked logged in users, but others think is does in some circumstances. WP:Autoblock is no help as it's aimed at blocked users. I cannot find a similar page aimed at admins. Could somone who actually knows how it works comment please. Theresa Knott | The otter sank15:33, 13 March 2008 (UTC)
OK, autoblocks effectively hardblock the last IP(s) used by the account that is blocked. So accounts that use that IP will be unable to edit for the duration of the autoblock (about a day, I think) unless it is lifted. I'm not entirely sure what happens if someone is actually already signed in to their account at the moment the autoblock happens, if that is what you are asking, but I doubt those are treated differently. There are two solutions to someone who uses IPs that are are frequently the subject of autoblocks:
Independently softblock the IP. As I understand it, softblocked IPs aren't affected by autoblocks and those using them can edit without being affected by them. The downside is that if someone already has a load of socks, they can then edit from that IP using those socks in sequence with impunity - autoblocks were after all designed to stop that very occurrence. It would take a checkuser to detect that.
Administrator accounts are unaffected by autoblocks - we can edit even from a hardblocked IP due to a feature called "IP-block exempt". There has been strong support for making "IP-block exempt" individually assignable to any user - see Wikipedia:IP block exemption. If implemented by the devs, it would allow us to give this right to those frequently caught by autoblocks or those who have good reasons for wishing to use proxies - e.g. editing from China.
Can someone please tell me how to create an infobox without using a template. I am creating a wiki (mediawiki based) for my work and the only information i can find is about utilizing the templates that are already made. I need a wiki page that really explains how to create an infobox and that breaks down the markup. I have already looked at the existing pages on wikipedia.
Thank you!
Infoboxes are usually just tables, see Help:Infobox#Basic Infobox for the basic markup usually used in an infobox or Help:Table for more details about the markup for tables. They're generally turned into templates here so the same style of infobox can be easily reused on multiple articles. -- Rick Block (talk) 15:47, 15 March 2008 (UTC)
Database contains an invalid page title
The MediaWiki software disallows the creation of pages with certain titles that are liable to abuse or difficult to parse; one of these is [[/.]]. (Notice that the link isn't parsed, even though I didn't escape it.) However, an entry for this title appears nonetheless to exist in the enwiki database. If you go to Special:WhatLinksHere/Slashdot&limit=50&from=390335&back=245947, the tenth item (as of this writing) is a link to [[/.]]. If you click on that link, you get taken to a "Bad Title" error page (even though the whatlinkshere page indicates that you should get redirected to Slashdot. Even odder, if you look at Special:Prefixindex//, the second entry in the first column (or the 4th overall) is also a link to [[/.]], but if you click on this one it is interpreted as /, which redirects to Slash (punctuation). So we have an entry in the database for a page that can't be accessed, can't be edited, and is displayed inconsistently on different special pages. I suspect this is going to take some manual mucking with the database tables to fix! --Russ(talk)22:26, 13 March 2008 (UTC)
But seriously, this is an issue for users of the api.php interface. If a query (for example, to generator=allpages) returns [[/.]] as one of the pages in the response, and the user then uses that result to form a new query that includes "/." in the titles= parameter, the new query will throw a Bad Title error. And that means that information about all the other titles in the query will be lost. --Russ(talk)09:52, 14 March 2008 (UTC)
I can confirm that the API does return pageid 414115, title "/.", in the result list when asked to generate the list of backlinks of Slashdot. Are you suggesting the API should filter these out? — Carl (CBM · talk) 14:45, 14 March 2008 (UTC)
which appears to be a bot edit... but it was actually a self reversion of the previous edit from minutes before.
The ip has already been blocked for a week, but I'm wondering if this 'impersonation' of a bot is a warnable offense, and if so: what warning? Geoff Riley (talk) 21:25, 14 March 2008 (UTC)
Interesting question. I'm not sure, however, how the IP could mark the edit as a bot edit. Minor edits have a tick box, but as far as I know there is no corresponding box for a non-bot account to mark an edit as a bot edit...without wishing to question you, are you absolutely sure you're reading the watchlist correctly? GBT/C 21:29, 14 March 2008 (UTC)
Add the page to your watchlist and you’ll see it too. IPs don’t even have a minor checkbox. This is probably an issue for Village Pump (technical). —Travistalk 21:41, 14 March 2008 (UTC)
Well, I'll get my knife and fork and tuck into a big helping of humble pie! Very strange indeed...whilst I'm not the most technically minded of people, is it perhaps possible that the bot wasn't logged in at the time...? GBT/C 21:43, 14 March 2008 (UTC)
(edit conflict)I was just about to say that as well. Probably either a bug in the software or the IP hacked into marking it as a bot edit somehow using the URL or some foreign editing thing (perhaps?). jj137(talk) 21:44, 14 March 2008 (UTC)
Yes, it's definitely as I depicted it above: at first I skipped over it because I generally ignore bot edits, but the fact that it was an IP drew my eye back to look again. If someone has found a way to mark vandalism as a bot edit then we'll have even more to examine in the chase. I'll wander over to the VPT and see if there's anyone interested. Thanks for the suggestion. Geoff Riley (talk) 21:48, 14 March 2008 (UTC)
There is definitely something wrong somewhere - I've taken a screen-dump of the result in case it's required for future evidence, as it won't be seen next time the page is edited. Let me know if you need it. — Tivedshambo (t/c) 22:41, 14 March 2008 (UTC)
That is used to hide mass rollbacks from RC, it doesn't set the user's contributions to be bot edits, and AFAIK it only works with rollback. Mr.Z-man17:15, 15 March 2008 (UTC)
How do you know that? They can't be seen on the watchlist anymore, since they aren't "top" edits, and they're impossible to find in recent changes. – FISDOF922:28, 16 March 2008 (UTC)
Never mind. I found the option in preferences. The first two links you gave were marked as bot edits because an admin reverted them with the bot option. I don't know about the last two, though. – FISDOF922:42, 16 March 2008 (UTC)
I think I've figured it out. When an administrator uses rollback with the bot option turned on, both the revert and the reverted edits are marked as bot edits. Since the edits to November 3 were self-reverted, when an admin attempted rollback, no edit was made but the "reverted" edits were still marked as bot edits. – FISDOF922:49, 16 March 2008 (UTC)
Template coding help
Can someone please help clean up Template:US county navigation box? The box is set up so that one lists the name of the county seat (for example, for {{Logan County, Ohio}}, simply type "Bellefontaine") in order to get it to link to it. However, in some counties, the county seats have more complex names. For example, {{Quitman County, Georgia}} displays the county seat as "Georgetown, Quitman County"; it would be more desirable to display as "Georgetown", but if we put in "Georgetown" in the template, it's a disambiguation page. Similarly, {{Jefferson County, New York}} has its county seat listed as "Watertown (city)". Of course the county seat isn't the surrounding town of Watertown (see Administrative divisions of New York for technicalities), but that's the way we have to list it. Therefore, could someone write a code fix to make it so that only the first part of the name displays (similar to how typing [[Bellefontaine, Ohio|]] results in Bellefontaine)? Someone already wrote one, but it didn't work, although I don't know if it was poorly written or if I put it in the code wrongly. I know that the template is a protected template; but I'm an administrator, so if you're a non-admin and write the proper code, tell me and I'll try it. Thanks much! Nyttend (talk) 00:22, 16 March 2008 (UTC)
How many counties have this problem? Are you looking for a sort of magical fix (ala your Bellefontaine example)? Or is it acceptable if some county templates have to be changed? The simplest solution would be to add a "seatname" parameter so that for Jefferson county the seat would be "Watertown (city)" for linking but the seatname would be "Watertown" for display. If that's an acceptable solution I'd be happy to code the change to Template:US county navigation box - and to test it via Special:ExpandTemplates first. Sbowers3 (talk) 11:22, 16 March 2008 (UTC)
A new parameter for footnotes
(Apologies if this has been posted before, but technical input is still needed - JB)
At Wikipedia talk:Footnotes#Page number support, a proposed change to the footnotes system is being discussed, to add a new parameter to footnotes that could be used for specifying page numbers or possibly notes. The editor who has coded up the change has asked other editors to review the coding before this is implemented.
The new feature opens the possibility of eliminating the two-section Harvard-style footnoting now appearing in many articles, making it much easier for readers to jump back and forth from a source to the article. More involvement of interested/knowledgeable editors would be welcomed. -- John Broughton(♫♫)16:31, 16 March 2008 (UTC)
Deleted archive page
I just had a Speedy Delete put on an Archive page I made: (User talk:Mattisse/Archive 16) It was explained to me that parentheses are not allowed in archive pages. The problem is that I do not know how the parentheses got there, as they are not in the code generating the archive. How does that happen? (I notice another archive has a parentheses but was not given a speedy delete.) What do do? Thanks! Mattisse (Talk) 16:57, 16 March 2008 (UTC)
The parentheses meant that the page was created in the article namespace rather than the User talk namespace, which is why it was nominated for deletion (and why I have deleted it). Pass on how it got there. If there are other archive pages in the article namespace, please let me or someone else know and they can be moved and resulting redirects deleted. Thanks, mattbr17:14, 16 March 2008 (UTC)
I do not know how it got there and barely know what you are talking about, so I am not a good person to pass anything on. I asked the question here because I don't know. Thanks, Mattisse (Talk) 19:03, 16 March 2008 (UTC)
Sorry, I meant 'pass' as in 'I don't know the answer'. After some digging, it appears you created User talk:Mattisse/Archive 15 in October 2007 with the text [[(User talk:Mattisse/Archive_16)]], which created a link to (User talk:Mattisse/Archive 16). This page ( (User talk:Mattisse/Archive_16) ) is in the article namespace because the software does not recognise it as having one of the defined prefixes for a namespace, such as Wikipedia: or User talk:, because it has the prefix (User talk: (with a '(' at the beginning). Wikipedia:Namespace has more details about namespaces. mattbr19:44, 16 March 2008 (UTC)
A suggestion re galleries
It would be useful if the <gallery> ..... </gallery> construction would display the image thumbnails in rows of as many can fit across the user's browser's window's current width, instead of always in rows of 4. Anthony Appleyard (talk) 17:23, 16 March 2008 (UTC)
I am new to this site, so I forgot the copyright tags. I fixed it, and it's showing up alright in the Commons, but not when you click on the photo from the Pine Barrens Tree Frog article. How do I get rid of the deletion tag, and does it really matter since I fixed it in the Commons? Lonerockalex (talk) 23:11, 16 March 2008 (UTC)
All the infoboxes are suddenly appearing left justified, on top of the article text - using Firefox. Anyone else seeing this? Doceirias (talk) 03:28, 17 March 2008 (UTC)
When I attempted to play a sound sample, I got a scary message from my browser (Firefox, running under Ubuntu) about somebody wanting to run an untrusted applet. I clicked "cancel" and tried the button again -- this time it played, wasn't clear whether it had anything to do with an applet. A little detective work suggests the applet it's talking about is "cortado.jar", hosted on up.wiki.x.io. I don't think the jar file was ever executed on my system (I think it was the totem plugin instead, which I reasonably trust), but it's a little disconcerting.
What concerns me here is whether people can just upload these Java files the same way they upload image files. Can they? I couldn't find a way to do it, but I also don't know where this jar file comes from. Can anyone tell me whether these files are the work of Wikimedia developers or of J Random Hacker? --Trovatore (talk) 09:00, 11 March 2008 (UTC)
The jar file is an applet that is capable of playing back ogg media files (It is called Cortado). The jar is hosted on up.wiki.x.io, but in a section where only high-level and trusted developers of the Foundation can change it. The documentation on this seems a bit outdated, but it is located here Wikipedia:Media_help. --TheDJ (talk • contribs) 09:15, 11 March 2008 (UTC)
Code signing on Java applets is only required for those that access the raw system, which this does not. You would have to have very unusual security settings for it to complain about an unsigned applet that's completely sandboxed.
Note that the reason the applet's .jar file is on up.wiki.x.io is precisely because of Java's security sandboxing; it needs to be able to load the media files there, and the sandbox prevents it from accessing files from anywhere else. Applets cannot be uploaded through the wikis by anyone else; it's put and maintained there by our server administrators. --brion (talk) 16:58, 11 March 2008 (UTC)
Well, my browser did complain, and I don't think it's all that unusual. Mozilla Firefox 2.0.0.12 running under Ubuntu 7.10; I haven't made any special security tweaks to the browser that I recall. What it said was that the classpath security code had not been finished, so maybe it was unable to detect that the applet was sandboxed. --Trovatore (talk) 17:32, 11 March 2008 (UTC)
Ah, well your problem appears to be that you've installed a wildly unsafe Java plugin on your computer -- the GCJ-based Java applet plugin has no security sandboxing, so should really *not* be used on the general internet. This would not have been installed by default (it's certainly not there on my Ubuntu 7.10 install); unless you need it elsewhere I strongly recommend uninstalling it.
Note that Ubuntu 7.10 does by default include native browser plugins for playing the Ogg Vorbis and Theora media files that we host. You can select the native player (which might show up not-quite-correctly as "QuickTime" in the player option lists), and based on a saved browser cookie it'll pick that instead of the Java player in future. --brion (talk) 17:50, 17 March 2008 (UTC)
Suggestion re diffs
I don't know how technically feasible this would be, but it would be great if when a diff is shown there was also an option to see what pages link to that diff (or perhaps also that link to wider/narrower diffs that are supersets/subsets of the diff being displayed). Obviously it wouldn't quite be analogous to the "what links here" button, as links to diffs would look like external links, but hopefully it could still be made to work. The motivation for this is that if there has been subsequent discussion regarding an edit, it would then be possible to see the diff in the context of the discussion that has already taken place, which could sometimes be very useful. Thanks. — Alan✉22:30, 16 March 2008 (UTC)
If you are referring to an example from Chess opening, the closing bracket is not blue. Look at the anchor tag in the HTML source (search for "often cataloged") and you will see the ")" is outside the tag. If your browser shows the ) in blue, that's not Wikipedia's problem. It's likely an optical illusion though. Franamax (talk) 07:39, 18 March 2008 (UTC)
Bizarreness
The links and tabs from the top of the page are now displaying down the side. Seems to be the same on all pages. Any idea what is occurring please? Screenshot here
The colours I like and chose, it was the edit/history/watch & my talk/preferences/watchlist etc moving about that was odd. As I say. thay have moved back to the right place now so all is well. DuncanHill (talk) 11:59, 18 March 2008 (UTC)
There were issues with wikipedia being down yesterday. Because of that, you probably loaded one of the CSS files only partially, causing the page to misformat. Once the caches of your browser or wikipedia cleared, the issue was gone again. I had a similar experience myself yesterday. --TheDJ (talk • contribs) 12:46, 18 March 2008 (UTC)
That makes sense thanks - I had been unable to get onto Wikipedia, and the bizarreness happened when I came back after that. DuncanHill (talk) 12:48, 18 March 2008 (UTC)
Display links to other language Wikipedias
Could you add an option / tab to unlock displaying the links to other language wikipedia pages?
I can understand if it's not everyone's cup of tea to have a list of links to languages they don't speak at the bottom of their page. However there are quite a few of us who speak more than one language. Currently the link can only be accessed through the "edit" page. I'm not talking Wikitionary here, but the pagelink. Looking up the corresponding expression in Wikitionary (hoping to find the right one) and then switching to a different language one and trying it there adds unnecessary steps and only works for one word literal translations. If you try to look up names you'd better know how they translate.
If you added a tab that displays the page with the links to other languages displayed everyone would win. People who don't want a cluttered page can stay with the usual one and those looking for more information in other languages don't have to start and "edit" process they don't intend to do. I'm not a programmer, but it should not take a lot of code and space to add that option, the links are already on the pages anyway.
Thks. Lisa March 18 08. —Preceding unsigned comment added by 71.236.23.111 (talk) 10:16, 18 March 2008 (UTC)
Thanks for your help. There might be a clearer heading for that. E.g. "Related articles in other languages" or shorter "other language pages."
I had tried that sidebar link before and landed at the main page for the respective language (maybe a broken link in that case) and had consequently not tried again. Now that I know it works it will save me a lot of time.
I'm interested in developing my bot skills, particularly to running bots which operate on a continuous basis, rather than the more script-oriented bots I'm already operating. I'm looking for a more experienced bot coder/operator who can help me get to grips with the extra knowledge and tools required to operate continuously-running bots. Kind of an adopt-a-bot-owner system :D. I can work in C++ and VB, but all of my previous bot-coding experience has been in python. Anyone interested and willing to give me a hand? Happy‑melon10:40, 18 March 2008 (UTC)
A suggestion about Wikipedia software: format of wikilinks
Currently, wikilinks can be in these two formats:
[[xxx]] :: link to page xxx, display xxx
[[xxx|yyy]] :: link to page xxx, display yyy
[[xxx]]zzz :: link to page xxx, display xxxzzz
[[xxx|yyy]]zzz :: link to page xxx, display yyyzzz
It would be useful if this form could also be allowed:
[[xxx|yyy|zzz]] :: link to page xxxyyy, display xxxzzz
[[xxx|yyy|]] :: link to page xxxyyy, display xxx
That would avoid much extra typing when linking to a page and displaying a word slightly different at the end: e.g. [[transsubstantiat|ion|ed]] rather than [[transsubstantiation|transsubstantiated]]. Anthony Appleyard (talk) 17:07, 16 March 2008 (UTC)
This would make a bit of a mess of inline image tags when they are converted to links (such as when they are blocked by the image blacklist. Does it solve any problems other than requiring a few extra characters be typed? Happy‑melon17:16, 16 March 2008 (UTC)
The code [[File:Bad Title Example.png|thumb|right|A caption]] displays as A caption when inlined, and "thumb|right|A caption" when "commented out" with a colon or trapped by the image blacklist. While the link is admittedly confusing, at least the intended image can be easily found by clicking on the link (or just hovering). Under your system, the link would display as Image:Example.pngright, which is extremely confusing, does not link to a valid image, and is difficult to accurately decipher without viewing the wikitext. Happy‑melon17:28, 16 March 2008 (UTC)
I like your idea, Anthony.
How about if Anthony's suggestions would apply to all links except links beginning with "Image:"?
I would prefer to have:
[[xxx|yyy|]] :: link to page xxxyyy, display yyy
(instead of display xxx as Anthony suggests), so that you could say things like [[User:|Coppertwig|]] or [[Wikipedia:|Criteria for speedy deletion|]] instead of [[User:Coppertwig|Coppertwig]] and [[Wikipedia:Criteria for speedy deletion|Criteria for speedy deletion]]. Things like these come up all the time. I'm not sure I can think of any example where I'd want to display xxx; but if so, maybe somebody can come up with an alternate syntax. --Coppertwig (talk) 17:51, 16 March 2008 (UTC)
Try [[Wikipedia:Criteria for speedy deletion|]]. It substitutes that for [[Wikipedia:Criteria for speedy deletion|Criteria for speedy deletion]] when you save. All of the suggestions are pretty syntactic sugar, but it'll likely serve to make wikitext more unreadable and more unapproachable for some editors and a majority of those interested in contributing. In addition, they destroy the invariant of having the left-most parameter always be the page that you're linking to. GracenotesT§18:32, 16 March 2008 (UTC)
Have to agree with the above - it's overloading something that goes away from common usage. That said, there's nothing stopping anyone from making a template variation on this, for example, it could be {{ll|xxx|yyy|zzz}} to give [[xxxyyy|xxxzzz]] which is minimal extra cost to the software and only adds 3 characters. (This is rather trivial to make, give me a shout if you think you want this) --MASEM18:48, 16 March 2008 (UTC)
(edit conflict) I was about to say that, but Gracenotes got here first; I agree with all the points. This edit is not useless, however, because Coppertwig has made a typo, which I have taken the liberty to correct (it was in the example—nothing to do with language). Interesting term, by the way, "syntactic sugar"... Waltham, The Duke of18:52, 16 March 2008 (UTC)
I am confused by Masem's response; wouldn't it require a software modification for a change in the way links work to be possible, in templates or not? In any case, double standards is something I do not believe would be helpful; creating confusion with little to no benefits is surely something undesirable. Waltham, The Duke of19:01, 16 March 2008 (UTC)
A template version of the approach would not require any software changes - it is just would spell out the long version of what the proposed addition is looking looking to do. It's a very simple template to make. --MASEM13:53, 17 March 2008 (UTC)
(outdenting) It is pretty common to want to display a link without the namespace qualifier, e.g. [[User:fred|fred]] or [[Wikipedia:Notability|notability]]. So a variation of the original suggestion would be [[xxx:|yyy]] would be linked to xxx:yyy but display as yyy. The rule might be if the first part ends with a colon, the link is both parts. Afterall, a link that is just "xxx:" is meaningless so the only way to make it meaningful is to add the second part to it. Sbowers3 (talk) 19:42, 16 March 2008 (UTC)
The main use would be in the many cases where a page's name is long and is the first part of related words, plus a bit. For example, linking from "reprocessed" to reprocessing as [[reprocess|ing|ed]].
By my suggested rules the form [[xxx||zzz]] would mean the same as [[xxx]]zzz and therefore would be redundant, so [[xxx||zzz]] could be given the meaning "display zzz, link to xxxzzz". Anthony Appleyard (talk) 22:54, 16 March 2008 (UTC)
Again, try [[Wikipedia:notability|]]. It substitutes the phrase. Implementing this feature, unfortunately, would break several pieces of software, not the least of which is AWB. GracenotesT§23:15, 16 March 2008 (UTC)
It may be enough to allow [[xxx|yyy|]] for "link to xxxyyy, display xxx"; if so, "link to xxxyyy, display xxxzzz" could be done by [[xxx|yyy|]]zzz , e.g. [[transsubstantiat|ion|]]ed . Anthony Appleyard (talk) 10:23, 17 March 2008 (UTC)
For the 1 to 10 percent of editors who do a lot of these types of wikilinks, having a more complicated syntax could reduce the amount of typing they do. For the other 90+%, you're proposing a further complication of the syntax (it's hard enough for a newcomer to understand templates, tables, piped links, and footnotes as is), making editing of Wikipedia even more difficult. Given that it's fairly easy, these days, to copy and paste text strings, I really think we don't need to further shift wikitext in favor of very experienced editors. We already have enough challenges in getting new, constructive editors for the project. -- John Broughton(♫♫)14:00, 17 March 2008 (UTC)
Totally agree. It is already bad enough that they get screwed by the syntactical difference between internal and external links. We should make Wikipedia easier to edit, not harder. —Ian Spackman (talk) 14:17, 17 March 2008 (UTC)
I shall make a parallel: light travels through the fastest route, not through the shortest one. A change like the one proposed here would have editors use less characters, but would actually delay the editing process. I do not believe this is to Wikipedia's benefit, and I prefer to think that this is unintended by the proposal's originator as well. Waltham, The Duke of16:26, 17 March 2008 (UTC)
Given that already, as well as the standard case "[[xxx|yyy]] = link to xxx, display yyy", there is the special case "[[xxx:yyy|]] = link to xxx:yyy, display yyy", and the "[[Image:xxx|yyy|zzz|etc]]" case, adding "[[xxx|yyy|]] = link to xxxyyy, display xxx" does not seem complicated to me. Anthony Appleyard (talk) 06:24, 18 March 2008 (UTC)
Regarding the special case "[[xxx:yyy|]] = link to xxx:yyy, display yyy", - let's be clear about this - it's is a shortcut for experienced editors. After saving an edit, the wikitext actually looks like this: [[xxx:yyy|yyy]], which is absolutely standard. The next editor coming along will have no idea (and no need to know) that the previous editor didn't have to type in the text to the right of the vertical bar (the second "yyy").
So if you're proposing a similar transitory wikicode, something that, upon saving an edit, will become standard wikitext, that's one thing - something much less controversial than permanently different wikicode that becomes yet one more thing that new editors have to learn (and Wikipedia has a high rate of turnover - so that's a lot of learning). If, however, you're saying something like "well, we already have a complicated syntax, so I don't see any problem with making it even more complicated" - then I can't agree. -- John Broughton(♫♫)12:57, 18 March 2008 (UTC)
Sorry, I did not know that the format [[xxx:yyy|]] is transitory: I did not know of it and I had never used it. I still think that allowing "[[xxx|yyy|]] = link to xxxyyy, display xxx", as permanent, would be helpful: e.g. to me "they believe that the Holy Host is [[transsubstantiat|ion|]]ed each time." is easier understood while editing than "they believe that the Holy Host is [[transsubstantiation|transsubstantiated]] each time.". Anthony Appleyard (talk) 06:21, 20 March 2008 (UTC)