This page contains discussions that have been archived from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.
That is in our LocalSettings.php; that code disables the counters. They're turned off on this hugely-popular website (the main page gets about 75 hits a second) because having them on would be a crippling performance drain. There is a whole website,
http://stats.grok.se which holds similar statistical information.
Happy‑
melon08:42, 25 October 2008 (UTC)
And also, I believe, because the heavy reliance on passive
squid caching makes the statistics generated by a server-side page counter somewhat meaningless. Note that dammit's stats are based on the throughput of the squid clusters, not the PHP scripts. -
IMSoP (
talk)
15:17, 25 October 2008 (UTC)
I think that there's a problem with {{ns:3}}: it returns User talk when the template seems to be looking for User_talk. I might be wrong though, the code is intricate... --
lucasbfrtalk14:14, 25 October 2008 (UTC)
Done I fixed it the dirty way... I guess there's a better way to do this but I couldn't find it... Please test and tell me if I broke something. --
lucasbfrtalk14:19, 25 October 2008 (UTC)
Umm, thanks, it looks OK in IE7, and the problem seems to be in Firefox, where I see each reference numbered twice AND in a scrollable (but disabled) listbox. Let me Google around a little. --
Rodhullandemu16:25, 25 October 2008 (UTC)
No, I've switched it to {{reflist}} for now, I think it's something to do with the lengths of some of the references causing bleeding. I've pencilled in to sort that out. Thanks. --
Rodhullandemu17:00, 25 October 2008 (UTC)
<noexactmatch-nocreate>
Anon accounts cannot create new articles, this policy has been in force for a long time now. However, earlier today, I wasn't logged in and searched for an article which did not yet exist. Instead of Wikipedia's helpful "do you want to create this?" response and a suitable redlink, which is found at
MediaWiki:Noexactmatch, I'm confronted with a meaningless <noexactmatch-nocreate> statement. If I am logged in and search for a redlink, I still see
MediaWiki:Noexactmatch, if I'm logged out, I get <noexactmatch-nocreate>.
How long has this been in place? Is it by accident? Where can I find the corresponding MediaWiki space page? If this is by design, then it's incredibly flawed, we should be suggesting that users sign up and contribute at every available opportunity. -
hahnchen15:10, 25 October 2008 (UTC)
Sounds like a synchronization problem with the translation files: the software was looking for the
noexactmatch-nocreate message, but couldn't find any, so it substituted a placeholder. These errors sometimes occur when the MediaWiki software is being updated to a new revision, and generally resolve themselves within minutes. (There's been some talk of making the update process more
atomic, but apparently no-one's got around to doing it yet.) Anyway, I've copied our customized
MediaWiki:Noexactmatch message to
MediaWiki:Noexactmatch-nocreate with appropriate modifications, so all should be well now. (Oh, and did you know that a section title with paired unescaped "<>" signs apparently messes up the default edit summary?) —
Ilmari Karonen (
talk)
18:38, 25 October 2008 (UTC)
Hi. When I click on an "undo" in a history the edit summary is automagicaly filled in. That's a good thing. But when I click on the "(xx changes)" in my watchlist then click "undo" the edit summary is blank. Not so good, though the boilerplate at the top of the page implies therr should be an auto-edit summary. Can that be changed? Thanks.
Saintrain (
talk)
17:59, 25 October 2008 (UTC)
Probably yes, as long as we can decide what the summary should say in the case where one undoes multiple edits by multiple editors over a long span of time. While we're at it, another thing that has annoyed me is that there's no easy way to redo a useful edit that got reverted along with a bunch of vandalism. Maybe the undo feature should prompt the user if it detects that the edit has already been undone. (The
patch utility has a similar feature, prompting the user if the patch appears to be reversed.) —
Ilmari Karonen (
talk)
19:49, 25 October 2008 (UTC)
Consecutive spaces or underscores in wikilinks are reduced to one underscore in the rendered URL. I have noticed this in the past but didn't report it. At the time I just edited the section heading to remove the multiple spaces so wikilinks would work (although that breaks section links in old edit summaries).
PrimeHunter (
talk)
00:26, 26 October 2008 (UTC)
Correction to the parenthetical remark: It appears that the section edit links in edit summaries also remove extra spaces, so removing the extra space from the section heading would fix edit summary links and not break them. That's not how I recall it from long ago when I tested the situation. As far as I currently know, the only place where links include the extra underscore is in the TOC, and that would automatically be updated if the section heading was changed, so I cannot think of a downside in removing extra spaces from section headings.
PrimeHunter (
talk)
00:36, 26 October 2008 (UTC)
On a practical level, I took the extra spaces out of everything, and it all seems to work, now. Thanks.—
Kww(
talk)
00:37, 26 October 2008 (UTC)
Offensive word in captcha
Offensive word in a Wikipedia captcha.
This image shows the
captcha I received when registering for Wikipedia. As you can see, it contains an offensive word. Probably, this was not intended, and the captcha was meant to be the words "heads" and "hits" put together. However, maybe the captcha-generating system should be changed to avoid accidentally creating offensive captchas?
GalNAc (
talk)
13:54, 23 October 2008 (UTC)
I dunno, it depends how rude the combinations can get; what if the word list contained, say, "pontiac" and "untold"? It's like realising that your URL reads as
expert sex-change... -
IMSoP (
talk)
18:54, 25 October 2008 (UTC)
It wouldn't be hard to add a simple
regexp to filter out the
usual suspects. However, as noted on bugzilla, it'd probably be easier to just remove the words that are likely to cause unwanted coincidences (specifically "hits", the only culprit reported so far) from the wordlist. —
Ilmari Karonen (
talk)
21:05, 25 October 2008 (UTC)
we'd probably get better result if we switched to random letter/number format. the odds of getting an offensive word from a randomized combo are very small. --
Ludwigs221:41, 25 October 2008 (UTC)
I think Ilmari is right. Using completely random captcha instead of dictionary wording would certainly narrow down the chance of something like this happening. Either that, or just remove "hits". - Bkid My talk/Contribs16:13, 26 October 2008 (UTC)
This may seem a ridiculous problem for some, but I know several people that would fret at the idea of having to type "head shits" in order to register at Wikipedia.
GalNAc (
talk)
21:00, 26 October 2008 (UTC)
I don't know, but all my scripts just stopped working. Oddly enough, they work when viewing article space and my watchlist, but anywhere else it looks as if my monobook has been blanked (clearing the cache didn't work). Also, the green text in my watchlist is now bold when it's a big change (which was never the case before, only the red). -
auburnpilottalk05:19, 25 October 2008 (UTC)
Wikipedia's pretty weird tonight, if I do say so myself. I'm guessing it's just a glitch in the scripts. Probably it'll be fixed in a matter of time. :)
Master&Expert (
Talk)
05:37, 25 October 2008 (UTC)
Some API modules started overloading the database after the software update and were disabled: allpages, backlinks, categorymembers, and logevents, according to
[2].
Mr.Z-man05:49, 25 October 2008 (UTC)
Twinkle down, Huggle down
As discussed above, Twinkle is down, and Huggle is not working either. With Huggle, I can do one revert after which it claims I have been logged out, and no warnings are given to vandals. Apparently Huggle too depends on these modules. –
Sadalmelik☎09:08, 25 October 2008 (UTC)
Apparently one can ignore the Huggle warning then, but it still does not give warnings to vandals:
Special:Contributions/L'Aquatique. The few warnings are manual, not Huggle-produced. Presumably the same thing also applies to reporting them to AIV. –
Sadalmelik☎11:52, 25 October 2008 (UTC)
See this makes me feel spoiled. I went in to warn the users manually, and remembered I have no idea how to do that without twinkle guiding me through the steps. There's quite a bit of vandalism going on but without twinkle and huggle, it's hard to fight. We should try to get this fixed as soon as possible... ~L'Aquatiquetalk11:56, 25 October 2008 (UTC)
I'd debug twinkle, but it seems to work just fine for me. Can you be more specific about what is wrong with it? If you want me to debug huggle you can port it to an open source programming language first. --
Tim Starling (
talk)
12:48, 25 October 2008 (UTC)
I tried to use it to warn users. It went through all the motions, even said that it was completed. But the warning never showed up. Does that help? ~L'Aquatiquetalk12:50, 25 October 2008 (UTC)
As a test, I tried to rollback the latest edit on
User:Frankenpuppy as vandalism with Twinkle. No reversion, no warning to my poor sockpuppy. I would still assume it has something to do with the disabled API modules (I know they should be up again) or with the revision that originally introduced the load issues. –
Sadalmelik☎12:57, 25 October 2008 (UTC)
Another problem: when searching to a non existent page, instead of having the red link below the tagline, there is <searchsubtitle> . CenariumTalk13:55, 25 October 2008 (UTC)
(undent) Twinkle worked for my cruel
master: see
[3] and
[4]. Hmm, should there be an automated warning, too, for vandalism reverts? There was just an edit window with all the current text of the user page but nothing else.
Frankenpuppy (
talk)
15:28, 25 October 2008 (UTC)
Please test these tools again. Revert of more stringent edit form field checks is rumored to help with such tools which submitted their forms incorrectly (missing some form fields). --
brion (
talk)
20:50, 25 October 2008 (UTC)
I've fixed twinkle now (I hope), and have reactivated the arv module; I should convert it to use the API fully, but havn't had any time at all lately to do anything twinkle related (if some one could implement a JS api for the api I would be happi happi :)) →
AzaToth23:55, 26 October 2008 (UTC)
I assume your post is regarding
[5] and
[6] which contained a link where you can download
Firefox, a free
Web browser. You have another Web browser
Internet Explorer which works for me and lots of other users. Wikipedia's servers sometimes have problems which usually don't depend on the browser of the user.
PrimeHunter (
talk)
20:29, 25 October 2008 (UTC)
By "it doesn't replace Internet Explorer" I think you mean "you can't actually remove Internet Explorer from Windows". If you use Firefox you will probably not have to use Internet Explorer again, at least not for browsing the web. —
Remember the dot(
talk)03:36, 26 October 2008 (UTC)
Yes, and you'll be the better for it. Internet Exploder is just about the most pitiful browser out there, in my opinion. I prefer
Safari for its blinding speed, but even Firefox is an improvement. Honestly, even if it did remove IE, you'd still be better off. {{
Nihiltres|talk|
log}}
15:37, 26 October 2008 (UTC)
In the days of IE5 or IE6, I'd agree; but IE7 is actually a pretty reasonable browser for most people's needs. Not to everyone's taste in its UI design, and lacking in certain features some power users would like, but I think the browser market is now really open to personal preference, with IE, Firefox, Safari,
Opera, and now
Google Chrome all offering their own take on what's "best".
So yes, I use Firefox, because I'm used to it, and because there are a lot of useful tools for
web developers. But to a new user, my advice would be to shop around, see what suits; just don't give IE special treatment because it comes pre-installed. -
IMSoP (
talk)
17:24, 26 October 2008 (UTC)
Unfortunately, that's not the main problem. We already know that the timelines are fickle and require blank lines to fix. The problem here is that the text disappears for no discernable reason. --
RattleMan05:31, 26 October 2008 (UTC)
I'm a little nervous about
this change, squeezing the TOC to the right of the world map at
Special:Booksources (per a user request). I use IE7, 1280x1024x19" and it looks OK to me. Can various others check this out and revert as necessary? It would be nice to move the world map a little farther down (bottom of the intro) but that seriously screws things up for me. Thanks!
Franamax (
talk)
10:42, 26 October 2008 (UTC)
It looks bad at 800x600 (TOC is below the map, text flows right of the map and then continues left of the TOC), ok at 1024x600 on FF3. Note that you can always test lower resolutions by simply sizing your browser window to less than full screen—some of us do that for normal browsing.
Anomie⚔14:39, 26 October 2008 (UTC)
Gigantic User Page Design Display
Today I have noted that the User Page Design Display I always use has blown up to an absurd proportion. Did something get tinkered with? Because there have been no recent edits to my template or user page. (
Mind meal (
talk)
16:51, 26 October 2008 (UTC))
add on
I seem to remember reading something about an add on tool that showed how long each part of an article had been in it. Does anyone know what tool I'm trying to remember?
RJFJR (
talk)
17:21, 26 October 2008 (UTC)
WikiTrust? Not an add-on, exactly (or not yet), since it runs on a dump of the database, not the live site. It colours individual words on a page based on the accumulated "trust" of editors that have changed it or words near it. -
IMSoP (
talk)
17:29, 26 October 2008 (UTC)
Behavior of {{ns:}} keyword has changed slightly, fix your templates!
This is something I should've advertised better before, but forgot to. Since
MZMcBride brought it to my attention, I'll do it now belatedly:
In the past, the
{{ns:}} parser function (which converts namespace numbers to prefixes, e.g. "{{ns:14}}" → "Category") used to return names containing underscores ("_") for talk namespaces (and for any other namespaces with spaces in their name, but we have none of those here on en.wikipedia). The
{{NAMESPACE}} magic word (which gives the namespace of the current page), on the other hand, returned names with spaces (" ") instead, as did the related magic words like {{FULLPAGENAME}}, {{TALKSPACE}}, {{SUBJECTSPACE}}, etc.
As a result of this mismatch, a lot of some templates and system messages that check for the current namespace have resorted to using code like this (note the extra "E"):
{{#ifeq:{{NAMESPACEE}}|{{ns:3}}|This is a user talk page.}} <!-- Don't do this, it doesn't work anymore. -->
This kluge used to work, at least here on the English Wikipedia, because the "E" variants of the magic words (which are meant to be used in URLs) coincidentally returned values with underscores in them. It never worked for namespace with non-ASCII letters in them, since {{NAMESPACEE}} URL-encodes those, but we don't have any here so there was no problem as long as nobody copied the code to non-English projects. (For an example of what happens on non-English wikis, see e.g.
fi:Keskustelu käyttäjästä:Ilmari Karonen/nstest.)
Anyway, as part of
rev:41876, I decided to finally fix the mismatch and make {{ns:}} return values with spaces instead. I figured that, since we're going to be messing around with namespace names anyway (cf. the upcoming
rename of the "Image" namespace to "File"), this was a good time to sneak that change in. It's taken a while, but it appears that the new code is now live.
This means that any templates and system messages using the "E" hack now need fixing. There are two ways to do that:
Just drop the "E". The code will then work again.
Don't use {{ns:}}. The only reason for using it is for writing code that is portable between wikis with different namespace names, and the "E" hack never was portable to begin with.
So, to recap, any of the following will work:
{{#ifeq:{{NAMESPACE}}|{{ns:3}}|This is a user talk page.}} <!-- This now works on all wikis, since
rev:41876 went live. -->
{{#ifeq:{{NAMESPACE}}|User talk|This is a user talk page.}} <!-- This has always worked on en.wikipedia, but is not portable to other languages. -->
{{#ifeq:{{NAMESPACEE}}|User_talk|This is a user talk page.}} <!-- This has also always worked, but is not portable. -->
Also note that templates that only use {{ns:}} to test for non-talk namespaces need no fixing, since there has been no change in behavior for those changes.
I'm sorry for not making more noise about this before the change went live. (As a limited excuse, I had no real idea when it was going to go live. :-) ) I'll try to fix any code I can find that has broken because of this, but all help is welcome. Again, sorry. —
Ilmari Karonen (
talk)
18:55, 26 October 2008 (UTC)
Hello everyone: Often, when I edit an article I push several times the "Show preview" button before I confirm the changes. And frequently I doubt thinking that perhaps I have changed something by mistake, so the article would get one or more errors. This could be solved if on top it would appear a comparison between the present content of the article and the proposed one, as it happens when the "Compare selected versions" button of history is clicked. Thanks, --
Edupedro (
talk)
22:00, 26 October 2008 (UTC)
Do you know you can click "Show changes" to the right of "Show preview"? Or do you think that is unsatisfactory and want to see preview and changes at the same time? I certainly don't think that should be default. I suppose there might be a fourth button saying "Show preview and changes", but I don't think that is needed.
PrimeHunter (
talk)
23:04, 26 October 2008 (UTC)
Sorry: They have told me about the "Show changes" button. I've been quite long editing Wikipedia and hadn't used it yet ... Now I know. Thanks, --
Edupedro (
talk)
23:01, 26 October 2008 (UTC)
Thanks, PrimeHunter: I wrote my own answer before yours appeared in my computer. I think all is OK as it's now. Regards, --
Edupedro (
talk)
23:27, 26 October 2008 (UTC)
Yes, you also saved it before me. I clicked edit before your answer and didn't get an edit conflict when I saved later so I didn't see your answer in time. This can happen when two users edit different paragraphs in the same section. Ironically, I spoke against showing preview and changes at the same time by default, but I previewed just before saving and if changes had also been displayed then I would have seen your post and not saved!
PrimeHunter (
talk)
00:01, 27 October 2008 (UTC)
Was Wikipedia having a problem today?
I've only had this much trouble with the Internet once before, when the Internet Service Provider had a problem due to
Hurricane Ike.
Here's something I've never dealt with. I was doing a long edit, and I clicked on Preview, and the circle just kept rotating, and the part of the page where it says to submit never appeared. I finally clicked on "Back" and because I had only previewed once, I lost everything. It took a while to find and retype. And since it was a copy and paste move, the edit history looks a little weird.
I also tried to get into a help desk page and that circle just kept rotating and rotating until finally it said "Internet Explorer cannot display the webpage" or whatever it says. I tried again and it wouldn't go. Other sites worked fine. I finally gave up and clicked on the red X.
On
Wikipedia:Requests_for_mediation the button doesn't work. Under "File request" the button is supposed to say "Click here to file Request" but it says only submit query, the default. This button title is created with the Inputbox MediaWiki extension, using the "buttonlabel", see
Help:Inputbox#Parameters.
<inputbox>
type=create
editintro=Wikipedia:Requests for mediation/Top/Instructions
preload=Wikipedia:Requests for mediation/Top/Sample
default=Wikipedia:Requests for mediation/
buttonlabel=Click here to file Request
bgcolor=white
width=50
</inputbox>
I have attempted to get the button title to work on numerous pages, without success.
What is really strange, is that if I remove the top template, {{H:h|editor toc}} (which can be edited
here) from
Help:Inputbox , the same problem happens on
Help:Inputbox, buttonlabel doesn't work.
Hello: I think that the possibility to search in the history of articles would be really very useful. In other words, in
Special:Search should be a selectable box for History. This would be very handy for example to see if in any past edition of the article appeared some information that we are thinking about but now is missing, to check why (if explained in the article, discussion or summary) it was deleted and make us agree with the omission or, other times, think that it should finally appear and think how to write the information in a more correct and agreeable (by the Wikipedist community) way. Thanks, --
Edupedro (
talk)
22:14, 26 October 2008 (UTC)
Thank you: I've checked it and it works. Now it's an external tool and slow but I hope that soon it could be integrated in Wikipedia and fast. Regards, --
Edupedro (
talk)
23:14, 27 October 2008 (UTC)
Idea Regarding Protection
Hey all. I've been thinking about something for a little while, and I figured that it might be something the community may consider implementing - if at all possible.
The following is an on-wiki scenario that happens all too often - a few editors, with different opinions on a topic of some sort, push some sort of agenda on that topic. They will re-add, revert, undo, sometimes
attack other contributors they are in the dispute with, and in some cases even
sockpuppetry - this is commonly known as
edit-warring. They aren't willing to compromise and discuss with other contributors, and they are only there with axes to grind and POV's to push. So what happens? An administrator intervenes by fully protecting the page for a set amount of time. Note that the majority of editors editing the page are edit-warring, so simply blocking involved parties won't work.
But consider this as part of the scenario - there is one or two editors, readily active in the article of interest (probably
wikignomes), who have done no edit warring whatsoever and are willing to discuss and compromise in a calm and
civil manner. They aren't active in that section of the page where the edit war occured. It's not particularly fair to them to have to wait to edit a page because other editors are so impatient that they can't seem to be able to discuss things in a civil manner. They have to wait out the protection as well, despite having done nothing wrong.
My suggestion, however, doesn't sound easy. In fact, it would be quite difficult to implement technically. What I suggest we attempt do is to enable the full protection of relevant sections rather than the full article. That way, the POV-pushers won't be edit warring and regular users who have done nothing wrong can still edit other parts of the article.
Note I know it says in the
protection policy"Isolated incidents of edit warring, and persistent edit warring by particular users, may be better addressed by blocking, so as not to prevent normal editing of the page by others. But if a large majority of editors are edit warring over specific areas, then the whole article would end up being protected anyways. I think it would be best to protect that area as opposed to everywhere else, so as not to block the wikignomes from copyediting the article or any other good-faith contributor from improving it.
This is not as simple or straightforward as it sounds, unfortunately; it opens the door to a host of security vulnerabilities. For instance, editing an unprotected section and prepending text (ie adding it before the title in the edit window) would probably add the content to the protected section, where it would be impossible to remove except by an admin... vandalism anyone? Plus there is the simple fact that edit wars inevitably expand to fill all available space: if you lock down a disputed section or sentence, then people will just find another section that says a similar thing to fight over. Sometimes even protecting a whole article is not enough; ArbCom have placed blanket bans on particular edits to whole topics before now, both temporarily and permanently. The correct approach is to be as precise as possible in controlling who can edit, not what can be edited. There have been mumblings of per-user page protection/per-article blocking (the two are essentially equivalent), which has enormous potenial as well as similarly huge concerns. But that's another story.
Happy‑
melon00:29, 27 October 2008 (UTC)
Hmm. I never thought of it that way before. That would be a concern. And also, if they click the usual "edit this page" tab, then they just scroll down and edit the section anyways. I should have figured.
Master&Expert (
Talk)
00:46, 27 October 2008 (UTC)
I'll add (just as a thought) that it's probably not ever a good idea to try to implement policy through technology. technology is thoughtless, and there's always going to be someone who figures out a way to subvert technical restrictions to their own ends (consider that an ironic paean to human intelligence...). if we can't get people to play fair, then there is nothing quite like the attention of a real live admin to make sure they play fair anyway.
Ignoring the technical aspects for a moment: Full protection should never be used as a long term solution on normal articles. There's plenty of other articles that need work if one happens to be protected. If its a very high traffic article or some other situation where protection would be inadvisable, the edit warriors can just be blocked, the number of people edit warring really doesn't matter. Its just as disruptive (arguably more disruptive) to have a dozen users edit warring as it is to have 2. There's also the option of non-technically-enforced sancations like
topic bans or
1RR restrictions.
Mr.Z-man01:21, 27 October 2008 (UTC)
Just to put the idea to rest (hopefully): this has come up repeatedly in the form of requests to be able to monitor just a section of a page. The problem is that a "section" is not at all well-defined; suppose I edit a section by changing the level of the heading, changing the section heading, and splitting the section by adding another (same-level) heading in the middle - then what should be tracked? Or suppose I rearrange two adjacent sections to be three sections, by moving text around? -- John Broughton(♫♫)14:36, 27 October 2008 (UTC)
problem with the timeline tag?
I'm having a bit of trouble with the timeline tag. I was editing the timeline on the "
distributed.net" article, and as soon as I changed the year from 2007 to 2008 (which was all I did), the text disappears.
Looks like something's gone wrong with the
EasyTimeline extension. Old cached timeline images continue to work, but any new timelines are borken. Even just adding a full stop to a comment is enough to break it. —
Ilmari Karonen (
talk)
20:28, 27 October 2008 (UTC)
I have written a program in C that I wish to transfer from my computer to my TI-89 Titanium, but the file is listed as "Incompatible type" in the transfer window (even after changing the extension to .89z). What should I do? (The program can be found on my userpage)
Lucas Brown (
talk)
23:50, 27 October 2008 (UTC)
I've been doing some template drafting on my userspace
here, and I'm getting some unexpected error messages popping up both there and on the template sandbox (such as
this diff). Now, the template seems to be categorizing as I intended it to, but if I could somehow resolve the error message, I'd be in hog heaven. Many thanks in advance,
Girolamo Savonarola (
talk)
00:08, 28 October 2008 (UTC)
With Win XP SP3 and Firefox 3.0.3, I see it rotated 90 degrees clockwise and tall-stretched, but the full-size image is correct. Just passing this along...
DMahalko (
talk)
04:21, 28 October 2008 (UTC)
Sounds like a caching problem. The image was re-uploaded recently, and sometimes the old dimensions stick around for a bit. --
Carnildo (
talk)
05:18, 28 October 2008 (UTC)
I saw it, but it's gone now. possibly someone made a bad edit on the template, and it needed to work its way through purge? --
Ludwigs204:35, 28 October 2008 (UTC)
For a couple of days the incidence of edits not showing after save has been very high. This means that I have to purge the page to obtain confirmation that the edit was saved ok. I would think this creates a lot of frustration for inexperienced users. __
meco (
talk)
17:19, 24 October 2008 (UTC)
I am still experiencing this. Can somebody who knows why this problem persists please acknowledge this report? __
meco (
talk)
10:07, 27 October 2008 (UTC)
Icons of search engine plugins should reflect language of Wikipedias
Hello: I think that many people like me look for information in different Wikipedias (different editions/languages of it). I believe that not few people, to save time, use the offered (in one of the first code lines, that begins with <link rel="search") search engine plug-ins for the browsers. If Firefox is used each plug-in comes with an icon (in this case a W). The problem is that, as the icons of the different Wikipedias are all the same, it happens to me often that I search in a Wikipedia that I didn't want to look for. I think that this could easily happen to more people. The solution is easy: to put a flag in a corner of the icon: for example the one of Italy for the Italian Wikipedia. Thanks, --
Edupedro (
talk)
22:29, 26 October 2008 (UTC)
But is English a UK, English, or USA language? Is Portugese a Portugese or Brazilian language? What flag would we use for Anglo Saxon? And the flag would be so small, it would be hard to work out what it depicted.
DendodgeTalkContribs22:33, 26 October 2008 (UTC)
Hello: I would go to the origins and use the flag of England for the English Wikipedia and the one of Portugal for the Portuguese version. I imagine that some Australian, ... and Brazilian ... people would prefer to use different flags: the solutions is easy, create a page with the search engine plug-ins (for example
Wikipedia:search engine plug-ins for the English edition) with different flags to choose from. For the Anglo-Saxon Wikipedia "ang" could be used instead of a flag. I've been using
personalized icons for the Wikipedia search engine plug-ins from some months ago and can confirm that the flags are easily seen (and my vision is normal, not excellent) and really help to search faster, in a more comfortable way and not confusing looking for in a different Wikipedia. You can see a similar solution for the Dutch and Nynorsk editions in
Mycroft and for Encarta in
the same web. Thanks, --
Edupedro (
talk)
23:22, 26 October 2008 (UTC)
A person can already restrict searches to a particular language Wikipedia by adding a search term, such as site:en.wikipedia.org for the English Wikipedia. You can even
create a "smart keyword" in Firefox so that you don't have to type as much. (Also, I think that the "W" icons you mention are actually part of the Firefox browser, rather than something provided by the Wikimedia Foundation; if so, it's somewhat pointless to make a suggestion on this page.) -- John Broughton(♫♫)14:44, 27 October 2008 (UTC)
That's not actually true. While Firefox does ship with a Wikipedia plugin, every WP page also uses
OpenSearch to provide a search plugin for auto-discovery by any compliant browser. OpenSearch allows a site to specify a default icon. The request is perfectly cromulent when discussing them. Smart keywords are undiscoverable and are basically deprecated in favour of OpenSearch (and the Add to Search Bar extension, which is due for integration into Firefox in future), and are a workaround for the issue described rather than a solution.
Chris Cunningham (not at work) -
talk15:01, 27 October 2008 (UTC)
Hello: In the past I searched sometimes from Google adding for example site:en.wikipedia.org in the end. I think it's useful, but the search engine plug-ins make the searches faster and more comfortable. I didn't know about the "smart keywords" of FF: I've tried them and find them OK. But if you usually search in more than 20 web sites or pages (like me) I find it difficult to remember so many keywords and when you use 2 or 3 versions of the same web (for example of WordReference or Wikipedia) it can be confusing. So I prefer the search engine plugins: they work with Explorer and Firefox. The first one only admits the OpenSearch format and doesn't show icons, while FF admits both OpenSearch and Sherlock and shows icons, which is really helpful. Every page of en.wikipedia.org has as one of the first lines of code this one: <link rel="search" type="application/opensearchdescription+xml" href="https://en.wikipedia.org/w/opensearch_desc.php" title="Wikipedia (en)" />. It offers the search engine plug-in, located where href says:
http://en.wikipedia.org/w/opensearch_desc.php. If we go to that page and open it with the notepad we can see the icon label (<Image height="16" width="16" type="image/x-icon">), that contains the URL of the icon for the plug-in: in this case
http://en.wikipedia.org/favicon.ico (the favicon of Wikipedia, shown if we enter this URL in the address bar of the browser). My suggestion would be to use instead an icon with something to distinguish it from other Wikipedias (languages). To avoid controversy it could be just the "en" Wiki (abbreviation). And to make it more visual I'd create a page called
Wikipedia:search engine plug-ins with several plug-ins, each one with the flag of England, UK, USA, Australia, ..... so anyone can choose the one which prefers. Regards, --
Edupedro (
talk)
23:57, 27 October 2008 (UTC)
In my opinion, it looks better, but I was told that I should discuss it here before doing it. Although I will not do it if there is not a consensus on it. --
IRP☎18:52, 27 October 2008 (UTC)
I didn't have an opinion until I saw Ilmari's statement; since it isn't a link, I have to agree that it shouldn't be blue. Besides, it's not like many people view the redirects with redirecting turned off...
EVula//
talk //
☯ //20:12, 28 October 2008 (UTC)
Hi. Nearly everytime I create a new page (or redirect) and click save page, I get taken to the "Wikipedia does not have an article with this exact name. Please search for...." screen. It must happen 99 times per 100 over the last few days. Has anyone else being experiencing this? Lugnuts (
talk)
20:47, 27 October 2008 (UTC)
I often see a similar effect when making edits: after clicking "Save page" I am presented with the old version of the page. I am currently using Opera 9.60. —
AlexSm20:59, 27 October 2008 (UTC)
This problem has persisted now since October 24. I mentioned the problem at the IRC channel #wikimedia-tech without anyone there actually acknowledging it, I just got some suggestion about making a change in my preferences to turn off caching. This is a most unfortunate situation. __
meco (
talk)
18:32, 28 October 2008 (UTC)
Sandbox header
Hello.. About the current sandbox.. How does it work (the invisible header)? New mediawiki feature? New extension? This can be used in other wikis which still using conventional-style sandbox. Many thanks.
Borgx (
talk)
04:56, 28 October 2008 (UTC)
all substitution does is hard-write the template code directly into the article. and why would you want it to not display properly? I'm not sre I understand the question. --
Ludwigs204:26, 28 October 2008 (UTC)
Ideally, yes; but I think there are still some bugs that cause certain things to work a little differently with subst, so people do exploit them to make templates that intentionally break if transcluded normally. -
Steve Sanbeg (
talk)
19:15, 28 October 2008 (UTC)
Thanks to you all four your help as to the substing thing, I needed it so a daily thing posted on talk pages could be identified as old and removed, as in
Template:Keg. I was able to use it because the contents are transcluded. However, I still want to be able to transclude sections and boxes on the main page.
Ipatrol (
talk)
00:14, 30 October 2008 (UTC)
strange font on wikipedia
I've been trying to figure this out for awhile. Wikipedia articles display in a weird font that doesn't show up on any other websites that I've found. The font is extremely difficult to read because most of the words overlap (especially problematic overlapping with linked words, for some reason). I use an iBook G4 and Firefox. I've tried changing my default font and size for my browser, with no luck. Does anyone know what font Wikipedia is supposed to use? I'm very confused. Thanks- —Preceding
unsigned comment added by
Trinachi (
talk •
contribs)
What article(s) are you noticing this on? Wikipedia doesn't use any special font; it's either an issue with specific types of articles (such as those that display Japanese characters) or a font issue on your iBook.
EVula//
talk //
☯ //14:48, 28 October 2008 (UTC)
I usually use my iBook G4, and I've never noticed a problem in either Firefox or Safari. if you can point to a page and a section where this occurs I can doublecheck on my machine. --
Ludwigs203:26, 30 October 2008 (UTC)
I am a student studying for an MSc in Telecommunications and Internet Systems at the University of Ulster in Northern Ireland. Wikipedia are difficult to get in contact with so I was wondering if someone here within the wiki community could answer my questions. My research includes analysis into the problem of ‘link rot’, that is, the progressive breaking of links on a website as content is deleted or moved. The web information company Alexa has named wikipedia in their top 10 rankings for global web traffic. I am interested in how you combat web decay and was hoping that you might answer a few short questions in support of this research. I would be happy to provide you with a summary of the findings when the project is complete and assure you that any information provided will be kept confidential and fictitious names used in my dissertation unless companies are happy for their details to be released. Anyone with an answer feel free to email me personally.
How often do you check your website for link problems?
How many people are involved in this work?
How serious do you see the problem of link rot, on a scale of 1-10 (1 being not serious / 5 being you don’t think about it / 10 being extremely serious)
How serious do you see the problem of web decay in general (information becoming out of date), on a scale of 1-10 (1 being not serious / 5 being you don’t think about it / 10 being extremely serious)
Do you use a solution you developed yourself or a commercially available package to tackle link rot/web decay within your organisation?
Additional Information (if possible)
Estimated budget to tackle link rot?
How long has link rot been a concern (if a concern at all)?
I appreciate you taking the time to look at this email and look forward to hearing from you.
Many Thanks
Alistair Shaw
cc: Professor DW Bustard, Head of Computing, University of Ulster
Wikipedia is written entirely by volunteers. We have 47,692,015 registered users, and if anybody finds a broken link, they fix/remove it. As we have so many editors, everyone will have a different opinion on 'linkrot' - we all dislike it, but it's inevitable and different people prioritize things differently. I hope that answers your questions.
DendodgeTalkContribs14:39, 29 October 2008 (UTC)
I've reformatted your request to make it legible and removed the email address (you can expect replies to be posted here rather than emailed).
The budget is zero. As in most things, Wikipedia primarily relies on volunteers to notice and correct broken links. Since Wikipedia is open to editing by anyone, essentially any reader can in principle contribute to this task. In practice, there is a core community of ~10,000 active editors on the English Wikipedia. I don't know how many of them actively try to address link rot, but I suspect that most of them would try to fix it if they encountered it in the ordinary course of their other activities. Though it can be an issue, I would personally say that link rot is only a minor problem on Wikipedia. Most links in Wikipedia go to other Wikipedia pages, and the fraction of links that are external tend to go to relatively respectable and durable sites where link rot probably less common.
Dragons flight (
talk)
14:50, 29 October 2008 (UTC)
I just click on the links on an article. When I find a broken one, I look for a copy on archive.org, or I google for the title or author to find an online copy of the original. The same information can be sourced from several sources, and notable, serious and encyclopedic information (the type of information that we want here) is usually available easily on paper books. I find the problem of link decay as not much serious for this site (3 on a scale of 10, but only so high because we use many online copies of newspaper articles). We try to relay as sources as good as we can, so a lack of online sources just means that someone will have to crack a paper book open to source the relevant information. (now, if books.google.com stopped that would be a problem, because it's tremendously handy for historical articles to be able to check complete versions of books 100 years old with just a few clicks. Idem for scholar.google.com and the online databases of scientific papers in relation to sociological/scientific articles) --
Enric Naval (
talk)
15:40, 29 October 2008 (UTC)
Do you mean sort by order of alphabet, based on local name of language, as political correct standard? If so, what's you opinion about
tally of language order poll? For me, this tally undoubtly pointing, that 2letters sort is better for mayority.
Encyresponse?23:05, 29 October 2008 (UTC)
"Hapsiainen 21:34, 2 Sep 2004 (UTC) The link lists like "Esperanto Español Eesti Suomi French" are awkward for the reader." It is also my opinion.
BartekChom (
talk)
19:47, 30 October 2008 (UTC)
Search malfunction
Does the search box work for anyone else because it's not working for me - it resolves to http://wikipedia.georgemoney.com/search.php?title=test&go=Go
Bremerenator (
talk)
23:15, 29 October 2008 (UTC)
Hey there. I think this has something to do with he following lines in your monobook.js
Is there any way the site notice above can be hidden from a single page? (I assume, with a magic word, just like the TOC can be hidden.) --
Mentisock12:01, 29 October 2008 (UTC)
If you have nothing else on the top row of the page, you can use this:
Thanks but if only some magic word which targeted the site notice specifically existed... that would be a lot more efficient. --
Mentisock12:51, 31 October 2008 (UTC)
That sounds awful. What's the use case for this? If there's a good reason for it, we might look into supporting it. —
Werdna •
talk14:03, 29 October 2008 (UTC)
What I asked for is awful? O_o There could be pages which are better excluded... or is the site notice an absolute necessity? In either case having an option available as an exception not the rule is better than having nothing, no? --
Mentisock12:51, 31 October 2008 (UTC)
The message board system is rather slow. When are they going to get chat? This will allow business to progress at a much faster pace. Message Board : Chat :: Carrier Pigeon : Telephone.
Libro0 (
talk)
20:32, 30 October 2008 (UTC)
En.wp has now been migrated to new servers and a new search backend. As a result, some nifty new features related to searching have been enabled here, including a "did you mean" option. See
this Wikitech-l post for more details. Graham8712:43, 31 October 2008 (UTC)
Yayy! I just saw that a little while ago, and it was a typo I'd made, and I did click on the "did you mean"! I thought it looked a little different/google-y. Excellent stuff.
Franamax (
talk)
12:47, 31 October 2008 (UTC)
In fact, maybe these changes should be written up and linked from the sitenotice for a while? This will be of interest to pretty much everyone who's ever used the search box.
Franamax (
talk)
12:52, 31 October 2008 (UTC)
Because
histories from before January 2002 made with
UseModWiki, the predecessor to MediaWiki, were manually imported into the database, so they got a higher revision ID number than edits that were there before (from February 2002 to September 2002). For example the
edit with revision ID number 1, made in January 2002, was probably one of the first edits made with the new software, but happened long after Wikipedia was founded in January 2001. The next and previous edit links in diffs, as well as the diff links from the user contributions, move by revision ID number, not date. Thus you can get diffs like
this one by going through
the early contributions of User:Jayjg. Graham8713:58, 31 October 2008 (UTC)
It says that there are 214 pages in that category (so does the magic word PAGESINCATEGORY. However, the first page contains 200 of them (or so the page says), but the next page contains 10 of them rather than 14. What is up with that? And yes I purged the cache several times.
Headbomb {ταλκκοντριβς –
WP Physics}
16:24, 31 October 2008 (UTC)
The problem is not that there are missing pages or pages that shouldn't be there anymore. It's that the number reported doesn't match the number of links. It's saying "This pages contain 20 links" while it contains 15 of them. This is not a problem of the page being up to date or not. This is a problem of not reporting accurately what's on the page.
Headbomb {ταλκκοντριβς –
WP Physics}
19:53, 1 November 2008 (UTC)
User creation log entry appearance in watchlist mystery
Last night I noticed a minor edit in a minor article which would be considered spam, i.e. a link was added to an "external links" section which promoted the website of a new user,
Sackrabbit (
talk·contribs). However, I found the linked site to be useful and relevant to the subject so I kept it, simply editing it to remove reference to his sackrabbit.com venture. I informed the user on his wikitalk, and then noticed something: the user creation log entry for Sackrabbit appeared in my watchlist. Is this normal? I did create his talk page, but I don't remember ever seeing a user creation entry appear in my watchlist prior to this. Please point me to information regarding why this has happened, as I searched but was unable to find an explanation. Thanks –
Sswonk (
talk)
17:12, 31 October 2008 (UTC)
Ya, but I gather Sswonk didn't CREATE the user Sackrabbit... that user existed already and had contribs before Sswonk left a note on their talk. I think, anyway. For reference [m:Special/CentralAuth] shows the account was created and unified on en.wikipedia.org at01:07, 31 October 2008 ++
Lar:
t/
c17:21, 31 October 2008 (UTC)
I don't think it is anything to be concerned about (although it is always good to ask!) My experience has been that the listing only appears if the user registers within the time period that your watchlist covers; that is to say, if Sackrabbit registers at 6 AM, you make an edit that adds his/her user or talk page to your watchlist at 6 PM, and your watchlist covers more than 12 hours, the user creation log will show up on your watchlist. If you edit, say, my talk page, I might end up on your watchlist but you won't see a user log entry because it was almost three years ago (and thus probably outside the range of your list.) I also see entries in the deletion, protection and block logs relating to users whose pages I have edited. --Ckatzchatspy17:33, 31 October 2008 (UTC)
Would the user creation log indicate I had created the user if I had done so? Here is a copy of the actual line from my watchlist:
(User creation log); 01:07 . . Sackrabbit (Talk | contribs) New user account
Works that way for me too - if I add a user talk page to my watchlist, the creation log entry shows up too, even if I did not create the user. –
ukexpat (
talk)
17:45, 31 October 2008 (UTC)
Sswonk: you could just do a simple test: go to
Special:Log/newusers, pick any new account (except those created automatically), add userpage to your watchlist and then reload your watchlist. And yes, when a user creates another user the message is different: just scroll through the same log. —
AlexSm18:03, 31 October 2008 (UTC)
This stems from how log entries are recorded by the server: it stores the time, the type of log entry, the user who did the action, the page affected, and the log comment. When a logged action affects a user rather than a page, the user's userpage is stored as the "page affected". This is why, when you watchlist someone's talkpage, your watchlist shows their account being created. --
Carnildo (
talk)
21:08, 31 October 2008 (UTC)
Well it's better than nothing for sure. I can function with that but a page tree feature would be nice to have, so if a dev wants to implement that, cool. But it's certainly not a priority.
Headbomb {ταλκκοντριβς –
WP Physics}
19:55, 1 November 2008 (UTC)
Experienced editors needed...
I'm building a team of editors for the development of a set of country profiles/outlines, including one for each country of the world. This set of pages (which is now well under way) is part of Wikipedia's outline of knowledge, and will serve as a useful tool for browsing country-related subjects and comparing countries. Editors experienced in the creation and use of bots, and editors experienced (or interested in) the use of advanced wikitools like
WP:AWB and
WP:LINKY are especially needed. If you'd like to help or would like to find out more, please contact me.
The Transhumanist00:38, 2 November 2008 (UTC)
Indiscrirminate diff engine? Specific diff text not showing
Please hold your nose at the content dispute involved, for anyone who looks at this. My question is about whether or not DifferenceEngine.php has gone colour-blind here. Looking at
this diff, for me (IE7, monobook, minimal preferences) the entire changed paragraph is red-on-yellow in the pre version, and red-on-green in the post version. This makes it impossible to identify the substantive change, whereas visual inspection confirms that almost all of the text is identical. The actual changes can be found by searching for "successive Canadian governments" (left(pre)-side) and "Liberal governments and constitutional" (right(post)-side) - this is where I warned you to hold your nose! The text exactly re-synchs within 1-200 bytes(/UTF-8 entities) or so.
My question is: wouldn't the diff display normally just highlight the changed text within the paragraph? The text synchs char-by-char both above and below a relatively small text change. Thanks to the dopey content conflict, there are several other diff's in the recent history with the same diagnostic. Why is the diff not pinning down the precise text change? Thanks!
Franamax (
talk)
08:31, 1 November 2008 (UTC)
I don't see any obvious whitespace change. That definitely looks like a bug in the diff engine, in the sense that it should be able to do better than that, even given the limitations of the two-pass algorithm used. I'm not sure why it fails here. —
Ilmari Karonen (
talk)
10:02, 2 November 2008 (UTC)
// Don't try to do a word-level diff on very long linesif(text.size()>MAX_DIFF_LINE){tokens.push_back(Word(text.begin(),text.end(),text.end()));return;}
It ought to be possible to fix the code so that any shared prefix and/or suffix is ignored when checking against the limit. It's a bit tricky due to the fact that you have to ensure the shared prefix ends on a token boundary, but it should be possible. I could try fixing it, but I don't really know C++ very well (just plain C). Of course, even better would be to write a new diff implementation that scales better for long input so that such hard limits can be dispensed with. —
Ilmari Karonen (
talk)
11:30, 2 November 2008 (UTC)
Move "New messages" tag above the title?
Has anyone given any thought to moving the "New messages" tag that shows up when one has new talk page messages to above the page title? That would seem to provide confirmation that one has a real message rather than a user's trying to spoof the message on their user page.
SchuminWeb (
Talk)
18:57, 1 November 2008 (UTC)
Yes, there is a bug filed in bugzilla to that effect. It just hasn't been resolved yet, possibly because it's not a very high priority for the devs. — Carl (
CBM ·
talk)
19:00, 1 November 2008 (UTC)
I may be doing something wrong ... but for some reason {{WikiProject September 11, 2001}} won't nest properly on
Talk:Mychal F. Judge. It does seem to nest on other articles and other project tags nest on this one. I'm guessing it's something with the template but I have not been able to sort out the fix. Any help appreciated.
-- Banjeboi21:51, 1 November 2008 (UTC)
It's a FireFox 3 rendering bug, if you can reduce it to a relatively simple test case you could report it to
[8]. Resizing your window often fixes it, and just hitting the tab key tends to work for me.
Anomie⚔15:03, 2 November 2008 (UTC)
Script checking whether you have edited a page?
Is there a .JS script (which adds a tab to the top of the page called "me" or something) that can show you your edits to a given page, so that you could check whether you had posted on a page, say, and if so, how many times? If you know of such a script, please do tell me! Thanks.
It Is Me Here (
talk)
14:19, 2 November 2008 (UTC)
The link labelled "Revision history statistics" in the page history will give you that information, among many other statistics about the page and the users who have edited it. The statistics about your edits will be accurate unless you are either out of the top 1,000 users who have edited a page, or your edit is outside the first 50,000 edits in a page history. So basically it will work for your purposes 99.999314159% of the time. Graham8715:17, 2 November 2008 (UTC)
I see what you mean now - but I can only see that line when I'm logged out; when I'm logged in, the links disappear. Is there something wrong with my preferences, would you say, or my
css or
js?
It Is Me Here (
talk)
20:47, 2 November 2008 (UTC)
API would definitely be your best bet. I have a script that tells me how long ago my last edit was (I'm thinking about combining this with an auto-unwatch tool). Sounds like you want more information than that though. What, something that looks just like the history tab but has only your edits? —
CharlotteWebb22:14, 2 November 2008 (UTC)
beware of code
function history_filter(){
if(wgUserName == null || wgAction != "history") return;
ul = document.getElementById("pagehistory");
if(!ul) return;
if(location.href.match("&myeditsonly") ) {
document.getElementById("contentSub").innerHTML += ' | <a href="' +
location.href.replace("&myeditsonly=1", "") + '">show all edits</a>';
x = new XMLHttpRequest();
x.open("GET", wgServer + "/w/api.php?action=query&prop=revisions&rvlimit=500&titles=" +
encodeURIComponent(wgPageName) + "&rvprop=ids|flags|timestamp|size|comment&rvuser=" +
encodeURIComponent(wgUserName) + "&format=xml", true);
x.onreadystatechange = function() {
if(x.readyState != 4) return;
rev = new DOMParser().parseFromString(x.responseText, "text/xml").getElementsByTagName("rev");
var difflink = function(r1, r2, a) {
return '<a href="' + wgServer + '?title=' + encodeURIComponent(wgPageName) +
'&diff=' + r1 + '&oldid=' + r2 + '" title="' + wgPageName + '">' + a + '</a>';
}
var oldidlink = function(r, a) {
return '<a href="' + wgServer + '?title=' + encodeURIComponent(wgPageName) +
'&oldid=' + r + '" title="' + wgPageName + '">' + a + '</a>';
}
var radio = function(r, n) {
return ' <input type="radio" value="' + r + '" name="' + n + '" />';
}
s = ""
for(var i = 0; i < rev.length; i++){
var oldid = rev[i].getAttribute("revid"); var bytes = rev[i].getAttribute("size");
var minor = rev[i].getAttribute("minor"); var sum = rev[i].getAttribute("comment");
var ts = rev[i].getAttribute("timestamp").match(/(\d{4})\-(\d\d)\-(\d\d)T([\d\:]+)\:\d\dZ/);
s += '<li>(' + (oldid == wgCurRevisionId ? "cur" :
difflink(wgCurRevisionId, oldid, "cur")) + ") (" +
difflink("prev", oldid, "last") + ")" + radio(oldid, "oldid") +
radio(oldid, "diff") + " " + oldidlink(oldid, ts[4] + ", " +
parseInt(ts[3].replace("0", "")) + " " + ["January", "February",
"March", "April", "May", "June", "July", "August", "September",
"October", "November", "December"][parseInt(ts[2].replace("0",
""))-1] +" " + ts[1]) + " " + '<span class="history-user">' +
document.getElementById("pt-userpage").innerHTML + " (" +
document.getElementById("pt-mytalk").innerHTML.replace(/>[^<]+</, ">talk<") + " | " +
document.getElementById("pt-mycontris").innerHTML.replace(/>[^<]+</, ">contribs<") + ")</span>" +
(minor ? ' <span class="minor">m</span>' : "") +
(bytes ? ' <span class="history-size">(' + bytes.replace(/(\d)(\d\d\d)$/, "$1,$2") + ' bytes)</span>' : "") +
(sum ? ' <span class="comment">(' + sum + ')</span>' : "") +'</li>';
}
ul.innerHTML = s;
if(!rev.length) document.getElementById("contentSub").innerHTML +=
'\n<div class="error">you've never edited this page, or else there is a bug somewhere</div>';
}
ul.innerHTML="scraping from api.php, please wait...";
x.send("");
}
else document.getElementById("contentSub").innerHTML += ' | <a href="' +
location.href + "&myeditsonly=1" + '">show my edits only</a>';
}
addOnloadHook(history_filter);
I got bored and came up with this, let me know if it doesn't work (probably won't in IE), or how it can be made better if it does work. —
CharlotteWebb00:28, 3 November 2008 (UTC)
Another template formatting issue
On {{Current}} the template doesn't seem to fit or size like the other templates do. That is, it seems to bleed off the page (on the right side) rather than wrapping the text to fit the webpage like the other templates do. See
Zeituni Onyango for an example. It's a fully protected template as well so either an admin could format or post {{editprotected}} to the template talkpage. Any help appreciated.
-- Banjeboi01:23, 3 November 2008 (UTC)
What browser and browser version are you using ? As far as I can see {{current}} is a standard {{ambox}} and those are quite thoroughly tested on many browsers. (I don't see your issue with
Zeituni Onyango when using Safari 3.1 for instance) --
TheDJ (
talk •
contribs)
01:59, 3 November 2008 (UTC)
I'm on Firefox at the moment. The other templates on the page (right above it) text wrap just fine but this one doesn't.
-- Banjeboi02:01, 3 November 2008 (UTC)
Correction, I do see it, but only when i make my window very (VERY) slim. The cause is the non-breaking spaces that seem to be part of the text. I suggest you discuss any changes to that on it's
talk page. --
TheDJ (
talk •
contribs)
02:02, 3 November 2008 (UTC)
It also seems a little silly to me. I just manually shorten the standard undo edit summary when I want more space for my own addition.
PrimeHunter (
talk)
17:20, 28 October 2008 (UTC)
For the record, I don't agree with the original change from
WP:UNDO, as I believe that keeping the link's target recognizable is more important than increasing the available edit summary characters by three. And as PrimeHunter noted, it's possible to edit the default message when additional space is needed.
It isn't something that I felt like arguing about (and a Cyrillic character that resembles the Latin letter "U" seems less confusing than a seemingly random punctuation mark), but I'd prefer that we simply revert to
WP:UNDO. —
David Levy21:45, 29 October 2008 (UTC)
Reverted pending further discussion. David, thanks for the attempt with the "Ц"; given the concerns that have been raised, it is best to go back to the easily recocnizable "UNDO" until this can be resolved. Thoughts? --Ckatzchatspy21:55, 29 October 2008 (UTC)
I pay attention to the little things and do a good amount of undoing. And, well, "Ц" confused the hell out of me. No sir, I don't like it. But I do like having 3 extra spaces to type in the edit summary. Therefore, I suggest stealing
WP:U as a redirect to
Wikipedia:Username policy and instead point it to
Help:Reverting#Undo. I guess that would need to go through
WP:RfD. Rgrds. --
Tombstone (
talk)
19:20, 30 October 2008 (UTC)
You know, if you're that desperate for extra characters in the edit summary, there's a "secret" trick that can get you up to 55 50 more of them. Just add this code to
your monobook.js (or equivalent page for other skins):
The only catch is that there's a slight risk of truncation: the actual limit in the
code is 250 bytes, which may not mean quite as many
UTF-8characters. As long as your edit summaries are pure
ASCII, all should be well, but any non-ASCII characters reduce the effective limit. (In principle, if your edit summary is long and has lots of non-ASCII characters, it can get truncated even without using this hack, but the 55-byte safety margin normally makes this less likely.) —
Ilmari Karonen (
talk)
20:50, 31 October 2008 (UTC)
Minor correction: the actual hard limit appears to be 250 bytes, not 255. Anyway, I've written a
safer version of the script. To use it, just add
I use a gadget that displays the current assessment of an article under its title. This works for all the pages I've come across except two that I've worked on recently:
The Singles 1992-2003 and
Tragic Kingdom. Both are Good Articles but The Singles displays start-class status and TK displays B-class status. This isn't a dire problem but I'm wondering if I'm seeing other articles wrongly or if these two articles appear wrongly to others. Can someone help me with this? -- Escape Artist SwyerTalkContributions14:38, 28 October 2008 (UTC)
I've got the same gadget installed, but get the proper coloring. Perhaps try turning it off, saving your preferences, and then back on?
EVula//
talk //
☯ //14:50, 28 October 2008 (UTC)
Did you try bypassing the cache while the gadget was off? (in theory, that would cause a fresh pull of the assessment data once you turn it back on)
EVula//
talk //
☯ //20:08, 28 October 2008 (UTC)
I designed the script in such a way that it should never run into compatibility issues with other scripts, so I doubt that's the problem. (Besides, a conflict with another script would probably crash one or the other, not change the assessment detected.) The two pages you linked to are working fine for me, with both showing up correctly as good articles. Are you still having this problem?
Pyrospirit (
talk·contribs)
02:49, 4 November 2008 (UTC)
I should not have said "conflicts". What I intended was, that if another javascript tool fails to run (for whatever reason), it might be that your browser never reaches some other javascripts. As such it's possible that a broken javascript, can cause other javascripts to fail. --
TheDJ (
talk •
contribs)
03:12, 4 November 2008 (UTC)
Predictive search and redundant articles
I noticed while looking for
Ecology of Minnesota that
Ecology of minnesota (note caps) is also an article, and that it redirects to the proper article. This is in keeping with Wikipedia's policies about
naming conventions. However, when using the search box with the new predictive search feature, both articles come up in the listings (as well as the other case sensitive redirects). This seems a bit redundant; if someone is searching for an article, they would surely want the proper article, not the redirect. It also clutters the search box; it may take up to double the amount of scrolling to get to the right article.
Hmm... not having looked at the suggestion code, I don't know which would be easier: excluding pages in
Category:Unprintworthy redirects or pages that redirect to a title that case-folds to the same value as its own title. Either would have advantages: the first would catch other undesirable suggestions (such as redirects from typos), while the second would also catch redirects to alternatively capitalized titles that haven't been categorized (which is probably most of them). Might be worth implementing both, if practical. —
Ilmari Karonen (
talk)
10:19, 29 October 2008 (UTC)
Keeping the redirects from typos in the auto-suggest is useful. If someone's likely to typo while searching, they're likely to typo while using the auto-suggest feature. --
Carnildo (
talk)
07:52, 30 October 2008 (UTC)
Yes, we do have a better engine, that excludes alternative capitalization redirects and such, sorts suggestions according to number of backlinks, etc.. It will be deployed once we figure out the hardware details on the WMF cluster. --
rainman (
talk)
14:56, 29 October 2008 (UTC)
While we are talking search, can I add a personal gripe. It would be so useful if there was a monobook setting that cause the engine to display Category:xyzs under the entries for xyzs- or a onekey shortcut to place Category: in the search input box.
ClemRutter (
talk)
10:01, 30 October 2008 (UTC)
Where do the bots get their information from? (I mean... I know that ultimately it's the raw RC where it comes from but why does VF receive it from the bots not directly?)
And another question... what's wrong with
this? Why doesn't it want to sort the usernames in the category with {{PAGENAME}}? --
Mentisock10:08, 1 November 2008 (UTC)
For debugging category sorting problems, the following API query is often useful:
[9]. In this case, a lot of the pages indeed have the sort key "Wikipedians interested in political science". Anyway, I've fixed the page linked by PrimeHunter above, so the category should be fixed shortly (as soon as all the pages get updated). —
Ilmari Karonen (
talk)
10:08, 2 November 2008 (UTC)
Is there any centralized list of all WikiProject templates that are placed on the talk page of articles? What I'm looking for is a computer-readable list with an entry for each WikiProject that includes the name of the WikiProject and the name of the template used to mark articles as being a part of that WikiProject. Also, if this doesn't already exist, is there a way to make one without having to write each entry manually?
Pyrospirit (
talk·contribs)
16:17, 2 November 2008 (UTC)
I wanted to update
Template:ArbitrationCommitteeChartRecent to include
User:Thebainer's reappointment for another year. I created a copy of the template
at my sandbox and
updated it accordingly. I'm pretty sure I made no mistake, but the result isn't what I expected. At first the timeline picture just vanished, but after a few minutes it did appear (I suppose the image needed to be generated first), but all the text is missing now
[10]. Did I do something wrong after all? And if not, what can be done to fix this? --
Conti|
✉23:08, 2 November 2008 (UTC)
Clicked on 'Preview' but Page was 'Saved' - Possible Bug?
I have had his happen about (at least) four times. At first I chalked it up to user error, or being cross-eyed or fatigued. However, I just went back to a page I made an entry on and was accused of not signing my comment - which I always try to do (yet this entry was bot-signed). A few minutes ago when I replied, I previewed my reply and I know I clicked preview. However, my comment was saved in its preliminary form (luckily, not before I was able to log my signature). Anyone else had this crop up? It's intermittant. --
VictorC (
talk)
23:44, 2 November 2008 (UTC)
It isn't "disabled". That is a very controversial and heavy parser function extension which has never been installed here and likely never will, see
bugzilla:6455. --
Splarka (
rant)
08:11, 3 November 2008 (UTC)
Tables in poems stopped working
I noticed this problem with
Philetas of Cos. It looked fine a few days ago, but when I rechecked it now, poems containing tables (in order to show transliteration) were preceded by huge amounts of vertical white space. For example, this markup:
Here is an example poem and transliteration:
<poem style='margin-left: 1em'>
{| cellpadding=0 cellspacing=0
| colspan=2 | ξεῖνε, Φιλίταϲ εἰμί· λόγων ὁ ψευδόμενόϲ με
|
| colspan=2 | ''{{transl|el|ISO|Xeîne, Philítas eimí. Lógōn ho pseudómenós me}}''
|-
|
| ὥλεϲε καὶ νυκτῶν φροντίδεϲ ἑϲπέριοι
|
|
| ''{{transl|el|ISO|hṓlese kaì nyktôn phrontídes hespérioi}}''
|}
</poem>
formerly used to look nice, but it now produces this output:
Here is an example poem and transliteration:
ξεῖνε, Φιλίταϲ εἰμί· λόγων ὁ ψευδόμενόϲ με
Xeîne, Philítas eimí. Lógōn ho pseudómenós me
ὥλεϲε καὶ νυκτῶν φροντίδεϲ ἑϲπέριοι
hṓlese kaì nyktôn phrontídes hespérioi
Currently there are several lines' worth of unwanted vertical white space between the "Here is an example poem and transliteration" line, and the poem. The generated HTML contains a lot of <br /> elements that generate that white space. A few days ago, it didn't do that. What is causing this regression? I assume something changed recently with the Wikimedia software? How would I track this down?
Eubulides (
talk)
23:31, 30 October 2008 (UTC)
Thanks; the response there seems to be "don't do that". I
worked around the problem by replacing <poem style='margin-left: 1em'> with <div style='margin-left: 1em'>. I haven't a clue as to why that works around the problem, or whether this "fix" is an advisable one; I don't know what defines the "poem" element.
Eubulides (
talk)
07:28, 31 October 2008 (UTC)
Well knowing brion, I think he will still take another look at some time, but yes, the general idea is: "don't use tables inside a <poem> , it's intended for single lines or blocks of text only." --
TheDJ (
talk •
contribs)
11:54, 31 October 2008 (UTC)
Poem mostly just adds a <br/> to the end of each line, and wraps it in a div, to make it easier to format blocks of text that require wiki formatting. I'd guess the <br/>s confuse the table parsing, but I don't know what else to expect; it seems like they shouldn't be necessary to use together. -
Steve Sanbeg (
talk)
20:25, 4 November 2008 (UTC)
Coincidentally, I created {{quote2col}} yesterday for original and translated texts. I could not get variables to work inside of the <poem>...</poem> tags, so I used the #tag magic word. --——
Gadget850 (Ed)talk - 20:59, 4 November 2008 (UTC)
Thanks. Is there some way to improve the way that quote2col works with wide or narrow margins? If you look at the original, it responds fairly well when the browser's window is narrow or wide: it always maintains a gutter (white space between the columns), and it keeps the paired lines in sync (though it doesn't further indent the lines if they wrap around, which is a minus). In contrast, quote2col doesn't maintain a gutter, and it doesn't use wide margins effectively; and it does an even worse job with indenting. I reproduced the problem with Firefox 3.0.3 but I expect that similar problems would occur in other browsers as you narrow or widen the browsing window.
Eubulides (
talk)
21:18, 4 November 2008 (UTC)
Need to move page from "IParty" to "iParty"; when I attempt to move it, it says that I can't move it to the same page
At first I thought it couldn't be done, but then I saw that the "iPhone" page has that done to it, and "iParty" needs it too since the business name is officially "iParty", not "IParty."
See
Help:Page name#Case-sensitivity and
Manual:$wgCapitalLinks. Page titles and user names are
case-insensitive for the first letter, as
MySQL has problems with sorting database entries based on the first character. For example, go to //en.wikipedia.org/wiki/IPod. You can make the page look like it has a lower-case first letter by adding {{
lowercase}}, which uses a
magic word that tells
MediaWiki to change the page's title when it builds the
HTML that it sends the viewer, but the actual title of the database entry is still "IPod".
Wiktionary has case-insensitive pages by request, seeing as it's a dictionary (see
sun and
Sun), but this causes problems such as breaking the search feature.
A more optimal solution to this has been long-desired; the aforementioned {{
lowercase}} template used to add a tag to trigger a bit of
JavaScript that made the user's browser change the page title after it was downloaded. Someone's
currently working on rewriting parts of MediaWiki to make capitalization more elegant, but this will understandably take a while, as the software's been written since pretty much the beginning under the assumptions that the first letter of any database key is case-insensitive, and that an underscore is equivalent to a space (check out "slowking_Man"'s
contributions, or if you're an admin, try
blocking him). So now you know more than you probably ever wanted to about MediaWiki's character handling. —
Slowking Man (
talk)
03:05, 3 November 2008 (UTC)
This much is intentional to the extent that we want to be able to write a sentence like:
[[Stupid]] is as [[stupid]] does.
Without having to explicitly create a redirect or a piped link. Wiktionary handles it differently, as there it is considered normal for "
Polish" and "
polish" to point to completely different content. I'd rather not to see the same flood-gate opened on Wikipedia. In fact I think it would be better if titles were less case-sensitive, to eliminate the need to manually create a redirect from
least weasel to
Least Weasel, for example. I think the best possible solution will be one which allows the "correct" title to use any amount of capitalization or spacing it wants to, prevent the creation of other articles with titles which are identical except for capitalization and spacing, and make unorthodox links such as
LeASt weASeL automatically point to the correct place. —
CharlotteWebb04:06, 3 November 2008 (UTC)
This is all true, for normal article titles. I still think an exception should be made for articles that begin with non-Latin (especially Greek) letters, because these are usually symbols, and it is quite possible for case to be significant for these. For example
Ω-logic is an entirely different thing from
ω-logic. --
Trovatore (
talk)
08:10, 4 November 2008 (UTC)
When I started writing this edit
[12], the previous edit you'll see by
User:Evb-wiki hadn't been made yet. I should have gotten an 'edit conflict', because Evb-wiki's edit came after I opened my edit window and before I hit the 'save page' button. I.e., edit conflict. Instead, it just added my edit after hers. I hit the back button a few times just to be sure I wasn't imagining things, and in my edit window her edit isn't there at all - instead the last edit was the previous one by Ferrylodge. I hope this makes sense.... Maybe a bug? ~
priyanathtalk04:39, 4 November 2008 (UTC)
Mediawiki does automatic conflict merging on a line-by-line basis. As you wern't conflicting with any lines being edited by them (and you seemed to be entering some whitespace above your comment, mediawiki was apparently happy with it. --
Splarka (
rant)
08:44, 4 November 2008 (UTC)
Looks a lot like a bug in Wikimedia's current
rsvg version: it seems to be ignoring styles specified with just a tag name, like:
path{fill:none;stroke-width:12;}
It seems to be a recent regression, since the old thumbnail for the first version proves that it used to work. I've reuploaded a version that works around the bug, but we should really try to get it fixed. —
Ilmari Karonen (
talk)
09:52, 4 November 2008 (UTC)
"You have new messages" - umm... not really
Anybody else being hit by the orange bar with no new messages to be found? I archived my talk page nearly a half-hour ago, and now every few edits or so, I get the orange bar with no new messages. *shrug*
JPG-GR (
talk)
17:08, 4 November 2008 (UTC)
Would it be possible to add an option to the protection expiry allowing admins to choose whether the protected page should return to no protection or semi-protection? This comes up w/ vandalism targets that temporarily go to full protection, but should go back to semi after a while (I'm coming at the issue from the presidential candidate biographies). Is there a patch already in the works, perhaps?--
chaser -
t02:38, 5 November 2008 (UTC)
Language
Not sure this is the right place to bring this up, but...having a need to look at the Old English Wikipedia for the first time in a long time, I went to
J. R. R. Tolkien, which has an interwiki, and I observed that the Old English link in the "other languages" list was "Anglo-Saxon". Am I remembering wrongly, or did it previously display as "Englisc" (the way that an Old English writer would write the name of his language)? And if it previously was this way, why/when was it changed; if I'm wrong, and it's always been shown as "Anglo-Saxon", is there a place where I could discuss this with a hope of changing it to "Englisc"?
Nyttend (
talk)
01:14, 4 November 2008 (UTC)
I don't know if it's always been that way or not, but one obvious reason to prefer "Anglo-Saxon" over "Englisc" is that the latter's similarity to "English" could be quite confusing to a lot of people who neither speak nor have ever even heard of "Englisc". I suppose "Eald Englisc" or "Engle-Seaxisc" might be possible compromises. —
Ilmari Karonen (
talk)
05:53, 4 November 2008 (UTC)
This is off-topic, but what's the best Modern English translation for the "ang" language code? A few weeks ago, I wrote a
script to translate interwiki languages in the left-hand column into their Modern English equivalents. The list of translations at the Meta suggested that "ang" should be translated into "Old English"; would you suggest something else? Any other suggestions for the script would be welcome. Thanks for your help,
Proteins (
talk)
19:24, 4 November 2008 (UTC)
Could i clarify my second question? How comes there seemed to be a server lag of that time (found it equates to roughly just over 5½ hours)...?
Simply south (
talk)
21:02, 4 November 2008 (UTC)
[
] didnt come out as I intended the first try. While it would be nice, I think, to have the image code separate in editing mode, I'm not sure whethter to report a bug as I cant find any docunetation on using "#"
Sparafucil (
talk)
04:03, 5 November 2008 (UTC)
there's some documentation at
wp:cheatsheet. in a case like this, the easiest thing to do is place the image link after the #, rather than on the line before (i.e. # [[image...]]). the image floats, so it will go to the proper position , and you need to have the numbered list paragraphs together so that numbering will be right. --
Ludwigs204:40, 5 November 2008 (UTC)
OK, so a banner has just appeared at the top of the page -- how do I kill it. I've seen it, and I am aware of the donation requirements -- now I want to dismiss it, permanently. It is extremely large (unlike all previous ones).
User A1 (
talk)
05:50, 5 November 2008 (UTC)
Also, and this probably doesn't belong here, why is the target 6 million, when the annual report prediction for community givings is 3 million?
User A1 (
talk)
05:53, 5 November 2008 (UTC)
See
above. As for the dollar target, hey why not dream? The more money the Foundation gets, the more the Board and Directors can raise their salaries. —
Slowking Man (
talk)
06:36, 5 November 2008 (UTC)
Thanks MER-C. I must have had my eyes shut. Slowking: Why not dream? Because you should be accounting correctly. I don't mean to be harsh, but the goal seems to be an arbitrary value -- accounting should predict the required funds for the project, then aim to raise them. The data shown on the banner, and in the annual report should correspond. If the project only needs 3 mil, then why ask for 6 -- thats just being cheeky.
User A1 (
talk)
As a side note -- I find the notice slogans a little bit slick, as in oily. Also, its a sad day when you Adblock a wikimedia site.
User A1 (
talk)
07:15, 5 November 2008 (UTC)
Any way that page (which is cool, by the way) can go to other pages of past contributions like in normal contribs pages? And as an extra question: will the new site notice suppress gadget work on all sitenotices (like the one we had recently) not just the fundraising one? --
Mentisock10:12, 5 November 2008 (UTC)
I am trying to make my company name into a hyperlink however I cannot figure out how to do so. If I cannot actually put an external link into the text where can I put it - the only place I could find was References at the bottom of the page but I don't think that is correct?
Hi Elaine, and welcome to Wikipedia. Actually, a better place to ask this is on the
Help Desk page. Anyway, if the company has an article in Wikipedia, the format is [[McDonald's]] (or whatever). However, if the company doesn't have its own article, then it may possibly be not appropriate to add a hyperlink - it depends on the circumstances. What's the context? --
A bit iffy (
talk)
18:30, 5 November 2008 (UTC)
I'm not sure if this is supposed to be like this or what, but at least on
this page, the link to "transwiki" leads to a disambig page, though I'm not sure which of the pages on that page is more apporpriate. Anyone think we should do something about this?
ErikTheBikeMan (
talk)
18:59, 5 November 2008 (UTC)
"Support Wikipedia: a non-profit project" omnipresent banner
Any way to suppress this for those of us who don't wish to see this message on every page we visit? If not, when will it end so I can just figure if a long break is a better choice.
-- Banjeboi01:37, 5 November 2008 (UTC)
I was just going to ask the same thing; that banner is obnoxious on an unprecedented level. It needs a hide button that will make it completely hide. -
auburnpilottalk01:39, 5 November 2008 (UTC)
Adding
#siteNotice{display:none;}
to your monobook.css works. That will suppress any other sitenotice as well, of course, but if this is the kind of garbage the slot is going to be used for, I don't want to see it.
Algebraist01:43, 5 November 2008 (UTC)
Anyone else notice that when they rolled that out, the servers got really slow? Is this just temporary while everything re-caches, or is it some sort of (inadvertent?) negative reinforcement - donate enough and we undo it? --
NE201:49, 5 November 2008 (UTC)
Any reason why it has to appear constantly for registered editors? The "hide" button doesn't remove it, unlike other top-of-page messages. (Not that we're "special", but doesn't time served count as a "donation" as well?) --Ckatzchatspy02:07, 5 November 2008 (UTC)
It's like going to the blood bank and being constantly told while you're donating blood that they need your money --
NE202:20, 5 November 2008 (UTC)
I'm imagining it's so slow because the database is overloaded by people looking up Obama/McCain. I agree, the banner is wayyyy too big, I don't know why we can't simply use the former one. ~one of manyeditorofthewikis(
talk/
contribs/
editor review)~
03:02, 5 November 2008 (UTC)
It was taken down until they could better optimize loading times for maximum donation to frustration ratio keep it from killing the servers. --
NE203:05, 5 November 2008 (UTC)
If i click hide, i expect something to be hidden.... I'd rather choose to have google ads for the benefit of wikipedia than this colossal piece of lost screenspace. --
TheDJ (
talk •
contribs)
05:32, 5 November 2008 (UTC)
To join in raining on the parade here, this is quite easily the most obnoxious donation banner ever. Perhaps if the Foundation weren't spending so much money on cross-country moves and over
US$300,000
on the Office of the Executive Director, there'd be more money for seeking grants, buying much-needed hardware, and hiring developers to work on any of the numerous MediaWiki features that have been
requested for years and years, and the Foundation wouldn't have to engage in unprofessional and obnoxious begging so often. —
Slowking Man (
talk)
06:27, 5 November 2008 (UTC)
I agree I think it looks more like a blood donation site. I am not at all impressed in the way in which this years fundraising has been addressed. What about the people who have already donated, must they suffer this glaring at them for every page of wikipedia they visit for the next few months? I agree with the comments above that it is not exactly the most luring of requests and is quite irritating. Can't it be done a little more subtlely? Also whats with the emotional blackmail "we are here when you need us. Now we need you". Having a "Donate Now" in red on every wiki page seems to be rather forceful to say the least.
Count Blofeld12:45, 5 November 2008 (UTC)
Couldn't agree more. Every article I turn to the red block is putting me off, can't imagine how it looks to new visitors and people who don't know how to disable it. If we must keep the banner even when "hidden" can't we at least remove that obnoxious red donate block and keep it as "Wikipedia is a non-profit project: please donate today"
Count Blofeld12:58, 5 November 2008 (UTC)
Just wanted to add my voice/opinion into the mix (because let's face it, you can never have too many). This banner is ridiculous, ugly and a pain in the arse. Thanks so much :D Alex Muller14:06, 5 November 2008 (UTC)
How is this campaign being run? It's obvious that the office who put that two-inch solid blood red chiclet up there has no competent graphic designers, no editors, and no readers of Wikipedia. Why are people who are so out of touch allowed to soil our project with this blot?
Who is this? How do I contact them? —
MichaelZ. 2008-11-05 14:32 z
I just sent the following to donate@wikimedia.org. Perhaps they would understand the problem if a few more participants let them know how this is going over:
Hi.
I'm a Wikipedia editor and administrator who regularly puts in a lot of work on the project.
The new ad banner with the huge blood-red chiclet reading "Donate Now »" is visually disruptive -- it actually makes it difficult to edit. It's accompanied by a "hide" button which doesn't hide it. I've had to develop a custom spam blocker for it, to hide it by adding code to my user style sheet.
It's insensitive towards participants in this project. It makes it seem that the people running the fundraising campaign are out of touch.
Please remove it as soon as possible.
In the future, please get a skilled designer to mess with the look of our project.
I got a reply from Rand Montoya (Head of Community Giving at the Wikimedia Foundation) that he understands and will see what they can do, but no promises. We will see what happens. --
Apoc2400 (
talk)
18:45, 5 November 2008 (UTC)
People don't seem to understand that they can disable the sitenotice in
Special:Preferences under the "gadgets" tab. There is no need for external software, and really no need to use your individual stylesheet page. Sitenotices are rare, so there will be no harm in enabling this in your preferences, as the gadget will be deleted at the end of the fundraiser anyways and there won't be any other sitenotices in the meantime. Also, as a side note, this is what keeps Wikipedia running. So stop whining about it ;-) . -
Rjd0060 (
talk)
15:02, 5 November 2008 (UTC)
Yes but it could be done a lot more tactfully and in rather better taste. How are most editors to know it can be disabled?
Count Blofeld16:48, 5 November 2008 (UTC)
The other problem with using Special:Prefs to disable it is that what if you forget to turn it back in a month or so, and the Foundation uses that space for an announcement (such as board member voting or something) that you probably want to see? I'm not as concerned with having a banner that cannot completely be hidden (think PBS during membership drives) but the button is what really makes this visually offensive. --
MASEM17:00, 5 November 2008 (UTC)
Did you read my comment two above yours? Sitenotices are rare, so there will be no harm in enabling this in your preferences, as the gadget will be deleted at the end of the fundraiser anyways and there won't be any other sitenotices in the meantime. -
Rjd0060 (
talk)
17:18, 5 November 2008 (UTC)
"People don't seem to understand..." Well, why the heck should they be expected to know this? By thought transference? Many of us are far more interested in editing than in technical arcana, and only come here when driven to do so by the sort of situation created by this imposition (where, to add insult to injury, "hide" doesn't actually mean "hide" at all). In any case, the preferences/gadgets solution you suggest doesn't work for me: I just get "Your password is invalid or too short. It must have at least 1 character and be different from your username." (Never have any other password problems on Wp.) I have contributed to most financial appeals in the past but certainly won't be doing so while this abuse of the sitenotices system continues. ---
Picapica (
talk)
17:39, 5 November 2008 (UTC)
People aren't expected to know - what I meant by that was, it has been repeated here, and various other noticeboards, about ten times in the last two days. I was simply suggesting enabling the gadget, rather than mess with .css pages, because the gadget is much easier. :-) -
Rjd0060 (
talk)
19:50, 5 November 2008 (UTC)
I should add that every single time we have this, people talk about how it's going to destroy wikipedia, everyone hates it, it's so much worse then last time (you'd think we must have a blinking popup by now given how it's always worse then last time), etc etc. Yet every single time wikipedia seems to survive. Sometimes the ads is tweaked a bit, but it stays and seems to actually work. I'm not saying there isn't a need for resonable discussion, but all the hypobolics are IMHO over the top.
Nil Einne (
talk)
03:14, 6 November 2008 (UTC)
Red button in collapsed stated
Anyone else think it would be a good idea to inline the div of the red button, when the collapsed state is active ? That would save some considerable and valuable vertical screenspace if you ask me. Should be simple to fix by adding
For some reason, whenever I go to a page whose URL has been URL encoded (that is, special characters like : have been translated into %3A, etc.), instead of the page itself, I get a double ad banner, one expanded, and another collapsed. In case that's not clear:
Is anyone else getting this problem? I'm using Firefox 3.0.3 on Windows XP, and I have the gadget adblocker enabled. This problem does not appear to occur in IE6 (when using IE Tab in Firefox).
Hersfold(
t/
a/
c)21:52, 5 November 2008 (UTC)
Appeal banner's show/hide obscures protection icon etc.
Can the show/hide link in the donation appeal banner be nudged to left, or at least to somewhere where it won't get in the way of the padlock icon, the spoken version icon (and perhaps others)? For example, on
Barack Obama it gets in the way of the padlock etc. (I know I can probably hide the appeal banner in its entirety via my preferences, but that's no good for anons or inexperienced users.)--
A bit iffy (
talk)
15:30, 5 November 2008 (UTC)
We're actively working on this. It's been a problem for years. If you really want an immediate solution, add "importScript('User:Splarka/topiconfix.js')" to your
monobook.js page. If you can wait a couple of days, the fix will (hopefully) be live on the entire site. More discussion can be found
here. Cheers. --
MZMcBride (
talk)
08:22, 6 November 2008 (UTC)
Thanks for that. I don't want an immediate solution for myself so I won't even bother with that script. It's good to hear that there's a general solution on the way. Cheers, --
A bit iffy (
talk)
12:28, 6 November 2008 (UTC)
A solution for this issue is now running, but you might have to clear your browser cache before you notice it. However this again shows that we really should fix this once and for all. --
TheDJ (
talk •
contribs)
22:54, 6 November 2008 (UTC)
Different password for saving preferences?
I just learnt above how to suppress the donation drive banner from the 'Gadgets' menu of the Preferences settings. Now, as I attempt to save, I get an error message stating that "Your password is invalid or too short. It must have at least 1 character and be different from your username." Now, this password easily lets me log on to my user account as it always has, why doesn't it work for Preferences? (Also, the advice given above about disabling sitenotices via the Monobook didn't work, and yes, I did use 'Shift-reload' after saving.) __
meco (
talk)
08:30, 6 November 2008 (UTC)
Are you using Firefox? Someone else mentioned a similar problem at
AN. The answer, provided by
Mr Z Man, is that Firefox automatically fills in the "old password" field, so when you submit the form, it thinks you want to change your password to an empty string. Clear the "old password" field before saving preferences. That said, I didn't have a problem with Firefox, and the gadget works fine.
Gwinva (
talk)
08:41, 6 November 2008 (UTC)
Because the little arrows wrap to the next line and the red button background wasn't big enough to begin with. Also, the button is on a new line, while it could have been easily placed on the same line as the title, freeing up much needed screen real estate.
Shinobu (
talk)
05:22, 7 November 2008 (UTC)
Signbot Sinebot Template
I brought this up with
User:Slakr, the operator of Signbot, several weeks ago. In the past few months the automatic signature which Signbot Sinebot adds to questions on the Ref Desks that are not signed by the questioner has become very large. If the text in general is about 12 point then the autosign is now showing as about 16 point. (I am looking at this in IE; perhaps it is not a problem in any other browser.) Slakr asked for examples and I gave them to him. (Look at almost any question on the Humanities Ref Desk, for example
Wikipedia:Reference desk/Humanities). I have heard nothing since. Just as a user's large signature is annoying to the reader, so are these automatic ones. Has anyone else noticed this? Is anyone else irritated by the obtrusiveness of these long signatures? Is this an easy fix? (I can find no discussion that led to the initial change from same size to large size.) —Preceding
unsigned comment added by
Bielle (
talk •
contribs)
22:11, 4 November 2008 (UTC)
With apologies, I must have missed the further the requests for information. I don't know how to provide a screenshot and my browser is IE 5 I believe. For your information, the {{unsigned}} appearing above also looks just fine to me. However, either the Bot isn't using that template or something on the Ref Desk pages is affecting the font size there. Here follows an example of the coding being used which I have cut-and-pasted from the current Ref Desk page: —Preceding
unsigned comment added by
71.126.229.125 (
talk) 01:30, 29 October 2008 (UTC) In my browser, this appears to be quite different from the one above. Not only is there a lot more coding (<span style="font-size: smaller;" class="autosigned">—Preceding [[Wikipedia:Signatures|unsigned]] comment added by [[Special:Contributions/71.126.229.125|71.126.229.125]] ([[User talk:71.126.229.125|talk]]) 01:30, 29 October 2008 (UTC)</span><!-- Template:UnsignedIP --> <!--Autosigned by SineBot--> ) but "smaller" appears to translate into "larger". Thanks for your help.
៛ Bielle (
talk)
00:18, 5 November 2008 (UTC)
I really don't want to be
brusque, but Internet Explorer 5 is horrendously old and broken (in software terms), and you're opening yourself up to a legion of
malware and security vulnerabilities by visiting any public website with it. IE 5 doesn't support entire swathes of the
HTML and
CSS standards, which is almost certainly why things look so odd for you.
Are you using some sort of corporate workstation or other locked-down device where you can't upgrade your software? If so, I would definitely recommend asking your IT department (or whoever the relevant people would be) to provide you with software released in at least the last decade. If you do have control over your system, please consider installing
Mozilla Firefox or at least a
recent version of IE. If you're stuck on an old version of Windows that recent IE and Firefox versions don't support, and can't upgrade Windows, consider trying a
GNU/Linuxdistribution such as
Ubuntu that will provide you with more recent software. —
Slowking Man (
talk)
06:12, 5 November 2008 (UTC)
I am sorry to have misled you, Slowking Man. My IE (now that hasn't changed) is Version 7, not 5. I don't know if, in the terms of such experts as yourself, this is any big improvement. If, however, it is my browser that is at fault, then there is no fix at the moment. (Not everyone can just whip off and change a whole system, painstaking learned and barely understood even now. We inhabit very different technical worlds, you and I.) I do have Sophos anti-virus running.
៛ Bielle (
talk)
18:03, 5 November 2008 (UTC)
IE 7 should do just fine. You can find probably find a "print screen" button at the top of your keyboard; somewhere above the "delete" button, to the left of the numpad. Press that key. You now have a picture of whatever your screen was showing in your clipboard. Proceed to open up Paint, which you can find in the start menu somewhere. Paste the picture - you can do so by pressing the "Ctrl" button and the "V" key at the same time, or through the menus at the top of the Paint window - and save it, for example on your desktop. Open
this link in your browser and upload the picture you just saved, and give us a link to it here.
Plrk (
talk)
13:27, 7 November 2008 (UTC)
Problem
It's been three weeks or so since my original WP account (Shshshsh) stopped working. Everytime I log in on WP, my window is stuck... I initially thought it was a temporary problem, or maybe some geek had hacked in my account. That's why I created this new account (Shshshsh2). It entangles me a lot. I tried changing my password etc, but nothing helped. The problem still exists, and I thought maybe you have a solution? Maybe there is some way to fix it via admin tools, or any other way like changing my username or whatever? What could be the problem?
Shshshsh2 (
talk)
10:32, 6 November 2008 (UTC)
When I log in (only on WP and only with my original account), after a few minutes online, unexpectedly, my window is stuck and unable to proceed. I mean, I cannot do anything but close it (that too with difficulty).
Shshshsh2 (
talk)
10:56, 6 November 2008 (UTC)
Hey guys I think it works!!! I'm not sure 100%, but as of now, nothing has happened. I ran several pages simultaneously, and it was not stuck (when it usually is...).
Try it and see which script exactly was causing it. Probably popups, since that was changed to use the api.php recently (it can't be changed back, the query.php is no longer available). What operating system and browser/version do you have? --
Splarka (
rant)
18:48, 6 November 2008 (UTC)
Your monobook.js didn't break my browsers (FF 3.0.3 and IE 7.0.5730.13 on winXP) so you can probably use it if you really want to.
Algebraist18:50, 6 November 2008 (UTC)
API is what the popups gadget now uses to access Wikipedia. It previously worked in a different way. It's possible this change was the cause of your problems, especially if you are using an obsolete browser. What version of Internet Explorer do you use?
Algebraist19:13, 6 November 2008 (UTC)
I tried to restore my monobook and it got stuck again. I don't get why it happens. What is specifically the problem with the popups?
Shahid • Talk2me21:45, 6 November 2008 (UTC)
Have you tried the two scripts in your monobook separately? Have you tried the popups without the parameters? Have you tried the popups gadget in your preferences?
Algebraist22:01, 6 November 2008 (UTC)
Well if I can't have the parameters and set my preferences then I am not interested in having them. I'll try to set the second script without the popups.
Shahid • Talk2me22:04, 6 November 2008 (UTC)
An IRC user ran a test for me, and it does indeed seem that popups is broken in IE7. I'll try to find a computer for myself with IE7 installed so that I can look into the problem. That will take some while however. --
TheDJ (
talk •
contribs)
22:51, 6 November 2008 (UTC)
Thank you, TheDJ, it's very kind of you.
But now I'm completely confused. As Algebraist suggested, I tried to check the two scripts separately. I removed the popups and left the Automated PR, and the problem returned -- it got stuck again. Then (I mean, what I have set now), removed the APR and left the popups and it seems to be working properly. How come? Shouldn't the problem be with the popups?
Shahid • Talk2me22:58, 6 November 2008 (UTC)
Well a problem might be that it may take a while before the website uses the "new" monobook.js file. Perhaps if you keep it like this for 2 hours and then force reload again, you will see the problem with the popups. Sorry, loading scripts is a bit of a mess at times. :( --
TheDJ (
talk •
contribs)
23:05, 6 November 2008 (UTC)
I tried again. [While my popups script is set on the monobook,] I tried to install the automated PR. The problem returned. I removed now the APR, and it's gone. So can we be sure the APR is the reason behind that? Do you know why? Were there any recent changes with that script too?
Shahid • Talk2me08:04, 7 November 2008 (UTC)
Template:Random portal component says it was made for portals, but I don't see why it can't be used elsewhere. It picks a quote based on the day of the year.
Template:Randomquote is another, but it isn't too user friendly, and has no documentation. It picks a quote based on the user's number of edits. It should be possible to modify the latter with a truly random number; you might look
here for guidance. --
A Knight Who Says Ni (
talk)
12:09, 7 November 2008 (UTC)
Topicons
If people notice that their userpage topicon has broken, then do not be alarmed. This is due to a change we have made to stabilize the position of the icons (and coordinates) on articles due to the sitenotice (and will only be enabled for that period of time atm). Unfortunately, MANY people do not yet use class=topicon for their topicons, and they even usually have "hardcoded" icons on their pages. I'm currently working on a meta template {{topicon}} which will help people fix this properly, and to prevent any further disruptions by changes in CSS etc. --
TheDJ (
talk •
contribs)
00:22, 7 November 2008 (UTC)
Thank you. Unfortunately, it's rather distracting for my templates to be borked, and I don't have the resources or knowledge to fix them. The sooner the Wikimedia Foundation raises 6 million dollars, the happier I, for one, will be.
CompuHacker (
talk)
02:22, 7 November 2008 (UTC)
I have converted
User:ArielGold's page (
diffview) to use this new {{topicon}}-template. Note that for the duration of the sitenotice, many of the "user talk"-topicons still won't behave nicely. This is because the icons are often of inconsistent size, causing them to easily cross the header line. When the sitenotice is no longer active however, the icons should return to their "original" position, and I can see if I can make further improvements to the template tomorrow. --
TheDJ (
talk •
contribs)
03:59, 7 November 2008 (UTC)
Because they are a hack, and hacks broken in userspace are less important then the same hacks being broken in articlespace where anonymous editors cannot hide the banner. Besides, they are not broken, people are just not using the topicon CSS class, like these elements should.
Actually it's weird... it's broken if accessed normally but the icons are seemingly restored if the page is purged. I tried putting class=topicon but nothing changed... and as an example, the gnome icon you fixed is still seemingly broken on my page. --
Mentisock18:14, 7 November 2008 (UTC)
TheDJ is completely correct here. The primary importance is the article namespace. If these hacks are broken in the User or User_talk: namespace, they can be fixed over time. But both
Barack Obama and
Stuy look correct. And after all, that's why we're here, for the articles. :-) --
MZMcBride (
talk)
16:46, 7 November 2008 (UTC)
Yeah, of course; ideally they all work everywhere though. Why would they break in one namespace not another anyway? --
Mentisock18:14, 7 November 2008 (UTC)
Mostly because the article namespace uses templates like {{featured article}} which can be fixed with one edit, while the User: and User_talk: namespaces generally use in-line code, which requires manually editing each broken page. For example,
my user page. --
MZMcBride (
talk)
18:18, 7 November 2008 (UTC)
Well, i think you are using a hardcoded version of {{userpageinfo}} right? There is no "fixed" version of that. But if the sitenotice is no longer active, all icons will return to their original position (regardless of any changes we are currently making as "interim-solutions"). --
TheDJ (
talk •
contribs)
23:04, 7 November 2008 (UTC)
Ah, forget that...seems like it's back. Did it really go, or was it some quirk for me? (I logged on and off a few times, and checked preferences, and it really had disappeared for a while.)
Gwinva (
talk)
If you are imagining things, then I am imagining the same things, too. (There might be some song lyrics in this.) It did go way for about 20 minutes, and the ugly banner came back. As you have noted, the fix is, once again, fixed.
៛ Bielle (
talk)
06:54, 7 November 2008 (UTC)
This gets asked periodically - and the last flurry that I am aware was around 18 months ago. I don't know if things have changed since. But... Is there any way of getting dates in articles formatted according to the setting in one's preferences? I ask this because I HATE the "December 25, 1999" format. I wasn't a problem when it was "
December 25,
1999" since that got formatted. However most date links seem to be being removed now. So what is the status of date formatting please. --
SGBailey (
talk)
13:40, 7 November 2008 (UTC)
Since the linking is being removed, you will have to deal with the format that is used in articles, whether you like it or not and there is no way around it. There's been a whole host of discussion surrounding it, including
this one which is mentioned as the reference on the
WP:DATE page. I don't think there can be any technical way since dates are now not being linked, i dont think its a big deal whether its December 25 or 25 December, and if thats your biggest worry in life, then i think you'll be fine :-)
211.30.109.24 (
talk)
13:45, 7 November 2008 (UTC)
(several months ago). It is common to have it at 10-40 seconds, every once in a while it goes up to extraordinary levels, usually 2000 to 10,000 thousand seconds, then comes down over a few hours. Try to use little server bandwidth when this is happening so it ceases, that is probably the cause. Clark89 (
talk)
05:42, 2 November 2008 (UTC)
Unless you're running a really fast-editing bot or are an admin deleting pages with large histories (the effect from user renames has been reduced), the effect that one user will have is virtually nil. The lag isn't related to bandwidth (Wikimedia gets 50,000+ requests/sec during peak time), but from the
slave database servers getting out of sync with the master.
Mr.Z-man06:24, 2 November 2008 (UTC)
I could have sworn it said replication lag, it's about page loading. Well, anyways, from what could be found, occasional server lag has been around a long time-the toolserver is the most severe (a month every once in a while), the techs probably already know about it, so one could check with them but it might not be productive. Clark89 (
talk)
03:50, 9 November 2008 (UTC)
Aha! Yes, you got me, I only tested it in monobook... :S That clause was to stop the loop unnecessarily traversing up out of the content area... I'm surprised it causes an error tho, surely it should just not make the early break? Regardless, the important thing is, is there a nice reliable way of making it break at a sensible time? Perhaps it should stop when it gets to the <body> element?
Happy‑
melon00:39, 7 November 2008 (UTC)
Doesn't appear in modern - they use "mw_content" instead. If we took the whole clause out, it would just ride the DOM tree to the top until it got to <html> which doesn't have a parentNode, right? Perhaps this would be both simpler and more efficient (more nodes to analyse, but fewer attributes to obtain as well). Seems to work over at test.wiki. Thoughts?
Happy‑
melon10:25, 8 November 2008 (UTC)
Probably that'd be the best. This is a curious bug: I think what's happening is that it's walking up the DOM tree until it hits the root ("HTML") node, whose parentNode is the document object, which doesn't have a getAttribute() method. Another solution would be to just use element.id instead of element.getAttribute("id"), but I think just leaving out the check entirely would be simpler. In fact, I've
just fixed it. —
Ilmari Karonen (
talk)
18:25, 8 November 2008 (UTC)
Image display problems in infoboxes
In recent days, I've noticed problems with certain images in US geography infoboxes displaying badly. For example, look at the metro area map in the infobox for the
Seattle metropolitan area, or the detailed local map at
Fairview Lanes, Ohio. Unless I'm remembering very badly, the words in larger lettering and certain lines in these images appeared quite obviously until recently: you could easily read the names of the counties in the Seattle area, and the streets of Fairview Lanes (although not necessarily their names) were clearly visible, rather than being nearly invisible strings of dots.
Nyttend (
talk)
15:12, 7 November 2008 (UTC)
I don't think it's related: the image stays in the same spot, even when I move or rearrange or resize or move my mouse over it.
Nyttend (
talk)
15:43, 8 November 2008 (UTC)
I suspect something is on the glitch with our image-rendering software (Magic or Wizard something), there have been more than the usual amount of complaints about images being on the fritz all over en-wiki at the Graphics Lab. Maybe the devs know what's up?
§hep •
¡Talk to me!23:57, 8 November 2008 (UTC)
Section Lines Slicing Through Userboxes
Just noticed this today on my userpage. Under IE-6, the seperator lines for sections are slicing through my userboxes that I have set up on the right side of my page. Under Firefox, there is no problem, the lines stop at the border of the boxen as normal.
ArakunemTalk18:20, 7 November 2008 (UTC)
Yes, this is an issue that has to do with some CSS we are using (for topicons) while the sitebanner is active. I'll see if we can disable that CSS for IE6 and older. Then those users will have broken topicons, but we can only do so much. --
TheDJ (
talk •
contribs)
22:22, 7 November 2008 (UTC)
It appears some of the characters used in this... i don't template? render in an unusual way. Two of them, the long and short "oo," take up a lot of vertical space (see
Thule; notice the unusual spacing - select the first line and you will see what I'm talking about). Also apparently on Macs they render as "o[]o" (see
this discussion page point). Can this be corrected?
Naufana :
talk00:26, 9 November 2008 (UTC)