Wiktionary:Grease pit/2006/December

I've been away for a while, but has something changed with {{IPA}}? It's outputting in a different font from before, which is fine, but it seems to have trouble displaying some accented characters, e.g. ɛ̃. Widsith 11:19, 2 December 2006 (UTC)

I acted on a request from a sysop from another Wiktionary (pl.wikt:, I think it was) who had a much nicer layout for IPA on their Wiktionary. I doubt the extra 10% is the problem, more likely my change (no, I don't see what the problem is, offhand) at: {{IPA fonts}}. Perhaps I was too bold, with what looked like a tried and true fix? --Connel MacKenzie 22:43, 2 December 2006 (UTC)

Well, if I'm not the only person now seeing boxes where there used to be characters, it would probably be better to revert whatever font changes you made, at least temporarily. The size is fine, obviously. Widsith 16:07, 3 December 2006 (UTC)

Are you not? Well, rollback done, +10% kept. --Connel MacKenzie 22:04, 3 December 2006 (UTC)

Ah – that's better! Thanks. Widsith 10:27, 5 December 2006 (UTC)

I have three related questions on tag and category names, which to avoid confusion I'm going to ask in three separate subsections. Hopefully this won't make things more confusing. —scs 00:20, 6 December 2006 (UTC)

How much of an implicit requirement is it that the tag name be the same as the category name? How bad an idea is it if they're different?

(This comes up because I started creating a template {{railroading}} and a Category:Railroading, then discovered that we already have Category:Rail transportation. But I think "(Rail transportation)" is a bit on the long and ugly side for a tag, so for the moment I've got {{railroading}} displaying as "(railroading)" but categorizing to Category:Rail transportation. More on this question below.)

—scs 00:20, 6 December 2006 (UTC)

It's easy enough to handle on the technical side. Your concerns seem reasonable: on the one hand length, on the other hand identity. I guess the only question is, do the two terms refer to exactly the same thing? I can certainly see cases where that's true. You wouldn't label apple pie with U.S. no matter how American it may be, since by that context tag people mean that style of English. So there are some cases where it's not a problem, at least. And so I should probably update {{cattag}} to deal with those cases more smoothly. DAVilla 00:36, 6 December 2006 (UTC)

And on the other hand, perhaps I should rename Category:Rail transportation to Category:Railroading.

It's intriguing that you consider it "easy enough to handle on the technical side" to "deal with those cases more smoothly", although doing so could introduce some new foibles and quirks... (What did you have in mind?) —scs 00:47, 6 December 2006 (UTC)

Template talk:cattag says, "If the name of a category differs from the name of the label, include the category explicitly". This makes sense, and it works.

But, when a sense definition should have two or more tags at once, the recommended thing to do (as far as I know) is to invoke {{cattag}} explicitly:

{{cattag|tag1|tag2}}

But that means that if either tag is different from its category, the caller has to remember to explicitly categorize the article. And this is a nuisance, negating the benefit of linking category and tag using higher-level templates like {{mathematics}}, and is likely to lead to lots of incompletely-categorized articles, because the explicit category link is too easy to forget in this case.

So I think I may have answered my first question: this seems to be a reason why having tag name different from category name is a bad idea. But see below.

—scs 00:33, 6 December 2006 (UTC)

No, that's poorly worded. Include the category in the label template explicitly. DAVilla 00:38, 6 December 2006 (UTC)

I understand that. But when a sense should have two labels, one can't use the "convenience" label templates, right? —scs 00:44, 6 December 2006 (UTC)

one uses cattag, and it works correctly, see below; and having the label name used in the source redirect to the "standard" name and category is fine, also see below Robert Ullmann 09:53, 7 December 2006 (UTC)

In other words (for anyone else reading this): I was wrong, and one can use the "convenience" label template names in {{cattag}} calls. —scs 14:55, 7 December 2006 (UTC)

At points sense 1 we find the invocation:

{{cattag|Rail transportation|british}}

This somehow expands to

(Rail transportation, UK)

And the article is placed into both Category:Rail transportation and Category:UK. What magic is this? How does that parameter "british" get translated into "UK", for both tagging and categorizing purposes? Is there a secret mapping table somewhere? —scs 00:42, 6 December 2006 (UTC)

Cattag calls the individual templates, doing the right formatting magic. So a label template with a differently named category works exactly as it should. So in this case, cattag finds no label template "Rail transportation", so simply categorizes it. Then it does find Template:british, which it calls, which redirects to Template:UK which does the right thing. (And this is why label templates have to invoke {{cattag/equate}} so they will work either way.) Robert Ullmann 09:53, 7 December 2006 (UTC)

Aha. Thanks. I was beginning to suspect that. Part of the problem was that I'd missed the existence of Template:british, and part of the problem was that I couldn't imagine that (say) Template:mathematics could call Template:cattag to do its work, while Template:cattag turned right around and called Template:mathematics to do its work, without it being a horrendous infinite loop. (But perhaps that's another reason why label templates must actually invoke Template:cattag/equate instead.) —scs 14:31, 7 December 2006 (UTC)

Currently Template:slang is categorizing itself in Category:Slang under the asterisk (*) header, which sorts after the slang entries starting with an apostrophe ('). I'd like to change this so it sorts under a space character, as I have done with slang (see the category for what I mean: slang appears first in the Entries list, above such actual slang entries as 'cause). I'm not sure how to best accomplish this (i.e., getting Template:slang to appear directly under slang) since the category tag is provided by the (extremely complex) Template:cattag/equate. Can someone help who knows more about complicated template parsing? - dcljr 16:36, 6 December 2006 (UTC)

Just add another category call (in noinclude tags), sort key overrides the first. Done. Robert Ullmann 10:01, 7 December 2006 (UTC)

Might wanna talk to Connel and DAVilla, too, because that categorizing under * is no accident, it's something DAVilla did explicitly (and at some cost), at Connel's request. (See User talk:DAVilla#Template:cattag again.) If it's not a good idea after all, the right thing to do is probably to undo it, not override it. —scs 14:49, 7 December 2006 (UTC)

I think the idea was just to make it appear first, not that there is anything special about *. It is a good idea, just doesn't work in the instant category. Maybe we should made cattag use a space? Robert Ullmann 14:55, 7 December 2006 (UTC)

My concerns are 1) having the template that is primarily used to fill the category, appear at the top or very close to the top of the category. 2) Consistency. A whole mess of templates use "*" as the identifier - if a "better" character sorts sooner, so much the better. But I don't think a space will pass from template to template consistently. Doesn't "!" sort before "*" and "'"? Is anyone willing to go through all the templates in the template namespace to convert them all? --Connel MacKenzie 10:08, 8 December 2006 (UTC)

The asterisk does have the advantage that it is fairly clear what is going on; space is a little too "magic" I think. (and * also kinda sorta means "special" which is a good thing ;-). For a few cats (like slang) we can just tweak it as we did; don't think there is any problem here. Robert Ullmann 10:27, 8 December 2006 (UTC)

Strangely, the internet seems like a pretty difficult to place to reliably find example usage sentences usage for words. Wiktionary is good for this but it is hardly comprehensive yet. Might it not be reasonable to write a bot like the City/State rambot which fills in such sentences from Google/Microsoft book searches? They typically return whole sentences in computer-readable form and I'd expect it would generally be judged as fair use (single sentence is a small part of a book). There might need to be some sort of disapproval system too where it would go dig for a better fit. Any opinions?Thadk 07:39, 7 December 2006 (UTC)

That's an interesting idea. The problem is that picking good example sentences is hard, even for people. (Many of our existing example sentences are pretty terrible.) You don't just want any random sentence that happens to use the word; you want it to be a truly realistic and demonstrative usage, and ideally you want it to help suggest the meaning of the word as well.

Something I've been meaning to do (and I commend the idea to others who are interested in improving Wiktionary's example sentences) is to be on the lookout for good sentences whenever I'm reading something. Finding a sentence, and going off and adding it to the entry for whatever word it's a good example of, is easy. (Starting with a word you want an example sentence for, on the other hand, and trying to find or construct a good one, is much more difficult.) —scs 14:17, 7 December 2006 (UTC)

Hippietrail (partially?) automated the collection of usenet "citations" for the citations page, but for some reason never did the same for http://www.google.com/search?q=site:gutenberg.org+... His process created /Citations subpages in the main namespace that were initially disliked. I personally enjoy finding a citations tab lit up, these days. (There's a Wiktionary preference for adding the tab.)

I like the separation of the sub-page, specifically so that possibly relevant citations can be sifted through faster. But obviously, the results of those searches is far too verbose, most of the time. --Connel MacKenzie 10:15, 8 December 2006 (UTC)

I think it would be fine to partly automate collecting more citations, adding them to the subpage under some kind of sub-head that identifies the automated collector? But it is going to be very hard to collect decent ones with any automation. Question: do we have real subpages in NS:0? Last I looked, they were just pages that happened to have a / in the title.

On scs's point: recall that a very large number of the OED citations were collected that way: reading books and picking out good citations for the words that appeared, not going out to look for the words. See The Professor and the Madman about William Chester Minor ... Robert Ullmann 10:39, 8 December 2006 (UTC)

• Note: this needs to move to WT:GP. --Connel MacKenzie 17:39, 6 December 2006 (UTC)

(Moved here from the Beer parlour)

Sorry, I've asked this before, but I can't find the page to edit to amend the pull-down menus at the bottom of the page when you are editing. It used to be Mediawiki:Copyrights, I think, but that's not right. Thanks. — Paul G 16:04, 6 December 2006 (UTC)

Ah, I see it's moved to MediaWiki:Edittools. — Paul G 16:10, 6 December 2006 (UTC)

Hm, I've just added characters with ogoneks to the Latin/Roman section, a raised tilde for nasalised vowels to the IPA section, and a new section for Latvian/Lithuanian characters. Only the first of these is showing up. Anyone know what's wrong and how to fix it? — Paul G 16:23, 6 December 2006 (UTC)

I added it to MediaWiki:Monobook.js#addCharSubsetMenu. --Connel MacKenzie 16:39, 6 December 2006 (UTC)

Thanks, Connel. The characters with ogoneks are there for me when I edit, but I still don't see the tilde in the IPA section or the new Latvian/Lithuanian menu option. I've tried reloading the page, logging out and back in, closing and reopening my browser and clearing the cache, all to no avail. Strangely, they are there if I edit MediaWiki:EditTools. — 193.203.81.129 17:29, 6 December 2006 (UTC)

Oops, I wasn't logged back in... I can see the Latvian/Lithuanian menu now, but still no tilde in IPA. Maybe it expects a preceding letter and doesn't appear if there isn't one. — Paul G 17:32, 6 December 2006 (UTC)

Yes, combining diacritics require a preceding letter. See for example how I did the acute accent for Cyrillic vowels under Cyrillic. —Stephen 07:18, 7 December 2006 (UTC)

Sorry to be pedantic, but I don't see a tilde here. So I don't understand what it is you are saying is missing. --Connel MacKenzie 17:38, 6 December 2006 (UTC)

Connel, the tilde is there when you edit the page, but does not appear in the displayed page. That's why you (and I) don't see it. It makes me think that it expects a leading character or else it won't be displayed. (Ah, I've just seen Stephen's comment above - let's see if I can sort something out.) — Paul G 09:36, 9 December 2006 (UTC)

Could we have an English section that includes æ, é, ï, ñ, ü, and their capitals as well as all the AHD symbols, and duplicate a lot of the symbols from Misc. into different language sections? DAVilla 10:36, 9 December 2006 (UTC)

That is what "Latin/Roman" is for - it is meant to cover all symbols based on the Latin alphabet, including English. I don't know why English would need its own section. (For that matter, I don't know why French, etc, have their own sections - isn't everything already under "Latin/Roman"? Or are these intended to make finding the symbols for a particular language easier?) If you let me know what other AHD symbols need to be added, I can add them. There is superscript KH (representing the final sound in Scottish loch). That can easily be typed as ^KH, but it could still be added for convenience. — Paul G 15:27, 10 December 2006 (UTC)

The French question is getting more at my point. Are we dividing by language or script or what? I can see some funny languages like Vietnamese, even though it's based on the Latin characters, really needing their own heading, but Hawaiian? Romaji? What's the difference there? I see two schools of thought on how to divide these: first, put everything a contributor would need under one heading. That would mean the duplication of many characters, as now, which is fine, and hopefully the duplication of many of the annoying "Misc." characters too. Hence my request to move AHD into an English section along with some of Connel's least favourite ;-) characters. Another option is to make it as compact as possible, a short list that can be easily referenced. Separating IPA characters falls into that category, and I think that's a fine idea too, but there can't be too many exceptions like that because a mixture of the two approaches would be quite cumbersome. That is, I'm not really cool with the way this is heading, and thank goodness the numbers are small now.

Another question: How useful are those buttons at top, really? Especially with the included text, it's more cumbersome to use than to just type, so it's really only targeted at newbies I think, and what newbie is going to need to know how to use <nowiki> anyways? I suggest we think about what's most useful there, that is, what we either use most (if anything, a double '' or tripple ''' in isolation would be enough) or what we most often forget how to do, e.g. the format for an IPA pronunciation or an audio file as our whenever-current standards dictate, not just some useless ] text. DAVilla 17:11, 10 December 2006 (UTC)

Um, how did I end up the maintainer of Hippietrail's baby? And what are my least favorite characters? Oh, you mean the general notion of the English language - how it never contains diacritics? Sheesh. I'm not making this stuff up...you know perfectly well my over reactions are directly counter to the abuses of people proposing that borrowed terms are English, when in fact, English normalizes its borrowings into the Roman alphabet? <shrug>

The buttons atop the edit box I find useful; Hippietrail has lobbied against them as "poor user interface" for years now. In WT:PREFS, I've added a search/replace button that is occasionally invaluable, as well as some others that I use a lot - and several that I use rarely or not at all. The MediaWiki default buttons are very difficult to override - I would recommend avoiding that route. Instead, going only part way by adding buttons to the secondary toolbar (which by default contains nothing, and appears simply as more buttons to the right of the default toolbar.) Perhaps I could refine WT:PREFS to allow individual choices of which/what buttons to display there. Would that make it more extensible? --Connel MacKenzie 08:33, 11 December 2006 (UTC)

The Special:Statistics show the job queue, which triggers a null edit on all pages that include a template by transclusion, whenever that template is edited.

There was a proposal a while back for protecting widely used templates. This was either forgotten or considered to be too unimportant.

I'd like to suggest that any template used by inclusion/transclusion (whatever term you want to use for it) more than 99 times be sysop-only protected. I find them from the XML dump - someone else may have a more up-to-date method of identifying them. Does 100 sound like a reasonable cutoff?

--Connel MacKenzie 19:04, 12 December 2006 (UTC)

The numbers should speak for themselves. I'm guessing there are a few templates that are used on as many as half the English pages, a number that have a range on the scale you're talking of, and who knows how many templates that are hardly included at all, hopefully most of them redirects if there are a bucketload. Maybe there's a pretty clear delineation. What are the kinds of templates that you'd want a regular contributor to edit, and the kinds of templates you'd only want a systop to edit? The supertemplates fall into the latter category for sure, but also some well-known templates like {{f}}. I would hope that the magic number is higher than 100 because I could see a user creating an inflection template, for instance, and then 100 pages that use the template, unassisted, and thereby automatically locking themselves out of editing it. Why not base the magic number on something concrete?

You didn't ask, but if you wanted to make a distinction between regular editors and new or unregistered editors, there is a natural lower-bound based on templates that are "in-use" but don't actually exist. You wouldn't want a user to create a template that has a mistake, and then find that they are immediately unable to fix it. Right now such red-link templates are low in number and stats, but may grow as a result of {{context}}, possibly even reaching the hundred mark. Just a heads-up, nothing to worry about for a while. But hey, no harm in looking ahead, right? DAVilla 17:15, 13 December 2006 (UTC)

Moved here from WT:BP#.7B.7Bcontext.7D.7D --Connel MacKenzie 05:06, 13 December 2006 (UTC)

The context template is ready to roll out. This is intended to replace cattag, which I will delicately redirect. The new template accomplishes language discrimination, selective categorization, and some nifty features with modifiers. It was marginally simpler before those additions but definitely more complex now, although in my opinion still more transparent and simpler to use. The documentation is already up and much more extensive than before. Any questions? DAVilla 21:46, 11 December 2006 (UTC)

Yes: Is there a way to make redlinks for brand new categories show? Are you going to create instructions on how to create brand new context tag template tags? (The {{cattag/equate}} thing never ceased giving me difficulty...and never did get documented, did it?) Creating new context tags properly is a fairly rare task, but needs to be simplified from the 'cattag' scheme. (I made it sortof work a couple times, but it was frustrating, to say the least.)

Also: since the documentation for the new template is lengthy (THANK YOU!) perhaps some redirects from the Help: and/or Wiktionary: namespace are in order? Or a brief summary somewhere? Are you planning on going all-out with documenting this version, once stable and accepted? --Connel MacKenzie 05:33, 12 December 2006 (UTC)

There's already some documentation on how to use context within a label template, but I think a separate page would be much more helpful. Any suggestions on location?

As far as going all-out on documentation, string functions would simplify this several fold, so I'm still reluctant to document the code. However, the naming I've switched over to (lang=, context/tag, context/cat) makes more sense. Thus the label templates themselves won't need to change, and any documentation on use can be fully fleshed out. DAVilla 19:49, 12 December 2006 (UTC)

This is the best thing I've seen here in a while! It makes labelling entires in general easier, never mind foreign langauge ones. I'm going to start depreciating old language specific label templates such as {{sco-vulgar}} and use context instead (I made a lot of them myself anyway!). I've also fixed a lot of the other label templates I've created and some others that seemed to be collecting dust in category:Label templates - so that they are now in line with the new {{context}} template.--Williamsayers79 16:27, 12 December 2006 (UTC)

Perhaps. I'm not 100% certain we should jump in right away, but I agree this looks promising. --Connel MacKenzie 17:07, 12 December 2006 (UTC)

I, obviously the clueless one when it comes to these super-templates, added the new context template to the new Russian word "пас," and the word was added to the volleyball category instead of (I guess non-existent) Category: Russian volleyball or Category: ru:Volleyball (too bad the standard around here is to use that second option with the language code instead of the language's name).  So, did I mess something up, or did it not work only because the Russian category isn't built? — V-ball 16:57, 12 December 2006 (UTC)

I really, really, really, really, really wish there were an easy way to tell "context" to show me the category redlink! --Connel MacKenzie 17:07, 12 December 2006 (UTC)

I'm working on this now. Because categories are named explicitly with context, whereas with cattag it was a kind of let's just look and see, I think this will be a reality very shortly. DAVilla 18:19, 12 December 2006 (UTC)

You did it correctly. This appears to be a bug. I'll have to look into it. It's just that we're in the transition from cattag to context. I'll smooth this over now. DAVilla 18:42, 12 December 2006 (UTC)

So much for "delicately". Since context is backwards compatible, but cattag isn't "forward-compatible", I've gone ahead and redirected ahead of ironing out any problems. I guess a note on context that it was in its final stages of development would have been wise. Oh well, no harm done. DAVilla 18:24, 12 December 2006 (UTC)

Erm, move to WT:GP? It shouldn't be announced and redirected while kinks are being worked out. --Connel MacKenzie 18:28, 12 December 2006 (UTC)

I'm not going to try to unanounce it, which is why I went ahead and redirected when I saw it being used. There are a few kinks perhaps, but no bugs that I've found, fortunately. DAVilla 18:42, 12 December 2006 (UTC)

Interesting to watch the effect this is having on the job queue (displayed at Special:Statistics.) --Connel MacKenzie 18:31, 12 December 2006 (UTC)

Things have been settled down for a while now. I just made a very minor addition to Template:context/modifier and saw the job queue shoot up to nearly two thousand. Not too smart I guess. Lots of work to determine the change affected maybe one page. DAVilla 20:33, 28 December 2006 (UTC)

• I'm seeing bugs on entries using "{{obsolete}}" at the moment...you're fixing this, right? --Connel MacKenzie 18:49, 12 December 2006 (UTC)
  • Or rolling it back? --Connel MacKenzie 18:56, 12 December 2006 (UTC)

    Thanks for the heads up. This should have been a result of the change I made at Template:cattag/equate. Unfortunately it would have been very difficult to simulate the transition without actually doing it, and this was a problem I didn't forsee. I've been unable to fix it cleanly, so I've rolled back that change, and I've also put in some other measures to try to prevent {{{subst}}} from showing as a language. DAVilla 19:16, 12 December 2006 (UTC)

    • Thank you for being so responsive. I'll chill out for a while while you fiddle with it. (I still think this should be moved to WT:GP though.) --Connel MacKenzie 19:25, 12 December 2006 (UTC)

      (GP if you wish.) The current version seems stable and I'm not going to make any more changes for now. (I'll test your red-linking off-line.) V-ball's problem above can be resolved by rewriting all of the label templates. I'm not going to try to do this through software again because a large job queue makes it difficult to see the effects of any changes. DAVilla 19:32, 12 December 2006 (UTC)

• One more comment before moving this section to Grease pit:

  I just edited absent. The "Templates used on this page" list is pretty daunting. I assume that's because it is a work in progress. The massive chaining of templates will be "subst:" consolidated when this is stable, right? --Connel MacKenzie 04:51, 13 December 2006 (UTC)

Any updates? --Connel MacKenzie 22:41, 16 December 2006 (UTC)

I've been making tweaks that I think should be innocuous, and without fail someone will message me within a few minutes that something doesn't look right (twice now not counting you), and then I'll have to go clean up after myself. Fortunately this keeps me from doing anything rash.

Right now everything in Category:Label templates is being converted to use {context}. There's absolutely no rush though.

I would still like to move the description on how to make label templates (and I could add a lot more) to a new location. It is referenced by some of the error-catching code. DAVilla 00:48, 17 December 2006 (UTC)

Sounds pretty reasonable. I would really like to see a full Help: page on creating the context templates, if it can't be significantly simplified. To me, a specific context template should either be a redirect, or a backup "independant label" thing for when it is called without "cattag"/"context". Does that match your general approach? --Connel MacKenzie 17:55, 18 December 2006 (UTC)

Yes, precisely, except that the independent labels have to use specific code so that they will work with context as well as independently. I am thinking of keeping Category:Label templates as a place to dump new label templates that are not formatted this way, if it really is too much trouble, so that they can later be converted by someone more knowledgeable. On the other hand context will take any arguments, whether the template exists or not, so is there really a need for that? DAVilla 19:19, 18 December 2006 (UTC)

In terms of learning curve, yes, I think the shorter tags will always be useful.  :-(   So I don't think a mass-conversion to context/cattag will fly. But I've been surprised before. --Connel MacKenzie 07:33, 19 December 2006 (UTC)

I'm not sure what you mean here. I'm not proposing eliminating the redirects. I'm just saying that ''(label)'' doesn't cut it as a label template, or even {{italbrac}}, since these wouldn't be compatible with certain uses of {{context}}. Not too long ago I thought we were using cattag almost exclusively, so the conversion is really from that to context. DAVilla 19:03, 25 December 2006 (UTC)

• What happened to the check for the category of ]? --Connel MacKenzie 01:59, 23 December 2006 (UTC)

Hmm... I had stripped a lot of that out thinking categories would be more explicit now. You could have spot-corrected the problem by creating the label template, but I guess using that as a general solution is asking too much. For when the template doesn't exist, the check is back in, as per the documentation. Where wasn't it checking that it should have, and is it fixed now? DAVilla 19:03, 25 December 2006 (UTC)

If most of the bugs are now worked out, then it would be very helpful to have examples for various cases of use in the documentation, specifically for those where the category and tag do not match. For instance, how would one ensure that azure is tagged (heraldry) but categorized in "Category:Heraldic tincture"? --EncycloPetey 07:28, 27 December 2006 (UTC)

Those are extremely rare cases. See Template:star for an example. DAVilla 17:35, 27 December 2006 (UTC)

I discovered commons:Category:CJK_stroke_order and made a small template to link to their images. The intent is to capture the most common case in a simple-to-use way. I stuck an example of it on 手. Does using this template on a large scale seem like a good idea?

If so, could someone instill some useful template magic in it? Particularly adding optional parameters for image size and exactly which image to use (there are some characters with different stroke order in Japan and China).

Cynewulf 07:50, 14 December 2006 (UTC)

Partly done, see {{stroke order}}, alt type for the Japanese stroke order soon (minutes? ;-) Robert Ullmann 18:03, 25 December 2006 (UTC)

There seems to be a problem with the {third-person singular of} template; if you provide a usage example below, it starts a new numeric list. See the example in wells#Verb and compare to wells#Noun. --Jeffqyzt 20:48, 19 December 2006 (UTC)

Did someone fix this? 'Cause I don't see any problem. --EncycloPetey 07:18, 27 December 2006 (UTC)

Not in relation to this bug report, yes, I routinely annihilate extraneous <CR>s in templates, as I find them. {{third-person singular of}} seems to have been one such case. I have not systematically reviewed this class of template, so there may be more similar errors floating around. --Connel MacKenzie 01:53, 28 December 2006 (UTC)

Wouldn't it be easier to just rely on {form of}? Simpler templates make it less likely to introduce errors. DAVilla 02:26, 28 December 2006 (UTC)

Very few (if any?) of the original templates have been converted to use {{form of}} yet. If you could convert them (since you are the one that understands the subtlties of it, the best) I'd appreciate it. --Connel MacKenzie 16:23, 29 December 2006 (UTC)

OK, I give in - what is this new number (+nn) in the recent changes listing? SemperBlotto 22:38, 19 December 2006 (UTC)

It's a new feature of Mediawiki, it shows the size in bytes of the change that was made. --Versageek 22:46, 19 December 2006 (UTC)

WOW! How long has this been available? --Connel MacKenzie 05:00, 20 December 2006 (UTC)

It even works on the really complicated templates, showing all the comments if you uncheck that option. Finally, a way to debug a live template without altering the output! Not that we would EVER need to debug templates live, that is. Our work is always without exception bug-free! :-P DAVilla 18:33, 25 December 2006 (UTC)

Yes, I certainly have a long history myself, of goofing up templates. I can't wait until there is an XML dump version that has templates iterated to completion. (Man, what a pain that is, trying to parse wiki-templates offline!) --Connel MacKenzie 07:46, 26 December 2006 (UTC)

I brought this topic up at the Beer parlour, and Connel Makenzie said i should discuss it in here.

The original discussion can be seen Here.

My idea is to basically organize the a translation project, into individual languages, with specific organized tasks.

My example page I've worked on a tad can be found Here. Currently at the stage its at it still needs to made more attractive, as well as be translated into spanish too.

Please give me some input on what you all think of this idea, it would be very helpful.

Bearingbreaker92 01:38, 24 December 2006 (UTC)

Where would you like to receieve input? I'd make that a link on the project page, even if it will simply be the talk page. I can (and do) help a bit with Spanish, but I'm working more heavily on Latin words and tidying the category structure of languages. --EncycloPetey 07:15, 27 December 2006 (UTC)

Leave input on here, the page that I linked to isnt really an official page or anything set in stone. Everything that is on there is subject to change, or move, so it would be safer to have it on here.

Bearingbreaker92 04:23, 28 December 2006 (UTC)

We kinda sorta don't want to have a long discussion of details here; I have a pageful of comments ;-). We do have a small number of projects, a few ongoing, with semi-standard names: I've taken the liberty of moving your page to Wiktionary:Project - Spanish ... that gives it a discussion page at Wiktionary Talk:Project - Spanish. (we can always refine this later) Robert Ullmann 04:45, 28 December 2006 (UTC)

(which is not to cut off any discussion anyone wishes to continue here ;-) Robert Ullmann 04:45, 28 December 2006 (UTC)

I think that it would be good if we had a status of Trusted user. One who, although not a Sysop has a reasonably long history of making good contributions. His edits would be automatically marked as patrolled, relieving us of one more mindless task. If it could be implemented, we would probably have to implement some sort of voting system - so would it be more trouble than it is worth? SemperBlotto

Do you use WT:PREFS? Connel built a 'whitelist' into the patrolling enhancements... --Versageek 10:08, 29 December 2006 (UTC)

Actually, I'd prefer to move away from my method, towards the allocation of a new class of trusted users. There are plenty of bugs in my Whitelist thing (e.g. User:16@r ah ha! URL encoded, of course!) and it is currently dependent on a sysop being around, with the whitelisting turned on, viewing 49 most recent edits, to refresh every couple minutes. It is a kludge atop a hack atop a work-around. So exploring other possibilities is a Good Thing. --Connel MacKenzie 16:20, 29 December 2006 (UTC)

(1) Trusted users would include some anonymous IP addresses. This would not apply to dynamic ones and would only be held while the address is consistently used, or otherwise there would have to be occasional review.

(2) Trusted users need not be voted in, providing the criteria are strong enough to determine that the user makes good contributions rather than long-term shady ones. Maybe it would be enough to have a period where a notice of eligibility was given after marking patrolled edits, and status granted automatically unless there were objections. DAVilla 16:33, 29 December 2006 (UTC)

Thanks for reminding me: currently very (VERY) few other sysops have (rarely) edited the whitelist. So it amounts mainly to a list of editors that I have happened to notice making lots of good edits. Certainly the community as a whole shouldn't condone *my* opinion involuntarily. Having people voted into the "trusted users pool" (perhaps from three other trusted users or sysops, or something) would be a much more Wiki-way to approach it all. --Connel MacKenzie 16:38, 29 December 2006 (UTC)

I don't like the idea of trusting anonymous IP addresses. Sometimes these are shared by good and bad users, and we need to look at each edit. I would be quite happy for a single Sysop to be trusted to set and unset the flag (as long as the change gets flagged openly). Though would we want the user himself to know that he was trusted? SemperBlotto 16:47, 29 December 2006 (UTC)

They already do, by merit of their edits appearing as patrolled or not patrolled. (Log out and check - non-sysops still have the "hide patrolled edits" button.)

Of greater concern is that I have not been diligent about checking the dynamic nature of IPs currently in my whitelist. I've tried, but I certainly have not been systematic about it.

--Connel MacKenzie 16:53, 29 December 2006 (UTC)

For IP addresses, perhaps it would make sense to trust them for up to thirty days at a time (depending on the ISP information available)? Actually, there are some IPs that are static.

Perhaps I should revisit additional JS enhancements (mark all contribs by this user/IP as patrolled; mark all edits to this page as patrolled; mark all edits to a page as patrolled if a sysop has since edited it.) I also should look at making it compatible with the MW built-in RC enhancements, instead of "simple mode only" as it is now. --Connel MacKenzie 20:19, 30 December 2006 (UTC)

I notice that neither the WT:COW nor the WT:TOW rolled over this morning to the new week. Is this a result of having the New Year on a Monday, or is something broken? --EncycloPetey 12:29, 31 December 2006 (UTC)

Is WMF using the wrong boundary for the start of the week? Sunday 23:59:59 instead of 00:00:01? I guess we'll know in a little less than two hours. --Connel MacKenzie 22:12, 31 December 2006 (UTC)

{{CURRENTWEEK}} rolls over on Monday UTC (or server localtime for some of the european- or other language wikis where default local time is set differently). As far as I know this is correct (eg w:ISO week date) though there might be multiple ways of counting week numbers (as there are multiple local preferences for when "the new week" starts in lay terms, Monday or Sunday etc). --Brion 23:18, 31 December 2006 (UTC)

Indeed, we are now on week one, having skipped "Week 53" entirely (which, I guess was supposed to appear for one full day.) --Connel MacKenzie 00:11, 1 January 2007 (UTC)

At the top reads: The following data is cached, and was last updated 23:18, November 30, 2006.

Chatting in #wikimedia-tech, the refresh has been disabled, pending the purchase of more DB servers. Time for {shudder} a fundraiser? --Connel MacKenzie 06:00, 13 December 2006 (UTC)

Ahhh, what blissful irony. --Connel MacKenzie 09:36, 2 January 2007 (UTC)

Archived at Grease_pit_archive/2006/October#Tooltips_for_standard_headings

If I'm in the wrong place, please forgive me. It is not made plain where I should continue/restart this thread. Looking at the original discussion, I'm not sure if this shouldn't be a Bug Report.

I use Opera 8.5 in Windows 98. On any word definition page, when my pointer hovers immediately above or below a section heading, anywhere across the page, I get a completely pointless tool-tip:

• L2 heading and TOC: "Title:The language this term belongs to"
• L3 and deeper: "Title:'''' is not a standard Wiktionary heading"
• L4 ("Synonyms" only): "Title:'undefined' is not a standard Wiktionary heading"

While this... "feature"?... appears to be harmless, it is ^*%(* annoying. Is there any way to disable it in the Wiktionary preferences? Or even better, just not have it at all: I do not believe it will achieve the stated object of helping newbies, who should be encouraged to print and USE an editing cheat-sheet; and all it does for experienced users -- OK, me -- is annoy.  Gordon | Talk, 06:18, 2 January 2007 (UTC)

The same thing happens using Safari on a MacBook. There should be a defined list of standard headers at various levels someone, but it doesn't seem to be working right. --EncycloPetey 15:10, 2 January 2007 (UTC)