Wiktionary:Grease pit/2012/June

Unifying some of the 'catboiler' templates?

Lately I've been wondering whether some of the category boilerplate templates such as {{poscatboiler}}, {{etymcatboiler}}, {{lexiconcatboiler}} and so on could not be unified into one single template. From a practical standpoint it would be easier to decide which of the templates is needed for any particular category because they would all use the same one. This means the barrier to entry is lower for new editors which is a good thing. From a technical standpoint there aren't really any problems either, this could be very straightforward to do because the underlying templates for all of them are the same. One possible downside would be that all the subtemplates for every category would now be lumped under a single main template, making it a bit harder to get a good overview of which goes where. A possible solution to that would be to add them to categories. So... would this be desirable? —CodeCa t 12:10, 2 June 2012 (UTC)[reply]

I've considered the same thing, from having used the wrong template in the past, for example {{lexiconcatboiler|fro|borrowed terms}} when it should be {{etymcatboiler}}, if you unified them and then used redirects, this sort of mistake would no longer be possible — This unsigned comment was added by Mglovesfun (talk • contribs) at 12:26, 2 June 2012 (UTC).[reply]

I don't think we would be able to unify all templates though. {{derivcatboiler}} and {{pbcatboiler}} for example would have to stay separate for technical reasons, as would {{topic cat}} which uses an entirely different system. And I think it may be good to keep {{tempcatboiler}} separate as well for now. There is just one prerequisite for this to work: the labels used need to be unique. As far as I know we already do this, but there is one point of conflict: the 'numbers' label is used by both {{symbcatboiler}} and {{poscatboiler}}. I suppose this is a good time to settle the old numbers-or-numerals question? —CodeCa t 14:48, 2 June 2012 (UTC)[reply]

I haven't really been able to think of a good name for such a unified template. Maybe {{lexicalcatboiler}} or {{lexcatboiler}}? (I am also thinking it may be nice to change 'catboiler' to just 'cat' with a space, like in {{topic cat}} but that's another matter) —CodeCa t 13:15, 4 June 2012 (UTC)[reply]

Cat boiler sounds a bit cruel, I'm an ailurophile myself. Mglovesfun (talk) 08:52, 6 June 2012 (UTC)[reply]

Bwahahaha...yeah I've always thought of the idea of boiling a cat when I read catboiler. xD 50 Xylophone Players talk 17:32, 18 June 2012 (UTC)[reply]

If possible the new template will take over the functions of the following existing templates: {{charcatboiler}}, {{compcatboiler}}, {{etymcatboiler}}, {{lexiconcatboiler}}, {{morphcatboiler}}, {{namecatboiler}}, {{phrasecatboiler}}, {{poscatboiler}}, {{shortcatboiler}} and {{symbcatboiler}}. So is {{lex cat}} general enough a name to cover all of these? —CodeCa t 17:45, 18 June 2012 (UTC)[reply]

Also, what should be done with the conflict in 'numbers' mentioned above? —CodeCa t 17:45, 18 June 2012 (UTC)[reply]

Russian Wiktionary Android version

Entry ru:airplane. Screenshot of the open-source Android software "kiwidict-ru 0.091" with data extracted from the Russian Wiktionary.

kiwidict-ru is offline multilingual dictionary and thesaurus based on the Russian Wiktionary. Welcome to play and test this dictionary. Your feedback is welcome!

An android version of the English Wiktionary is coming... :) -- Andrew Krizhanovsky (talk) 16:48, 2 June 2012 (UTC)[reply]

Proto-Austronesian

I've just spent most of my weekend killing hundreds of bogus {{proto|Oceanic}} uses for non-Oceanic languages (see WT:RFC). In the process, I replaced a lot of those with {{proto|Austronesian}} where the etymology has been taken that far back. This, in turn, has put a lot of redlinked [[Category:Xyz Terms derived from Proto-Austronesian]]s at the bottom of a lot of entries.

I would like to create those categories, but we don't seem to have a language code for Proto-Austronesian to feed to {{derivcatboiler}}. It's true that the language family has historically been know as Malayo-Polynesian, but the addition of the Formosan languages has made that name a branch of Austronesian (albeit one with the vast majority of the speakers for the grouping), rather than the correct name for the grouping as a whole.

Do we have a designated poz-type language code for Proto-Austronesian, and, if not, can we create one- and all the template mechanics that go with it? Chuck Entz (talk) 14:11, 4 June 2012 (UTC)[reply]

We do have a code for it: {{langrev/Proto-Austronesian}} —CodeCa t 14:24, 4 June 2012 (UTC)[reply]

Thank you! By the time I got to this after work, Amir Hamzah 2008 (talk • contribs) had already been busy creating categories for various Austronesian languages- which were all causing {{proto}} to puke its subtemplates all over the place. Fortunately {{proto:map-pro}} was simply lacking a script subtemplate, and I added it after looking at how {{proto:poz-pro}} was configured. Now Category:Terms derived from Proto-Austronesian has 41 subcategories, and the {{proto}} template works fine with all of them. Chuck Entz (talk) 12:43, 5 June 2012 (UTC)[reply]

Pywikipedia bots

I keep having more questions about how to run a pywikipedia bot on OSX, so when I think of them, I'll ask here. First of all, should I add the line to user-config.py about API (use_api_login = True)? --Μετάknowledge^{discuss/deeds} 18:01, 5 June 2012 (UTC)[reply]

Category:Entries which need grc script

I made this change to {{rfscript}} to change this category to Category:Entries which need Greek script. It should work, shouldn't it? The more I look at it, the more it looks ok, but clearly, it isn't ok. Mglovesfun (talk) 18:35, 7 June 2012 (UTC)[reply]

Fixed. —Ruakh_TALK 19:43, 7 June 2012 (UTC)[reply]

I never noticed this category before, so I'm cleaning it out now. --Μετάknowledge^{discuss/deeds} 20:25, 7 June 2012 (UTC)[reply]

Weird bit of code, seems we should use {{uc:}} or {{lc:}} if anything, that way whatever case a user uses, it will work out ok. Mglovesfun (talk) 20:29, 7 June 2012 (UTC)[reply]

Weirdnesses in Template:reference-book

Why does this wikilink the accessdate value? There is zero utility in linking through to 2007, or to June 7. We don't have, and (so far as I know) don't want to have any such entries in Wiktionary.
Why does this template's display end in a colon? So far, I have only seen this template used in reference lists, when it is followed by nothing -- and as such should end in either a period, or nothing.

Curious if anyone has any insight. -- Eiríkr Útlendi │ Tala við mig 21:09, 7 June 2012 (UTC)[reply]

I could be wrong, but I thought that this was deprecated in favor of {{quote-book}}. I suspect nobody's upkept it. --Μετάknowledge^{discuss/deeds} 21:35, 7 June 2012 (UTC)[reply]

Essentially the reason is because nobody's fix it yet. 21:36, 7 June 2012 (UTC)

Ruakh had some additional info over at Template talk:reference-book. I don't want to use {{quote-book}} because I'm looking for something that's just for a bibliographic reference, without all the formatting that {{quote-book}} entails, and often without any quote. But yes, Ruakh suggested much the same thing -- this was probably copied from Wikipedia without ever being fixed for Wiktionary purposes. Ah, well. :-/ -- Eiríkr Útlendi │ Tala við mig 22:40, 7 June 2012 (UTC)[reply]

Python corpus searches

I'm looking into the possibility of creating an Ancient Greek inflection bot. This bot will differ from existing ones in that it will only create inflected forms which it can cite from existing works. However, I'm not exactly sure how to accomplish this. One option I considered was to simply go through works one by one, check each word, figure out what its lemma form might be, then check to see if that lemma exists on Wiktionary, check if that existing lemma does in fact have the inflected form in question, and then write the inflected form. Trying to extract a lemma from an inflected form is tricky enough, but this approach just seems incredibly inefficient, as most lemmas won't exist here. Another approach I considered was to read through all available works ahead of time, and create a database of words in them, which could then be checked when creating entries. I'm not entirely sure how such a database would be structured. The simplest would of course be simple text document(s), but I suspect that SQL would get me rather better performance and flexibility, even if it might be a little trickier to code. Right now I'm sort of leaning towards the latter option, but I really have no idea what I'm getting myself into, having never done anything like this before. If anyone has any thoughts, tips, links, etc., I would be quite grateful for them. -Atelaes λάλει ἐμοί 00:57, 8 June 2012 (UTC)[reply]

If you only need to be able to check whether a word is attested or not, a database might work, but you can also use a word list that's alphabetically sorted, and use binary search (if you know how that works). —CodeCa t 01:26, 8 June 2012 (UTC)[reply]

Hmm....that's not a bad idea. I'll keep that in mind. -Atelaes λάλει ἐμοί 11:00, 8 June 2012 (UTC)[reply]

Request: French plurals ending in -ss

Could anyone find all of the entries in Category:French plurals that end in -ss? Some of these may be bot mistakes. There might be a couple of valid ones however, the French Wiktionary suggests business for example (plural of business). Mglovesfun (talk) 14:24, 8 June 2012 (UTC)[reply]

There are only 4: edelweiss, infarctuss, Paloiss, plâtrass. 70.162.10.166 23:11, 8 June 2012 (UTC)[reply]

Job done. Mglovesfun (talk) 10:22, 9 June 2012 (UTC)[reply]

Note that we don't yet have business#French. --Μετάknowledge^{discuss/deeds} 18:15, 9 June 2012 (UTC)[reply]

Also fitness, which fr:fitness list as 'singular and plural', but I doubt the plural is ever used. Mglovesfun (talk) 19:40, 9 June 2012 (UTC)[reply]

Except it is used. Mglovesfun (talk) 19:44, 9 June 2012 (UTC)[reply]

business#French added. Jamesjiao → ^{T ◊ C} 23:12, 18 June 2012 (UTC)[reply]

γράφω

The declension templates on γράφω display things like {{{4}}} and {{{6}}} for me. Is anyone else getting this / any ideas on how to fix it? It Is Me Here ^{t / c} 16:09, 10 June 2012 (UTC)[reply]

I can't conjugate in Hellenic languages to save my life, but this is probably a sign that sombody didn't put enough arguments in when they called the template. {{{4}}} and {{{6}}} are the 4th and 6th arguments (bits seperated by pipes ( | ) that come after the first pipe), necessary, I suppose, for complete conjugation, which the author neglected. If you know what they ought to be, you can add them in, or get somebody else to do so. --Μετάknowledge^{discuss/deeds} 17:13, 10 June 2012 (UTC)[reply]

I've added some of the missing forms, but parameters {{{4}}} and {{{5}}} in the perfect table can't be added because they should be different things in different contexts. Basically, {{grc-conj-perfect}} isn't equipped to handle this verb yet. —An gr 17:57, 10 June 2012 (UTC)[reply]

I've raised the issue at Template talk:grc-conj-perfect#Problems with the template. —An gr 18:24, 10 June 2012 (UTC)[reply]

For verbs like γράφω the Template:grc-conj-perfect-labial is the right one, I guess. --flyax (talk) 19:57, 10 June 2012 (UTC)[reply]

Great! But {{grc-conj-perfect}} is still being used on many consonant stems where it should be replaced with more specific templates like that. —An gr 20:05, 10 June 2012 (UTC)[reply]

Plural for `{{te-noun}}`

Is there a way to include the plural mechanism into the template? Rajasekhar wants me to help out with it. --Lo Ximiendo (talk) 15:37, 13 June 2012 (UTC)[reply]

Done. Can you check that it works? (By the way, I also made WT:TE TR and wrote the template's documentation.) --Μετάknowledge^{discuss/deeds} 16:38, 13 June 2012 (UTC)[reply]

Yep, it works. --Lo Ximiendo (talk) 06:22, 15 June 2012 (UTC)[reply]

Template talk:cardinalbox#Bug

Mglovesfun (talk) 10:20, 15 June 2012 (UTC)[reply]

Fixed. (See there for details.) —Ruakh_TALK 23:11, 15 June 2012 (UTC)[reply]

A template for linking English terms to their definitions

Right now we have a few established templates like {{term}} and {{l}} to link to other entries. But there is one place where we still customarily use raw wikilinks, and that is when linking to English entries in a Wikipedia-style fashion to explain the meaning of an English term in running text, for example in a definition of another word or in a usage note. I have created a template for this purpose now, at User:CodeCat/d. The d stands for 'definition' because it is used to provide a link to a definition to a term. The template itself is very simple even compared to {{l|en}}: it always links to the English section. The advantage is that it would be a bit shorter to type, but it still links to the proper section on the page, which is good for a usability point of view, especially in combination with tabbed languages. It also means we no longer need to distinguish cases where the definition is on the very same page, such as when water in Dutch is defined as water in English; normally the link would turn bold black. Another advantage may be that it becomes easier for bots to recognise definition-style links if a separate template is used. I can't really think of any disadvantages except that people would type {{d|something}} instead of [[something]], which is not really a major change. What do you think? —CodeCa t 17:36, 16 June 2012 (UTC)[reply]

Well, I type definition wikilinks a lot. Honestly, two extra characters, plus the fact that 4 of the characters need shift ( [ to { ) would be really annoying. In any case, you'd have to choose a new letter, as there's already a {{d}}. I suppose this would come in handy for limited use, and extended use in certain specific contexts, but I strongly oppose any attempt to replace the standard [[]]s in definitions with a template. --Μετάknowledge^{discuss/deeds} 04:20, 17 June 2012 (UTC)[reply]

Oppose strongly. Standard linking suffices, we don't need this, which (besides Metaknowledge's objections) eats resources. (If the issue is that Tabbed Languages can't handle standard links, fix TL.)—msh210℠ (talk) 07:17, 17 June 2012 (UTC)[reply]

Why do you think it suffices? I don't think it does for the reasons I mentioned above. And TL can't be fixed in this way because it has no way of guessing which language a link is supposed to link to... not without using vastly more resources than this solution. —CodeCa t 11:41, 17 June 2012 (UTC)[reply]

The explanation you give above for its insufficiency are (1) that then "we no longer need to distinguish cases where the definition is on the very same page, such as when water in Dutch is defined as water in English" and (2) that "it becomes easier for bots to recognise definition-style links". Item 2 I think is very weak, as any link in a definition is (or should be) of this sort except those in {{term}} or a form-of template. (Besides which, why do bots need to distinguish these types of links?) Item 1 is weak, because it's easy enough to type [[#English|water]]. (We can (if desired) even have a periodic scrape for self-links (<strong class="selflink">, or whatever the code is) for elimination purposes.) TL can guess what language a link is supposed to go to: anything in a definition not in a form-of template (or {{term}}) is English.—msh210℠ (talk) 16:17, 17 June 2012 (UTC)[reply]

I, too, oppose this. It's needless complication; to use DCDuring's phrase, it's the tail wagging the dog. And yes (echoing msh210's comment), I note with amusement that if this template would always link to English, Tabbed Languages needn't do anything more complex than link to English by default, to accomplish the same thing. - -sche (discuss) 17:02, 17 June 2012 (UTC)[reply]

{{jump}} (I haven't read the whole discussion, so forgive me if this is not relevant). Mglovesfun (talk) 17:10, 17 June 2012 (UTC)[reply]

I don't see how it is, no.—msh210℠ (talk) 17:52, 17 June 2012 (UTC)[reply]

It’s a good idea, but personally I wouldn’t use it. Clicking a tab or scrolling a little isn’t the end of the world; in the few cases where it is, using {{l|en}} is less effort than having yet another template to remember. Ungoliant MMDCCLXIV 18:28, 17 June 2012 (UTC)[reply]

If it did something like link to a particular gloss, then it would be worth using. It would even be worthwhile to be able link to a PoS section other than the first-occurring, eg, the second Verb section. I'd like to encourage such a thing if there is a chance of feasibility. Alternatively, would it be terrible to have headings like "Verb 2"?

"Tail-wag-dog" is open source. DCDuring TALK 18:45, 17 June 2012 (UTC)[reply]

Yeah, if it linked to a sense, it might be worth using. (Especially if that sense were then highlighted.) Perhaps that's what Mg meant, too.—msh210℠ (talk) 17:23, 18 June 2012 (UTC)[reply]

Hmmm... --Yair rand (talk) 17:31, 18 June 2012 (UTC)[reply]

OK. Perhaps we could have a list of the most highly polysemous English words for deployment of the target template and then produce a clean-up list of all those FL terms that use those highly polysemous English words in the definiens for deployment of the referring templates. DCDuring TALK 18:33, 18 June 2012 (UTC)[reply]

At the very least, we could use an update of Special:MostLinkedPages which hasn't been updated for 2.7 years. DCDuring TALK 18:39, 18 June 2012 (UTC)[reply]

On `{{obsolete spelling of}}`

In Latvian, kaŗš is an obsolete, pre-WWII spelling of karš 'war'. With {{obsolete spelling of}}, it is categorized in Category:Latvian obsolete forms, which is then given as a subcategory of Category:Latvian terms with obsolete senses. But words with obsolete spellings are not words with obsolete senses: the sense (in this case, 'war') is quite non-obsolete, and the word in its other, more modern, spelling continues to be used with this sense. Doesn't this cause problems with obsolete spellings in other languages? Shouldn't this template be changed to avoid this categorization? --Pereru (talk) 07:44, 19 June 2012 (UTC)[reply]

As I understand it, by "sense" we just mean each line that starts with a #. For example, if the word foo had two senses, one of which meant bar before WWII but is now obsolete, and another sense which continues to mean quux, then foo would be a term with an obsolete sense, even though it would not be obsolete altogether. Does that make sense? --Μετάknowledge^{discuss/deeds} 17:24, 19 June 2012 (UTC)[reply]

I see what you're saying, Pereru. The "term" in this case is kaŗš: it's an obsolete term, and therefore all of its senses are obsolete. But that's not a very intuitive use of the word "term", since normally we think of kaŗš and karš as being two forms of the same term. Nonetheless, I don't think this is a big problem. Categories represent a fairly loose association between entries — a good way to navigate between entries that have something in common — and should not be taken too strictly literally. For example, Category:en:People is a subcategory of Category:en:All topics, but that doesn't mean that people are topics. —Ruakh_TALK 18:12, 19 June 2012 (UTC)[reply]

Well, at least my first thought when I see a category named Category:English terms with obsolete senses is that I'll find there words that lost one or more of its senses (#'s, as Metaknowledge above says). I wouldn't go there to look for words that changed spelling -- so if amydst became amidst, without any other changes (the pronunciation is the same, all the senses -- the #'s -- are the same), I certainly wouldn't think of looking for it in Category:English terms with obsolete senses. I see what you're saying about kaŗš and karš (and presumably amydst and amidst being treated as two different terms rather than two forms of the same term, but I don't think this is what is done here -- after all, amydst and amidst are linked not as homophones or synonyms, not as different terms, but as alternative forms of the same term. (Again: no changes in meaning or pronunciation, only the spelling was altered.)

So, even though categories may only be a good way to navigate between entries, I think this particular categorization wouldn't help (at least me) navigate. As I said, I would have never thought of looking for amydst under en:Category:English terms with obsolete senses. This is different from Category:en:People being under Category:en:All topics, which is not at all surprising or counter-intuititve. --Pereru (talk) 01:19, 20 June 2012 (UTC)[reply]

Accelearation?

I've just found out about WT:ACCEL, which sounds like something I would like to add to my form-of templates (like {{lv-inflection of}}. How do I do that? WT:ACCEL only tells me how to modify my settings so that I can see the green links, but not what my template has to have in order to automatically create form-of entries. --Pereru (talk) 02:48, 20 June 2012 (UTC)[reply]

Template introduction for Gaulish

Hello,

Because I want to enhance Wiktionary in the Gaulish language, I need a declension grid for the nouns, otherwise they're useless due to that Gaulish is extinct and rare. I didn't create any templates yet, so how can I do that? Could you give me something like a "sketch" and answer me those questions, please:

How can I increase the grid to contain additionally Vocative, Instrumental, Locative?
How can I know the code, which I have to put into the edit of a noun's page?
How can I be sure that it is the correct code for the input of the noun in the code?

Please answer me & excuse my English Greetings HeliosX (talk) 13:16, 20 June 2012 (UTC)[reply]

You can copy the declension template from any other language (make sure you indicate in the edit summary or talk page which template it's from, because copying Wiktionary requires attribution (IANAL)) and modify it. I'd tell you how to add vocative, instrumental, and locative, but I don't know what you're adding them to: what other cases are there? The code would be {{cel-gau-decl-noun}}, say, assuming you save your template as [[template:cel-gau-decl-noun]]. (The template name should start with cel-gau-.)—msh210℠ (talk) 22:13, 20 June 2012 (UTC)[reply]

They have:

Nominative, Genitive, Dative, Accussative, Instrumental, Locative, Vocative

But where shall I put the ending in the template, there in -->(((pal|?suffix?))) ? Please answer me & excuse my English Greetings HeliosX (talk) 11:57, 21 June 2012 (UTC)[reply]

I did something like a sketch, could you edit it to make it correct? http://en.wiktionary.org/wiki/Template:cel-gau-decl-noun (I actually can't do templates...) Please answer me & excuse my English Greetings HeliosX (talk) 16:56, 21 June 2012 (UTC)[reply]

Is this even a good idea? How many Gaulish nouns are even attested in multiple cases? We shouldn't really be listing purely hypothetical forms here, should we? —An gr 18:26, 21 June 2012 (UTC)[reply]

Even though: you won't find every noun in Gaulish attested in every case, but you can rebuild them. Maybe we can add such a box like in Proto-Germanic at the top of every Gaulish noun to clarify that the declension is rebuilt, if there is a part in the declension rebuilt. I think it's good to give the most information we have about Gaulish nouns. And the unattested forms can also be marked or something in the grid. Please answer me & excuse my English Greetings HeliosX (talk) 09:58, 22 June 2012 (UTC)[reply]

You might want to see what I've done with {{fro-conj}}, which addresses more or less the same issue, just Old French is much, much better attested than Gaulish. Mglovesfun (talk) 11:01, 22 June 2012 (UTC)[reply]

IP range blocks?

Now that we have IPv6 activated, it would be useful to know if we can block a range of IP addresses, and if so, how. IPv6 addresses are normally given to customers in a whole block rather than just a single address, so if we want to block a user, we should be able to block their address range, because otherwise they could just change address and come back. In fact, it may well be useful to make the IP blocking tool automatically block the entire range for a single user (which I believe is the second half of the bits). —CodeCa t 00:31, 21 June 2012 (UTC)[reply]

Yes, absolutely, you should never block just a single IPv6 address. See m:User:Jonathan de Boyne Pollard/Guide to blocking IP version 6 addresses#Range blocking is the norm. —Ruakh_TALK 00:49, 21 June 2012 (UTC)[reply]

Do you think we could make a change to the software somehow so that it automatically sets the block to (IP)/64 by default? I would imagine most administrators would forget it otherwise. —CodeCa t 17:07, 21 June 2012 (UTC)[reply]

`{{IPA}}`

It strikes me that it would be quite useful if {{IPA}}, which currently allows for several different alternative transcriptions, also let you specify arguments like *label1=noun*, *label2=verb* and suchlike, for cases such as (deprecated template usage) outcast, which crop up fairly regularly. Ƿidsiþ 08:58, 21 June 2012 (UTC)[reply]

How would this be better than say {{qualifier|noun}}? Mglovesfun (talk) 09:09, 21 June 2012 (UTC)[reply]

Better? I don't know. Unless we think it's better for one template to be able to handle these eventualities, rather than calling on two. Ƿidsiþ 09:15, 21 June 2012 (UTC)[reply]

Size of edit summaries

Just how big should the longest edit summary be? I haven't found mention of a limit in any discussion. I had thought, based on some truncated summaries that I had noted that there was some Twitter-like size limit, but have established that the limit is greater than 240 characters. I will experiment to find out whether there is currently a limit, but there are limits to my ability to test all aspects of this.

I assume that the limit is in the number of characters (Unicode?).
Is there any technical limit on how many characters can be entered? Stored?
Are there any technical limits on how many characters can be displayed, say, on Special:Watchlist?
If there are display limits of some kind, then should there be limits on input?
If there are varying display limits, would a character counter help?
Could we use some kind of behavioral norm to limit any quantitative aspect of the use of edit summaries?
Are these silly questions?

-- DCDuring TALK 12:32, 22 June 2012 (UTC)[reply]

There's a limit, just try typing a really long edit summary and at some point it won't accept any more. It's happened to me. Mglovesfun (talk) 15:55, 22 June 2012 (UTC)[reply]

I think it's 250 characters for display at least. DCDuring TALK 16:57, 22 June 2012 (UTC)[reply]

It is 200 chars, according to the HTML source code: <input class="mw-summary" id="wpSummary" maxlength="200". This does not prevent a technically adept user from sending more than 200 characters (which may or may not be rejected by the server at the other end), but any well-behaved browser will enforce the limit. Equinox ◑ 17:08, 22 June 2012 (UTC)[reply]

Is Firefox not well-behaved? DCDuring TALK 17:41, 22 June 2012 (UTC)[reply]

Awesome edit summary, by the way. Mglovesfun (talk) 17:41, 22 June 2012 (UTC)[reply]

Actually, the underlying function restricts the maximum length to 255 bytes. This means either 255 English characters, or 127 Cyrillic characters, or 85 Chinese characters, or 63 Gothic characters. -- Liliana • 17:43, 22 June 2012 (UTC)[reply]

I think this leaves mostly the questions of, 1., whether a character counter could be implemented in JS, at least for a given character set, and, 2., should be used to encourage compliance that browsers don't regularly enforce. I would personally like to know when my edit summary exceeded the length users would see. DCDuring TALK 17:49, 22 June 2012 (UTC)[reply]

The number of characters available for an edit summary message seems to be:

250 with no header

250 less (the number of characters in the header + 6 [for "/* " and " */"])

I don't know what happened to the other 5 characters the function allows. Counter for characters remaining, anyone? DCDuring TALK 16:21, 25 June 2012 (UTC)[reply]

Special:UncategorizedTemplates

There are more than 5,000 of these, probably nearly 10,000. Can't some of these be bot-categorized based on the naming convention followed, eg, the name contains "/family", "/name", or "/script"? DCDuring TALK 16:14, 23 June 2012 (UTC)[reply]

The templates that you mention can definitely be categorized by a bot. However:

Do we even have categories to put them in?
These templates weren't created with doc-pages and whatnot, so there's no way to categorize them without actually editing them. In the case of the /family and /name templates that's not a huge deal, but in the case of the /script templates it means that the average page on the wiki will be added to the job queue something like (warning: random guess ahead) a half dozen times. Suffice it to say, we should really think about this before undertaking it. It's not something to do lightly or more than once.

—Ruakh_TALK 18:01, 23 June 2012 (UTC)[reply]

I believe that if edits are made only to the noinclude part of a template, it doesn't trigger a refresh on pages that transclude it. But I may be wrong. —CodeCa t 18:34, 23 June 2012 (UTC)[reply]

I don't know for sure, but I don't think that's the case. (It certainly didn't use to be.) —Ruakh_TALK 19:57, 24 June 2012 (UTC)[reply]

@Ruakh: If it can't be done for the script templates in question en masse without rendering the site unusable, could they be done one at a time at times of relatively low utilisation?

I was looking for a template and didn't realize how many were in these Augean stables. Couldn't all the templates of standard naming format be documented with boiler-plate templates? This might make that task of template cleanup less herculean, though also less worthy of a hero's effort, I suppose. DCDuring TALK 18:42, 23 June 2012 (UTC)[reply]

Let's not let the tail wag the dog now. Let's not categorize thousands of perfectly valid templates which work ok to empty Special:UncategorizedTemplates to the detriment of the templates. Mglovesfun (talk) 19:37, 23 June 2012 (UTC)[reply]

@MG: But the special page is intended to assist in recognition and cleanup of templates. This is somewhat analogous to flooding "recent changes" with bot activity. One man's tail is another man's dog.

@Ruakh: I would expect that the distribution of number of transclusions of the "/script" variety of template has a Pareto-type distribution, especially since a large number of this type are not transcluded on any content pages. Excluding the most commonly transcluded ones would be fine. DCDuring TALK 20:18, 23 June 2012 (UTC)[reply]

@DCDuring, I seem to think you once completely emptied Category:English terms needing attention once by fixing a load of entries, and removing the unfixed entries as well simply so that Category:English terms needing attention had no entries. This feels quite similar. Mglovesfun (talk) 20:20, 23 June 2012 (UTC)[reply]

Whenever I get a feeling like that I just lie down until it goes away.

But the objective of maintenance categories should be to dispose of the issues and empty the category. Using a maintenance category like Category:English terms needing attention for items that belong on a talk page is counterproductive. I try to empty out categories I watch, like Category:Entries with level or structure problems, Category:Translation table format problems, and, formerly, Category:English words that lack inflection template. I usually stop paying attention to categories that I can't successfully clean out, like Category:English terms needing attention. DCDuring TALK 00:19, 24 June 2012 (UTC)[reply]

Here's a naive newbie question: why do we even need to categorize most subtemplates? After all, I would suspect that we would deal with the subtemplates only in the context of their parent templates. Who cares if af/family and af/script are uncategorized if we always go to af first? I'm sure there are cases where this doesn't apply, but in those where it does, couldn't we just filter them out (i.e. whitelist them, or tell the code that builds the list to ignore them)? That way we we can clear out the irrelevant clutter and have a clear view of the ones that really need attention. Chuck Entz (talk) 22:43, 23 June 2012 (UTC)[reply]

I think it is Mediawiki standard-issue code that populates the list. I don't know whether we are asking for future maintenance problems with your approach.

One reason to categorize may be that the subtemplates could be reusable or serviceable as models. But the more important reason is maintainability. Special:SpecialPages includes lists of upto 5000 uncategorized pages from certain namespaces. If we have some pages that are deemed legitimate residents of the "Uncategorized" state, we could not distinguish those from pages that should not be there because they should have been categorized for finding and maintenance purposes. I'd be find if there were a Category:Templates in need of no categorization, but Category:Language family templates, Category:Language script templates, etc seem more like what would be useful. In the current situation some thousands of uncategorized templates (those exceeding the 5000-page limitation) are not visible except on Special:AllPages. Not all of them are of the types under discussion, eg, {{wikivar}}, not used in any content pages. DCDuring TALK 00:19, 24 June 2012 (UTC)[reply]

For what it's worth, I've assembled at User:DCDuring/UncategorizedTemplates a list of all uncategorized templates that aren't subpages. —Ruakh_TALK 16:58, 24 June 2012 (UTC)[reply]

Thanks. There are about 700 of them. Some look easy to categorize. I'll look for the older ones not transcluded in content pages as candidates for deletion, too. Others can join the fun of course. DCDuring TALK 17:11, 24 June 2012 (UTC)[reply]

Bot for mythology

Hello,

Could someone tell me how to create a bot, which adds those figures from Greek or Roman or Egyptian mythology in Swedish, Danish, Norwegian, Faroese, Icelandic and German automatically from a list with declension grids? I ask this, because I see many entries that for missing and to add them by hand would just be inneccessary, wouldn't it? I don't have collected any experience with bots yet.

Please excuse my English & please answer Greetings HeliosX HeliosX (talk) 17:35, 25 June 2012 (UTC)[reply]

Not unnecessary, that's how the rest of us do it! Just look at the recent changes. Mglovesfun (talk) 10:49, 25 June 2012 (UTC)[reply]

Okay, if you think so. I added this project to my page, you may look there to its state. HeliosX (talk) 17:35, 25 June 2012 (UTC)[reply]

Great Mandarin shift

If you see Special:WantedCategories, there are a lot of Mandarin... in (simplified|traditional) script categories as red links. Most category boilers aren't equipped to handle them, and the code is complicated enough that I can't do it. Also, this is another reason to unify some category boilers; when all category boilers need an update, the fewer the better. Mglovesfun (talk) 11:23, 25 June 2012 (UTC)[reply]

Indicating "mathematics" or "law"

I don't know how to indicate that a word is used in a certain field, like mathematics or law.

I recently added to the definition of 'tail'.

I wanted my definition to say:

(law) limitation of inheritance to certain heirs.

There was an existing defintion that said:

(mathematics) (of a sequence) All the last terms of a sequence, from some term on.

It was coded like this ~~(backslashes added so it doesn't get changed)~~:

{{mathematics}} (of a sequence) All the last terms of a sequence, from some term on.

So I put {{law}}. But it changed it to Lauje.

So I just made it (law), but it's not a link.

How do I make it like the link to 'mathematics'? — This unsigned comment was added by Ckendall86 (talk • contribs).

coding modified so it won't interfere with the rest of the page Chuck Entz (talk) 14:20, 27 June 2012 (UTC)[reply]

The best way to do that is with the {{context}} template (if you want to show a template here, use {{temp|xyx}}, where xyz is what you want to show between the curly brackets. You can also surround it with <nowiki></nowiki>, but that also turns off formatting between them as well). After "context|" put "legal", "mathematics", "UK","US", etc. If you have multiple contexts, just put them all in, separated by "|".

There are separate templates such as {{legal}} and {{mathematics}}, but context uses those when your word matches one of them, and just repeats what you put when it doesn't, so you're covered either way. Chuck Entz (talk) 14:38, 27 June 2012 (UTC)[reply]

See Category:Topical context labels if you want to know what defined labels are available (to show categories here, add a colon after the square brackets, like this: [[:Category:Topical context labels]]). Also, don't forget to sign your comments with four tildes (~~~~) Chuck Entz (talk) 14:48, 27 June 2012 (UTC)[reply]

Thanks to SemperBlotto

SemperBlotto sent me a message. I don't know how to reply to him privately, so I'll put my reply here in the hope that he'll see it:

Thanks very much for fixing up my edit of the 'tail' entry, and for teaching me how it should be done.

Now that I know to use the '+' tab, I won't be messing up the format of the Grease Pit again.

Fanks! — This unsigned comment was added by Ckendall86 (talk • contribs).

Just reply on your own talk page, or on the talk page of the user you want to talk to. And you can sign your talk with four tildes ~~~~. SemperBlotto (talk) 15:22, 27 June 2012 (UTC) (love him or hate him)[reply]

Tips for keeping edit history when "moving" only part of a page?

I was doing some entry cleanup this evening and ran across [[panya]]. The Japanese portion is misfiled, as this should be at [[pan'ya]] instead. The page was created by mistake back in 2008, and shortly thereafter acquired a Swahili entry. Due to the rules of romanization, [[panya]] would be read as [pä.njä], which is not a word in Japanese. The intended Japanese word is パン屋 (pan'ya), read as [pã̠ɴ.jä] instead.

Since the Swahili entry means that I cannot simply move the [[panya]] page in its entirety, I was just about to create the [[pan'ya]] page, and then cut the Japanese content from [[panya]] and copy it over, when I recalled a previous discussion with Ruakh and -sche where they explained that doing so is a bad idea due to the provisions of the Creative Commons license, which requires that edits be attributable to the contributing editors.

With that in mind, how does one move part of a page while still maintaining the relevant edit history for that part? -- Eiríkr Útlendi │ Tala við mig 07:15, 28 June 2012 (UTC)[reply]

(1) In some cases, it is possible to temporarily delete all the edits which modified the non-Japanese sections of the page, leaving only the Japanese-section-modifying edits, then move the trimmed page (from [[panya]] to [[pan'ya]]), then restore the deleted revisions (of [[panya]], in situ).

(2) You could remove the Japanese section of [[panya]] (because it is wrong), and then create [[pan'ya]] using your own knowledge of Japanese. What you write may end up being identical to what was written at [[panya]], but no infringement occurs, because you're not copying [[panya]], you're creating something using your own knowledge of Japanese.

(3) Idea which just occurred to me: copy-and-paste the content from [[panya]] to [[pan'ya]], and simply link to the old page, especially to its edit history, in the edit summary you use when creating [[pan'ya]]. - -sche (discuss) 08:50, 28 June 2012 (UTC)[reply]

Thank you, -sche, I think I'll take Door Number 3. :) -- Eiríkr Útlendi │ Tala við mig 16:42, 28 June 2012 (UTC)[reply]

Gender templates for other gender-like classifications?

We have templates like {{m}}, {{f}}, {{n}} for genders but some languages use other classifications outside that scheme. The Bantu languages in particular use a classification system based on numbered noun classes rather than genders. Would it be desirable to create gender templates for these? Like {{1}}, {{2}}, {{3}}, {{4}} and so on? —CodeCa t 11:29, 28 June 2012 (UTC)[reply]

I don't think it would be a good idea to use just the bare numbers, because different languages no doubt have different arrangements of noun classes, so noun class 1 in language A might be completely different from noun class 1 in language B. Chuck Entz (talk) 12:45, 28 June 2012 (UTC)[reply]

Of course, but it's that way with genders too. A feminine noun in German is not the same thing as a feminine noun in Russian or in Arabic. —CodeCa t 16:05, 28 June 2012 (UTC)[reply]

I don't speak any Bantu languages, and it's been a long time since I've read anything about them, so please correct me if my understanding is wrong . . . but from what I remember, in the traditional numberings of noun classes in Bantu languages, each singular noun class has a corresponding plural noun class; that is, if a noun's singular is in noun class 1 or 3 or 5, then its plural will be in noun class 2 or 4 or 6 (respectively). So a noun as a whole has two classes. Given that we don't label e.g. Spiegel as "masculine singular", I'd think that we wouldn't want to label a Bantu noun as "noun class 1", but rather as something like "noun classes 1–2". Also, even for languages where a single noun does have a single noun-class number, I think that {{1}} isn't a great name. Nothing about it shouts "noun class". Maybe something like {{g1}} would be better? —Ruakh_TALK 17:21, 28 June 2012 (UTC)[reply]

Chicago timestamp bug

Whenever I set my preferences to the America/Chicago timestamp, it signs everything as if it was posted a year ago. For brand new posts it gives something like this.
posted 11 months and 31 days ago on June 28, 2012 at 10:25 am
So I changed my timestamp to America/Mexico City. My format is set to (Month day_number, year), but it doesn't seem to be displaying properly. I am also using Firefox 13.0.1 as my browser. Is this a known issue? -- AnonymousDDoS (talk) 15:30, 28 June 2012 (UTC)[reply]

N.B. There are two relevant preferences: this user has set his/her time-zone at Special:Preferences#mw-prefsection-datetime to America/Chicago, and has checked the “Change UTC-based times and dates, such as those used in signatures, to be relative to local time” checkbox at Special:Preferences#mw-prefsection-gadgets. That checkbox enables User:Bequw/comments in local time.js, and it is almost certainly there that the bug lies. —Ruakh_TALK 15:51, 28 June 2012 (UTC)[reply]

Also, note that this bug affects everyone's timestamps, not just your own -- I see the same for everyone's posts. However, also note that this only occurs for posts on the same day -- if you wait 24 hours, all those posts magically become yesterday's, not yesterday's one year ago. I suspect that there's an off-by-one error in counting days; this is also hinted at by the wording, "11 months and 31 days ago", as "31 days ago" would perforce have to be at least one month prior, and up to one month and three days ago.

FWIW, I've just learned to ignore this particular oddity. -- Cheers, Eiríkr Útlendi │ Tala við mig 16:47, 28 June 2012 (UTC)[reply]

I filed it at Bugzilla a while ago and was told that it's something in our local JS, but I forgot to raise the issue here. Ruakh, I'm sure you're right - but are you able to fix it? --Μετάknowledge^{discuss/deeds} 10:52, 12 July 2012 (UTC)[reply]

The date math is really just fundamentally broken; it makes completely broken assumptions about the way calendars work. I would have to completely rewrite it. I don't really have any interest in doing that, sorry. That said, keep in mind that User:Bequw/comments in local time.js is based on a two-year-old version of a script on Wikipedia, namely w:User:Gary King/comments in local time.js, which has undergone many changes since then. You might want to ask Bequw if he's willing to incorporate those changes, in the hopes that they'll include a fix for this issue. —Ruakh_TALK 20:12, 12 July 2012 (UTC)[reply]

English Wiktionary Android version

Entry boat. Screenshot of the open-source Android software "kiwidict 0.096" with data extracted from the English Wiktionary.
List of Russian words (marked by red oval) from the English Wiktionary with typo: Latin "c" instead of Cyrillic "с", fixed: diff1, diff2 and diff3.

kiwidict is offline multilingual dictionary and thesaurus based on the English Wiktionary. Welcome to play with this dictionary. Your feedback is welcome!

This software could be used in order to search and correct :) errors in the Wiktionary (see second figure). -- Andrew Krizhanovsky (talk) 12:54, 29 June 2012 (UTC)[reply]

`{{yue-hanzi}}`

This template doesn't seem to apply the correct fonts, the Chinese characters display as boxes to me whereas {{infl|yue|verb}} displays correctly. Can anyone figure out where the problem is? -- Liliana • 15:31, 29 June 2012 (UTC)[reply]

I'm not 100% sure, but it looks like the first logic clause in the <font> tags have traditional and simplified backwards:

{{#if:{{{tra|}}}|{{Hans|{{PAGENAME}}}}|{{#if:{{{sim|}}}|{{Hant|{{PAGENAME}}}}|{{Hani|{{PAGENAME}}}}}}}}

I believe that should be {{#if:{{{tra|}}}|{{Hant| and {{#if:{{{sim|}}}|{{Hans|. The #if: statements later in the template seem to correctly correlate the {{{sim}}} and {{{tra}}} parameters with the {{Hans}} and {{Hant}} script templates. -- Cheers, Eiríkr Útlendi │ Tala við mig 17:34, 29 June 2012 (UTC)[reply]

I think that that aspect of the logic is correct: the idea is that if a tra= parameter is specified, then the current headword isn't the traditional-characters version, so it's simplified-characters-only, i.e. {{Hans}}. (And so on for the other two branches.) —Ruakh_TALK 18:31, 29 June 2012 (UTC)[reply]

Aha, yes, it looks like you're right. I was mistaken. -- Cheers, Eiríkr Útlendi │ Tala við mig 21:12, 29 June 2012 (UTC)[reply]

Protecting Maintenance Templates

Looking through the logs, I noticed Liliana-60 (talk • contribs) protected {{rfc}} with the comment "how on earth is this not protected". After clicking "Edit" on many of the templates in Category:Maintenance templates and its subcategories, I would guess that the majority of them are also not protected (though the most critical ones are, with an exception or two). Are there any in that category and subcategories that shouldn't be protected? if not, I'd be happy to make sure they're all protected ASAP, if no one beats me to it Chuck Entz (talk) 07:00, 30 June 2012 (UTC)[reply]

I went ahead and protected most of them, since I figured it would be better to have to unprotect a few than to have some vandal read this and change {{rfm}} to read "Request for Mommy" Chuck Entz (talk) 08:15, 30 June 2012 (UTC)[reply]

Autoconfirmed should be enough, right? Mglovesfun (talk) 10:11, 30 June 2012 (UTC)[reply]

Template cleanup

I've tried to make it as easy as possible to determine use of uncategorized templates at User:DCDuring/UncategorizedTemplates. All templates that are not to be considered for deletiond will be categorized (on the documentation page, please!) as soon as someone figures out a good category for each one. The templates that are transcluded on at least a few pages, but not too widely, will get {{documentation}} added on the main page. Templates that are very widely transcluded will eventually get {{documentation}} on the main page.

Category:Templates includes templates that did not seem to have a more specific category to the person categorizing, sometimes me. Category:Experimental templates may have templates worth keeping, but perhaps under something like Category:Interesting template experiments to differentiate those that have some redeeming value as examples of creative approached or something. Perhaps we need Category:Temporary templates to house, by default, all new templates as well as those intended for relatively short-term use. Such a category could be occasionally reviewed, but would, in any event, provide warning that the items so categorized were not intended for general use as is. DCDuring TALK 15:50, 30 June 2012 (UTC)[reply]

Perhaps someone with good knowledge of the technical possibilities could come up with more categories suitable for characterizing templates by technical features or the nature of use (use only be other templates, use only within a specific family of templates, subst only use, use only in certain namespaces, etc). DCDuring TALK 15:54, 30 June 2012 (UTC)[reply]

I'd say it's important not to nominate templates for deletion merely because it's easier to do so than to improve them or add documentation and categorization. In other words, laziness shouldn't be a reason for nominating something for deletion. Mglovesfun (talk) 15:55, 30 June 2012 (UTC)[reply]

After edit conflict, Category:Templates that must be substituted. Mglovesfun (talk) 15:55, 30 June 2012 (UTC)[reply]

Yes that category would be very important because it's the only way we would be able to tell that a template is being used despite having few transclusions. On the other hand, how do we know whether a template that is meant to be substituted is still being used by anyone, except by rfding it or bringing it up for discussion somewhere else?

As for DC's post... I think that temporary templates that are used just by one user would be better suited to a user's namespace. That way they won't get in the way of organised cleanup activities. On the other hand, it would make it easier for us to lose track of it and have it linger around unused for years. —CodeCa t 16:00, 30 June 2012 (UTC)[reply]

I don't suppose we could have a category for templates whose documentation does not reflect the actual action of the verb? (e.g. {{fr-noun}} documentation does not mention the function of an optional 2nd and 3rd positional parameter) SemperBlotto (talk) 16:07, 30 June 2012 (UTC)[reply]

@MG: Any template not categorized and transcluded nowhere in content space; not transcluded more than, say, ten times anywhere; with no documentation suppage or talk subpage; with no in-template documentation; and with no clear, currently valuable purpose is at least highly suspect, it seems to me. That the sole substantive contributor is no longer active here is also good evidence for marginal cases. DCDuring TALK 16:11, 30 June 2012 (UTC)[reply]

@SB: Something on the template talk page would presumably get on the watchlist of any contributor to the template. Are there other contributors who should be asked for help? i'm concerned about not overloading our relatively scarce crew of technically knowledgeable types. For language-specific templates perhaps heavy users might be able to help, I suppose, but there should be more specific ways of catching their attention, perhaps {{attention}}. (Admittedly, it is says it is for "terms".) DCDuring TALK 16:20, 30 June 2012 (UTC)[reply]

I would propose a new template {{rfu}} which means 'request for update', which could be used when something doesn't reflect the current state of affairs. It might be used for templates and their documentation, but also for other things like policy pages or tutorials. —CodeCa t 16:29, 30 June 2012 (UTC)[reply]

@DCDuring, a little-used template is sometimes merely one that people don't know exist. It doesn't necessarily mean that the template is poor or unuseful. Mglovesfun (talk) 16:59, 30 June 2012 (UTC)[reply]

@MG: Sure, but how can you defend one "not categorized and transcluded nowhere in content space; not transcluded more than, say, ten times anywhere; with no documentation suppage or talk subpage; with no in-template documentation; and with no clear, currently valuable purpose"? They still have to go through RFDO, anyway, where, of course, no item is ever summarily dispatched. DCDuring TALK 17:34, 30 June 2012 (UTC)[reply]

No they don't, if they're good (that is, useful) just use them instead of nominating them for deletion. I find that argument a lot like nominating terms for deletion because nothing links to them, even when they meet WT:CFI. Mglovesfun (talk) 10:31, 1 July 2012 (UTC)[reply]

The difference is that terms are meant to be looked up, which can happen without links, while templates are meant to be used, which can't happen without transclusions unless they are substituted. —CodeCa t 11:25, 1 July 2012 (UTC)[reply]

One of the most important things to do for this is cleanup is to categorize any uncategorized subst-only templates into Category:Templates that must be substituted. I did so yesterday with about 15, but they were relatively easily identified: the names began "new-". Any help in this regard would be appreciated.

Because there seems to be no convenient way to tell whether a subst-only template is used, they seem likely to get a free pass for now unless they are obviously early or inferior versions of other templates. I suppose there could be ways involving analysis of change logs or installation of counters, but there are not now so many of them that any costly effort is worthwhile. Those not complying with WT:ELE could be brought into compliance. DCDuring TALK 17:55, 1 July 2012 (UTC)[reply]