Template talk:dlc

This is an invented code which doesn't exist according to SIL. -- Liliana • 17:17, 22 June 2012 (UTC)
 * I fail to see your point; isn't the question 'does the language Elfdalian exist'? Mglovesfun (talk) 17:43, 22 June 2012 (UTC)
 * Both are questions, actually. Do we want to treat the lect as a separate language? If so, what code should we use for it? If "dlc" isn't a valid code, we need to create a code following our "exceptional code"-creation pattern: so, "gmq-elf" or similar. This is because "dlc" could later be assigned by ISO to some other language. - -sche (discuss) 18:07, 22 June 2012 (UTC)
 * Yes, I should've said that. Mglovesfun (talk) 18:24, 22 June 2012 (UTC)
 * I have to agree with this sentiment. dlc is an invented code and as such, should either be moved or deleted because of that. Razorflame 18:27, 22 June 2012 (UTC)
 * is the LINGUIST List code for the Dalecarlian dialects, of which Elfdalian is one. The LINGUIST List code specifically for Elfdalian is, so I would recommend either using   for Elfdalian or lumping all of Dalecarlian together as a single language and using   for it. (I would prefer the second option myself.) —Angr 19:50, 22 June 2012 (UTC)
 * Incidentally, "dlc" for Dalecarlian was in a draft version of ISO 639-3, but was removed (along with "scy" for Scanian) before publication under pressure from the Swedish government, who wanted those two languages to be considered dialects of Swedish rather than separate languages (see Requests for new languages/Wikipedia Elfdalian and GerardM's blog). GerardM also says, "According to the rules of the ISO-639 standard, the code dlc will not be used for anything but Dalecarlian", so we don't have to worry about it being assigned to something else. —Angr 17:50, 25 June 2012 (UTC)
 * The second link says According to the rules of the ISO-639 standard, the code dlc will not be used for anything but Dalecarlian. If that's true, then we can safely use that code as there is no danger of a conflict with another language that is later assigned the code 'dlc'. Whether or not it's not valid ISO-639 is not really that important, as we can use any code we like as long as it conforms to the rules for the HTML lang= attribute (which allows codes other than ISO-639). A few days ago I did some researching about Elfdalian and added some entries and even without knowing Swedish too much I can already tell this language is almost as close to Old Norse as it is to Swedish, and much less like Swedish than standard Danish or Norwegian are. There is no way a Swedish speaker would understand it. As for the other varieties of Dalecarlian... well I don't know too much about them and they aren't very well documented, so I suppose we could use the name 'Dalecarlian' with the assumption that it will primarily be Elfdalian, but also allow other varieties if attested. 17:56, 25 June 2012 (UTC)
 * Re: "the HTML lang= attribute (which allows codes other than ISO-639)": I believe you're mistaken, or rather, irrelevantly correct. The HTML specification defines valid language tags by pointing to 1766 or BCP47. 1766 required the first part of a language tag to be either i, <tt>x</tt>, or a two-letter code assigned per ISO 639. BCP 47 (currently 5646) allows a few more possibilities, but it still forbids a language tag from starting with a two- or three-letter code that is not a "shortest ISO 639 code", unless the tag is one of a very short list of grandfathered tags (such as <tt>sgn-CH-DE</tt> and <tt>zh-min-nan</tt>). So while it's true that it's possible to construct a valid HTML <tt>lang=</tt> attribute that is not an ISO 639, <tt>dlc</tt> is not an example of this possibility. —Ruakh <i >TALK</i > 19:15, 25 June 2012 (UTC)
 * Nevertheless, Wiktionary does have language templates that don't correspond to ISO 639, and indeed we recognize languages that ISO 639 does not, such as Jèrriais and (among North Germanic languages) Gutnish . I do think our best bet is to move this template to  and change the content of it from "Elfdalian" to "Dalecarlian" so that other Dalecarlian varieties (to the extent they're attested in written form) can be subsumed under it. (On the other hand, we also have a precedent for calling a language by the name of its best known dialect: we call  "Cantonese" even though the code covers all Yue dialects, not just the Cantonese standard dialect. So in principle we could get away with calling  "Elfdalian" as well.) —Angr 20:51, 25 June 2012 (UTC)
 * I think it's fine to recognize languages that ISO 639 does not, but I think our use of nonstandard-language-codes-that-look-like-real-ones is a mistake. (And CodeCat seems to agree with me, since she says that "we can use any code we like as long as it conforms to the rules for the HTML <tt>lang=</tt> attribute"; she was mistaken about how to do that, but acknowledged that it's something we should do.) Fortunately, it's a mistake that we can rectify, once we acknowledge that it is a mistake worth rectifying. In the case of Dalecarlian, we should use (or similar): <tt>gmq</tt> is a valid language subtag, but <tt>dlc</tt> is not a valid extension code, so we need to mark it as "Private Use" by using the <tt>x</tt> prefix. —Ruakh <i >TALK</i > 21:19, 25 June 2012 (UTC)
 * Part of me disagrees with this on principle, I don't feel right with the idea that a standards body like ISO can tell us what codes to use, especially not when a national government has influence on it. What basically happened is that the Swedish government told us we can't use 'dlc'. It's a bit crazy... :/ 21:23, 25 June 2012 (UTC)
 * Ruakh, if you say  is better than , I'll take your word for it, and if there's consensus that you're right then we should move the other non-ISO code templates to names with an "-x-" in them as well. CodeCat, I understand your frustration and agree that it sucks that ISO caved in to political blackmail, but HTML uses the ISO codes for better or for worse, so we can't just start making up codes that influence the HTML of our pages. —Angr 21:35, 25 June 2012 (UTC)
 * I do say so, but I don't ask you to take my word for it. I encourage you, and other editors, to look through the relevant standards (W3C technical reports on HTML/XML/XHTML; IETF RFCs on language codes; the IANA language-code registry), or perhaps this informal guide by the W3C if you prefer a less technical document, and form your own opinions. —Ruakh <i >TALK</i > 21:43, 25 June 2012 (UTC)
 * One thing I see in the informal guide is that we have another option for picking a code:  using the ISO 3166-2 country and subdivision name as an extension will get us a properly defined "Swedish as spoken in Dalarna County". The drawback to using that is that there may be a local variety of Standard Swedish spoken there in addition to Dalecarlian. —Angr 22:10, 25 June 2012 (UTC)
 * No, that's not allowed, because subtags are separated by hyphens, and can't contain hyphens, so <tt>SE-W</tt> is two separate subtags: <tt>SE</tt> would mean Sweden, but <tt>W</tt> would be an invalid extension subtag (invalid because unregistered, and because extension subtags always have to be followed by one or more additional subtags). But if we're O.K. with treating Dalecarlian as a form of Swedish, as <tt>sv-*</tt> implies, then we may be able to register a "variant" subtag with IANA, such as <tt>dalec</tt>, such that Dalecarlian would be publically defined as <tt>sv-dalec</tt>. (I think. I've never really looked into how variant codes work.) —Ruakh <i >TALK</i > 22:30, 25 June 2012 (UTC)
 * Well, it wasn't supposed to be more than a kludge, so if it isn't allowed anyway, I'd rather use  for linguistic reasons (Dalecarlian linguistically isn't just a dialect of Swedish, whatever Stockholm says). —Angr 09:07, 26 June 2012 (UTC)


 * [@CodeCat, after e/c] Well, but the main purpose of these codes is interoperability with other standards-compliant systems. I don't like that political factors play a role in the assignment of codes, but we don't really accomplish anything by silently subverting it. We can fight the injustice by expanding our Dalecarlian coverage and treating its words as ==Dalecarlian==, but assigning Dalecarlian its own pseudo-code <tt>dlc</tt> will not help and should play no part. (Imagine that your corrupt local government was bribed to split up your telephone area-code into two, and you don't like the new area-code you've been assigned. Does that mean you'll continue to give out your phone number as using the old area-code, even though it won't actually enable long-distance callers to reach you?) —Ruakh <i >TALK</i > 21:43, 25 June 2012 (UTC)


 * I think I get it, dlc was removed as Wikipedians put it 'with prejudice'. Since we keep because although ISO 639-1 retired the code, we still find it useful, and we should aim to make use of ISO 639, but not to be a slave to it. So I'm fine keeping this per the KISS principle; why use a seven or nine character code well three will do. Mglovesfun (talk) 21:39, 25 June 2012 (UTC)
 * There is a difference: sh used to be valid. dlc never was. -- Liliana • 21:41, 25 June 2012 (UTC)


 * @Ruakh: can the codes Wiktionary uses in its HTML be different from the template titles/codes, or are the template-values automatically the HTML values? I.e., can we (per our current naming scheme) name the template, but set the HTML to "gmq-x-dlc"? What HTML codes are currently associated with our other exceptional codes, such as ; do they include "-x-"? When the WMF creates exceptional codes for new wikis in languages without ISO codes, what naming scheme do they follow? - -sche (discuss) 22:35, 25 June 2012 (UTC)
 * Re: first two questions: I guess it depends what you mean by "can" and "automatically". Currently, we design templates under the assumption that these will be identical; we don't have any mechanism in place for mapping from language-template-name to HTML-language-tag. If is <tt>Dalecarlian</tt> and  is <tt>Latn</tt>, then  will be <tt> foo  </tt>. But there's no technical reason that we can't institute such a mapping.
 * Re: third question: As I'm sure you can guess from the answers to your first two questions, the answer to this one is "e.g. <tt>roa-jer</tt>; no, no <tt>-x-</tt>".
 * Re: last question: They no longer do that — see Language proposal policy — but back when they did do it, their naming scheme was basically "invent something realistic-looking". Needless to say, this eventually led to problems. The Swiss German Wikipedia is <tt>als.wikipedia.org</tt>, but ISO/SIL assigns <tt>als</tt> to Tosk Albanian and <tt>gsw</tt> to Swiss German. Our scheme, inventing codes like <tt>roa-jer</tt>, is not nearly so bad as that, IMHO; but it's still far from ideal.
 * —Ruakh <i >TALK</i > 23:11, 25 June 2012 (UTC)

Sorry if this is a stupid question but, who forces us to use syntactically valid language codes? As long as browsers can parse them, everything should be fine? -- Liliana • 23:27, 29 June 2012 (UTC)


 * If they aren't syntactically valid, then browsers can't necessarily parse them. I don't trust them not to give out dlc. (For example, Romania found rom to be troublesome, as it imply the Roma in their eyes, so they demanded a new country code and got rou, even though that was reserved, as the Republic of Uruguay has apparently used it in some context.) And if it's given to say a language written in China in the Chinese script, indicating the text is in dlc may make a browser preemptorily change to a Simplified Chinese font.--Prosfilaes (talk) 00:44, 10 October 2012 (UTC)


 * Tagged to be speedied as a deprecated template. --ElisaVan (talk) 21:02, 3 November 2013 (UTC)