User talk:Stephen G. Brown/A translation template

Definition from Wiktionary, the free dictionary
Jump to navigation Jump to search

A translation template[edit]

  • Could I pass this comment on my talk page off to you? DAVilla 10:35, 21 October 2006 (UTC)

I think it is a pity there is no template for translation, why not create one like the french one : Template:trad.

It would be usefull to create links like :

Sinsedrix (talk) 14:22, 20 October 2006 (UTC)

I have seen this used, but I don’t see the benefit. It will produce a blue link to the French page regardless of whether the page exists. I don’t think we should add, for example, "ma tante (°)" until fr:ma tante comes to exist. Otherwise, it just produces a lot of frustrating dead-ends. —Stephen 11:02, 21 October 2006 (UTC)
Note that the fr.wikt doesn't just have one template (trad) they have two: trad and trad-. One generates a blue link and the other one a red link. They have a 'bot that goes around changing trad- to trad when entries appear, and trad to trad- when someone adds a translation to a non-existing entry. Robert Ullmann 11:30, 21 October 2006 (UTC)
Oh, I think I see what you mean. The bot needs a point of reference in order to work properly, so a blank "trad- " template is inserted by hand after each item. That way it makes sense. I suppose the blank template would have to be language specific, like {{trad-zh}}, so the bot would know which Wiktionary to point to. —Stephen 11:42, 21 October 2006 (UTC)
It isn't blank; trad- is just like trad, except that it forces the link to be red. All the 'bot does is check the explicit link. I think most editors actually put in trad, and it is the bot that fixes it. Just to be clearer, this is the first bit of the translation table for fr:ordinateur:
* {{af}} : {{trad-|af|rekenaar}}
* {{sq}} : {{trad-|sq|kompjuter}}
* {{de}} : {{trad|de|Computer}} {{m}}, {{trad|de|Rechner}} {{m}}, {{trad-|de|Datenverarbeiter}}
* {{en}} : {{trad|en|computer}}
* {{ar}} : {{trad-|ar|كمبيوتر}} (kambyutr)
* {{br}} : {{trad-|br|urzhiataer}} {{m}}
* {{bg}} : {{trad-|bg|компютър}} {{m}}
* {{ca}} : {{trad-|ca|ordinador}} {{m}}
* {{ko}} : {{trad|ko|컴퓨터}} (keompyuteo)

And compare rw:mudasobwa Robert Ullmann 11:54, 21 October 2006 (UTC)

Hmm, I see. I don’t like the Rwandan set-up very much, because the language codes look confusing. I have to stop and think whether, for example, "computador (es)" is an indication of a plural or a language code. I would prefer (*) or (^). Otherwise, it seems okay (although quite a lot of work).
I’m not sure some of them will work properly. Can we reliably nest templates? Arabic would *Arabic: {{trad-|ar|{{ARchar|الكمبيوتر العربي|الكُمْبِيُوتَر العَربِي}}}} (al-kombyúter al-‘áraby) {{m}} in simple cases; in more complex cases, only part or parts of an expression might be wikified. (Also happens with Japanese, Russian, and other languages on occasion). —Stephen 12:34, 21 October 2006 (UTC)
You can if the templates are set up right. I don't get your example; ARchar doesn't take two parameters?:
The rw templates are/will be set up to use script form templates automatically if they exist; e.g. if ar-char is defined, it will be used. The language codes aren't so confusing when you are used to seeing them (in blue) on every inflection and translation line. But we might try some other syntax. The problem with ^ or * is that they don't mean anything. Robert Ullmann 14:45, 21 October 2006 (UTC)
In Arabic, Persian, Hebrew and Urdu, you can enter a word as usually written, without short vowels, double consonants, etc., like Template:ARchar, or you can point the word with all the matres lectionis Template:ARchar. Regardless of whether any or all the matres lectionis are used, the word must be directed to the page without them. To achieve this, we put {{ARchar|[[كمبيوتر|كَمْبُيُوتَرٌ]]}}.
Okay. You left the wiki brackets out of your previous example. And isn't this example backwards? (oh right, the WM s/w does scary parsing of RTL languages even when imbedded in LTR! ;-) In the fr.wikt (rw, etc.) you have to point to the page name, they don't use the typographical alternation you are using here. (Wasn't someone complaining about that?) The translation template for the en.wikt would have to have another parameter for actual page name. (Once you start making the page name different from the word, all sorts of stuff gets horribly difficult; I was complaining to EncycloPetey about doing this in Latin) Robert Ullmann 15:44, 21 October 2006 (UTC)
I think that fact that (^) means nothing is what makes it so useful. It cannot be confused for an abbreviation or a bit of grammar (some languages often toss in little grammatical bits, such as Breton -ou and Dutch de). Once you click on (^) a time or two, you realize what it is, and it looks the same on every entry. With the existing pl, f, sg, c, -ou, de, p4, nc 5, and all the other bits that are in use, adding cryptic language codes everywhere really makes a mess. Besides that, I keep finding errors produced by a misreading or misunderstanding of language codes...I’ve seen Navaho words thus mislabelled as Dutch (nv, nl), and then the first Dutchman to see "*Dutch: ndee biyáti' bígoch'il'aah" simple deletes it as nonsense. Just yesterday I came across Kashubian mislabelled as Kashmiri because of language code confusion. —Stephen 15:14, 21 October 2006 (UTC)
This may depend a bit on the language; in the en.wikt, we have names for all the languages (at the very least, IS 639 gives us "English" names.) Kinyarwanda doesn't have a word for Navaho; and there are hundreds of others that will need to be recoded: (sot) right now is Sesotho, but will get changed (even if the appropriate name is coined), probably Igisotho. The isemura template forces all the "odd bits" into parens not wikified, so the ordinateur (m) (fr) syntax isn't too bad. Yes, the en.wikt is different ... Robert Ullmann 15:32, 21 October 2006 (UTC)