Talk:Using templates in Wiktionary

From Meta, a Wikimedia project coordination wiki

I use these templates in the Latin Wiktionary (except for the {{-xx-}} kind, there's little reason to differentiate them from =={{xx}}==: translations can be as much a definition as a full entry). The main reason I use them is because my Latin skills aren't perfect and someday someone may come along with a better name for certain languages, and it'll be easier to change them all at once.

Anyway. It isn't true that the translations can be copied without any changes. Alphabetization is an issue: the effort saved in not having to translate is transferred to moving things around so that they still appear in alphabetical order. This is also made slightly more difficult by the fact that presented with a code like {{eu}} one needs to already know that 'eu' will expand to (e.g.) 'Vasconice' (which admittedly isn't a large problem, as it unpacks on preview).

Another benefit of templates (generalized from above) that you don't mention is that for lesser-known languages, the preferred designation of a language may change. Even in languages like English this can happen, like Persian -> Farsi, Slovenian -> Slovene, etc. But a developing language community may have started out with (e.g.) "Monapice", or even just untranslated "Manx", and decide to change it later to "Mannensis" based on developing trends. Changing the template word will save on having to get a bot to change it (or worse, having it done by hand), though there is still the alphabetization problem.

A variation of this benefit is that people who don't know the language they're working in can just use {{eu}} and automatically have it replaced by the proper name.

Why is a new namespace needed? As I understand it, the template namespace was designed for translationary purposes like these. Incidentally, one can use a title like Template:en on any wiktionary—the software will automatically translate the word "Template" to the particular language's equivalent. —Muke Tever 16:23, 2 Sep 2004 (UTC)

The case for using -xx- templates[edit]

The useage of the -xx- templates (xx is a ISO 639 code) is not lexicological; it is used to indicate that the next part is going to be about a word in this language. Using this methodology allows you to create a list of words in that language. Something that without this method needs to be done by hand.

Another thing is that the -xx- template refers to something different from the xx templateon the nl:wiktionary; here it refers to a page that gives you a choise to:

  1. the article for the word in wiktionary.
  2. the article for the word in wikipedia.
  3. the intention is to refer to the list of words in that language as well.

To make this clear I need some code that gets the "what links here" output for the template, does some filtering and alphabetises the list and publishes the new list. This is on my "todo" list.

  1. In what way is the normal use of {{xx}} (viz., {{xx}}: [[woord]]) not indicating that the next part is about a word in the language?
  2. The {{-xx-}} template's whatlinkshere is less useful, as it only lists words that have a full article on them, while the wiktionary will have many more words in that language defined with {{xx}} in translations sections. Using {{-xx-}} instead breaks this list, so that one cannot find all the words of a language in the wiktionary with one whatlinkshere.
  3. nl:'s {{-xx-}} templates don't even use the {{xx}}, which might help the previous point, instead they spell out the word in full.
  4. There's no reason that the {{xx}} template can't link to the same kind of page that nl:'s {{-xx-}} ones do. In fact, it's a good idea and I may implement it on la:. —Muke Tever 16:55, 2 Sep 2004 (UTC)
On the nl:wiktionary there is a difference between translations and articles about words.
  1. {{xx}} is exclusively used with translations.
  2. {{-xx-}} is exclusively used to indicate information on a word in that language.
These two templates, are both used in nl:wiktionary one template is not used in the other. Why would they?
The two templates can refer to the same content. That is up to you. The point is that the technically the usage is different. By making a distinction between the two you will not get double entries and you will only get the words for which there is an article.
I hope that from reading this you appreciate that there is a reason why there are both xx and -xx- templates. Thanks, GerardM 17:27, 2 Sep 2004 (UTC)
I still do not understand how a translation is somehow not "information on a word". It is, in fact, the definition of a word: Having "{{es}}: [[tres]] on a page called Three is no less defining the word than having "Three" on a page Tres with an {{-es-}} header. —Muke Tever 22:33, 2 Sep 2004 (UTC)
The xx template INDICATES a translation in a specific language. the -xx- template INDICATES a word in a specific language. The indicator is universal, the content is wiktionary specific. The useage means that you have a discreet list of all translations and a discreet list of words in the xx language.
This is even less clear :(
The thing is the template is there for a technical reason: {{en}} identifies as well as "English" does. However a template can be used in another wiktionary and it will say ᏲᏁᎦ/ᎩᎢᏏ in the chr:wiktionary. It will make perfect sense if you know Cherokee. GerardM 23:22, 2 Sep 2004 (UTC)
I think the latter word should be ᎩᎵᏏ gilisi, not ᎩᎢᏏ giisi. (I updated wiktionary:en:English to reflect that. If it's been copied anywhere else it might want to be fixed there also.) The yonega word seems okay though. (I understand the difference between them to be like the difference between 英文 (yingwen) and 英语 (yingyu) in Chinese.) In any case, I don't see {{en}} and {{-en-}} having different content in the same Wiktionary, even if it is Cherokee. If there is a counterexample I would like to see it. —Muke Tever 01:28, 3 Sep 2004 (UTC)
You are looking at it from the wrong angle. It is not about what its content is, it is what it is used for. The content on a wiktionary can be exactly the same. It must be correct, but the one thing that is relevant is what it is used for: indicating a translation or indicating a word. This means that it its use is technical; as said earlier it can be used to generate a list with all the words in a language.
The angle that I'm trying to explain is that all words in a dictionary are translations. The only difference between what you call a "word" and what you are call a "translation" is the format. And I don't see such a differentiation between these formats being useful. —Muke Tever 13:15, 3 Sep 2004 (UTC)

Have a look at the article wikt:nl:Nederlands, there are two occurences of the -xx- template. One to indicate the beginning of Nederlands the other for the Afrikaans the description of the word Nederlands.

There are some 67 occurrences of the xx template for the translation of the meaning of the noun Nederlands. Functionally they are different. On the en:wiktionary, you can not tell me how many Dutch or English words there are defined or how many words have a translation in Cherokee, Thai, Tamil or whichever. This can be done on nl: THAT is the function. Dutch words, English words, Cherokee translations, Thai translations and Tamil translations.

I try to make it as clear as possible, it is not about the content but about what the template can be used for. GerardM 13:55, 3 Sep 2004 (UTC)

I understand what it is these templates are doing, but I am not understanding why you want them separated. If I am looking for words in Cherokee (perhaps I want to contribute by spell-checking Cherokee that other people have added) I have to look at both -chr- whatlinkshere and chr whatlinkshere—and I don't need two separate lists. (Heck, with the list of Cherokee words divided between both lists, it'll be easier for me to just put "Cherokee" in the search box and go from those results.)
If all you want is statistics about "how many Dutch or English words there are defined" or whatever, then what you want to use is Categories, which will add it up and format it alphabetically and in a nice table for you, and you can add a better description than "De volgende pagina's verwijzen hiernaartoe".
If you have a different way you expect this list to be useful, I would like to know what it is. I am still not seeing a functional difference between those two or the need to have separate lists. Both wikt:nl:nederlands and wikt:nl:ᏁᏓᎳ tell me what Cherokee "ᏁᏓᎳ" (nedala) means. —Muke Tever 19:37, 3 Sep 2004 (UTC)
Read the article, it is a kludge. Categories do not do the job, as they do not translate between languages. I do not get statistics out of it I get a formatted page that looks good, I get data that can be massaged into a list. I only get a word in both instances when they occur both as an article and as a translation. On nl:wiktionary there are only -nl- and no nl occurrences. This gets me a page that can be copied across that will be functionally correct but missing the definitions of the word in the local language. This prevents editors particularly in the younger wiktionarues from doing a *load* of unnecessary work. GerardM 21:24, 3 Sep 2004 (UTC)
Categories translate perfectly well. In Template:-nl- you will put, besides the text =Nederlands=, also [[Category:Nederlands]], and all pages with {{-nl-}} will appear in Category:Nederlands. On en: the same template would have =Dutch= and [[Category:Dutch]], and all pages with {{-nl-}} will appear in Category:Dutch. No problem.
And there are several occurrences of {{nl}} on nl:. How do they interfere with your scheme?
Can you give a concrete example where a user will need to see the difference between -xx- and xx templates? —Muke Tever 00:25, 4 Sep 2004 (UTC)
Ok, categories translate. This would however result in a categorie with tens of thousands of entries. A definite nono technically speaking. It will surely cost much of the database capacity. Again, what I propose is a kludge and it will be replaced when the database is changed to accomodat certain things.
Okay; contrariwise, the way you have it now you are working with a whatlinkshere page with tens of thousands of entries—is this much better? The very nature of the information you are trying to organize means that you will have a lot of data. The kind of thing you want done is what categories are for. The only problem is that Wiktionary is aiming to have thousands more pages than Wikipedia will ever have, and yes, it will lead to huge categories; perhaps this will give developers to rethink the category interface and make it better.
A user needs to see a diffence in layout between the two. I need the -xx- list to generate a list with words in a language. Do you appreciate that what I propose works in the way I describe it ? (Proof of the pudding can be found on nl:) Thanks, GerardM 07:30, 4 Sep 2004 (UTC)
All right, I see what you are trying to do now. However I suggest that in any case your -xx- templates should include the xx template for the language's name, instead of storing it separately. —Muke Tever 13:56, 4 Sep 2004 (UTC)

As to Cherokee, the words are from the Cherokee wikipedia. Typically they should know best, they speak the language. Beside that, the content was copied TO the English wiktionary. Thanks, GerardM 06:14, 3 Sep 2004 (UTC)
Cherokee speakers can make typos too. (Especially as, if I understand correctly, there are no native speakers working on the chr: Wikipedia.) The stub article for English, though, is at the properly-spelled chr:ᎩᎵᏏ gilisi, not chr:ᎩᎢᏏ giisi. I'll fix chr:Main page. —Muke Tever 13:15, 3 Sep 2004 (UTC)

Thanks, GerardM 13:55, 3 Sep 2004 (UTC)

Sorting the translations[edit]

I am aware that the translations should be sorted according to the way needed in a particular wiktionary. For me this has a lower priority then getting what I am currently working on which is promoting the scheme, content and the uploading of two glossaries. Anybody is welcome to write the code, but untill it is done, it is on my "todo" list. :) GerardM 17:04, 2 Sep 2004 (UTC)

ISO 639 codes (and SIL)[edit]

ISO 639 codes are all well and good but do not cover all the languages Wiktionary should cover. Before it becomes a large issue I suggest that SIL codes (such as those used in the Ethnologue) be used for languages not in ISO 639. The semiofficial designation for using these in ISO 639 is x-sil-XXX, which is clunky; I suggest using the codes plain, and to distinguish them from ISO-based templates which which they might conflict their titles should be in all caps (thus for example Template:SPQ for Loreto-Ucayali Spanish). However this might not be useful should case-insensitivity be implemented; perhaps a different, yet still compact notation might be used. —Muke Tever 17:07, 2 Sep 2004 (UTC)

There is both discussion on the validity of the ISO 639 codes and the SIL codes. The first thing to agree upon is to use codes in the fist place: both the xx and the -xx- templates. Also many language with a ISO 639 code are not covered in any wiktionary. My first project is to cover these as much as possible and have as many translations available as possible.
Yes, ISO 639 is not perfect. However these can be added, certainly when because of things like GEMET we may get us into a position to request the inclusion for codes in the ISO 639. Untill that time even with SIL, life is not perfect. GerardM 17:33, 2 Sep 2004 (UTC)
This template will point from the discussion pages of all the different proposals for a single Wiktionary DataBase to the one page where all discussion on the subject of a single Wiktionary Database is conducted, to create a discussion of that purpose, rather than of each proposal separately. User:Aliter