Module talk:Language/data/wp languages

From Deep web, the free encyclopedia
Jump to navigation Jump to search

A correction and an addition[edit]

Two changes to request. In the section commented ISO 639-2 codes, please correct line 65 from

["roa"] = {"Jèrriais"},

to

["roa"] = {"Romance"},

. Jèrriais has an ISO 639-3 code of nrf, shared with Guernésiais. Also, please add (in the same section) a new entry

["bat"] = {"Baltic"},

the ISO 639-2 code for the group of Baltic languages and as recommended in Template:PIE. It might be worth adding the other ISO 639-2 codes listed in Template:PIE also. — OwenBlacker (talk) 23:27, 24 December 2017 (UTC)

Not done, well, sort of  Done
I am hesitant to make fixes and especially additions to this module. Because the provenances of the data items in this module are unknown, I think that this module should be frozen and changes limited to fixing (by commenting out) code/name pairs that are unequivocally wrong (mla, nrm, nsd, and roa – this last an oversight on my part). Instead, we should add items to the override table in Module:Lang/data (as the name implies, data in the override table wins). If at all possible I would prefer to not make additions to the override table that don't agree with the international standards that define the language code/name pairs unless we have very good reason to do so.
roa already exists in Module:Lang/data though at present it isn't used by the module unless it is used by {{lang}} ({{lang-roa}} does not use Module:Lang). Similarly, bat already exists in Module:Language/data/iana_languages and works in {{lang}} (there is no {{lang-bat}}). On the still-to-do list for Module:Lang is to properly handle ISO 639-2 codes for collectives or collections. Because this is on the to-do-list, I would prefer to make no commitments about them just yet – the solution I think requires renaming existing categories and some special code in the modules.
I think that you will find that most, if not all, of the codes listed at {{PIE}} are listed in Module:Language/data/iana_languages.
I will add bat to Module:Lang/data so that if you use it, it will make a more or less sensible category name.
Trappist the monk (talk) 01:48, 25 December 2017 (UTC)

Edit request[edit]

I'm surprised I'm seeing this protected, knowing how suspect its data generally is. If this is going to remain being protected, could someone remove the entry for "naz" please? – Uanfala (talk) 14:37, 6 January 2018 (UTC)

The protection was added to act a an impediment to additions.
I guess I have to wonder if this change is immediately necessary. {{lang-naz}} has been nominated for deletion; shouldn't that process be allowed to run to completion before any changes are made?
Trappist the monk (talk) 15:25, 6 January 2018 (UTC)

Request to undo an edit[edit]

/w/index.php?title=Module:Language/data/wp_languages&oldid=724280169

This edit was done by an IP without a reason given, and is different from the name used in the Sinhala language article and other Deep web articles on Sinhala. Would a template editor please undo this edit? Danielklein (talk) 12:13, 8 April 2019 (UTC)

Are you wanting this change merely because 'Sinhalese' is not the name currently used at Sinhala language or because it is causing other problems? If the latter, what are those problems? Is there a consensus for this change? Sinhalese is an alternate name supported by all of ISO 639-1 (si), -2 (sin), and -3 (sin), and also by IANA.
I have no objections to making this change as long as there is consensus to make it.
Trappist the monk (talk) 13:16, 8 April 2019 (UTC)
You are correct that Sinhalese is an alternate name, however, the data this table comes from lists "Sinhala" as the main name. The main problem is (in)consistency. Articles on Sinhala call the language Sinhala, yet Sinhala text such as {{lang|si|කොහොමද}} කොහොමද (when hovered over) and {{lang-si|කොහොමද}} Sinhala: කොහොමද currently display the language as "Sinhalese", as well as linking to a redirect page, instead of directly to the appropriate language page as most other languages do. There is currently no consensus for either "Sinhala" or "Sinhalese", but there is consensus to use the original name (i.e. "Sinhala") until consensus about "Sinhala" vs "Sinhalese" is reached. See Talk:Sinhala language. This module should match its source data as it did when it was created, and if consensus is reached to use a different name, that alternate name should be specified in the override module, not here. If this module had been protected from arbitrary changes when the anonymous user wanted to make their change, it would still be called "Sinhala" here. Danielklein (talk) 02:27, 11 April 2019 (UTC)
the data this table comes from lists "Sinhala" as the main name. Are you saying by that declaration that you know the provenance of the language names in Module:Language/data/wp languages? If you are, please tell us so that it can be researched and documented.
Linking to an article through a redirect is the preferred method when, for whatever reason, the language name used by {{lang}} does not match en.wiki's article name. This method is preferred because redirects are well understood by most editors and there is procedural and technical support for maintenance of those redirects.
What is past is past at en.wiki. There is no point in saying 'if only ...'
Per the no-consensus consensus at the RM discussion, I have added si to Module:Lang/data so that the language name rendered by {{lang-si}} and the tool-tip provided by {{lang|si|...}} in your examples above should be 'Sinhala'.
Can we expect further discussion about how Module:Lang should deal with pre-1948 Sinhal(ese) text?
Trappist the monk (talk) 11:23, 11 April 2019 (UTC)
There is a comment at the top of the code that says ISO 639-1 above the column of two-letter language codes. When the two-letter codes end the three-letter codes start, there are two further comments in quick succession: ISO 639-2 codes and ISO 639-3 codes. The comments next to each entry match the column called "English name of Language" found here: https://www.loc.gov/standards/iso639-2/php/code_list.php, except where subsequently edited. The original data was sourced from the International Organization for Standardization's standard ISO-639, and should ideally still be synchronised to it, except if errors have been identified in the source data. I don't quite see the point of protecting the module to prevent future damage if you're unwilling to repair past damage. Danielklein (talk) 04:19, 12 April 2019 (UTC)
All of those comments were added by me; you can see that in this diff. Before that, there was nothing; see the old version that you want reverted. The module did not have a documentation page until I created it in December 2017 and where I have noted my best guess as to the source of the data; no one has been able to confirm or refute that guess. It is not true that all of those codes come from ISO 639; see the non-standard codes at the bottom of the module.
I am not unwilling to repair past damage. Where did you get that idea? The module is protected to prevent un-discussed changes (like the Sinhala / Sinhalese change). What I want for ~/wp languages is for it to go away. The modules that use ~/wp languages, can be, I think, weaned away to use something better: the ISO 639-1, -2, -3, -5 and IANA language modules (where the provenance of the data is known) or MediaWiki's built-in list (provenance not so certain but universal to all of the various wiki language editions and has all of MediaWiki's made-up language codes).
Trappist the monk (talk) 12:00, 12 April 2019 (UTC)
Getting rid of ~/wp languages might actually be a good idea, as long we set out to replace it with a comprehensive list of those cases where the name preferred on wikipedia differs from the one used by the ISO 630 custodians. What we need is a systematic way to keep the labels, category names, etc. used by the lang templates in sync with the titles of the corresponding wikipedia articles. And I take is as self-evident that this is a desirable thing to do: ISO 639 custodians are authorities on the denotation of the language codes (what code corresponds to what language), they're not authorities on how languages should be called and how their name should be spelt (and Trappist the monk, I really don't think it would have escaped your notice that whenever the issue has been brought up the last couple of years, you have been alone in insisting that we should disregard how languages are named on wikipedia and use the names exactly as provided by the ISO custodians). – Uanfala (talk) 22:58, 12 April 2019 (UTC)
I totally agree that weaning other modules off wp_languages and eventually deleting it would be a good thing. However, while it is in use, it should agree with ISO-639 as much as possible, since that seems to be where it came from originally. The fact that the two and three-letter codes (mostly?) match ISO-639 is a strong hint. As well as providing for accurate data in the present, it will allow a smooth transition to ISO-639 data when other modules make the transition. The change from "Sinhala" to "Sinhalese" was undiscussed and should be reverted. The change from "Sinhalese" to "Sinhala" has been discussed, and it has been agreed to change it back to "Sinhala" until such time as consensus for a different name occurs. Danielklein (talk) 03:59, 13 April 2019 (UTC)
Why do I think that we are not communicating? So far as Module:Lang is concerned, the Sinhala / Sinhalese issue is resolved:
{{lang|si|කොහොමද}}කොහොමද – float your mouse cursor over that and the tool tip should report 'Sinhala language text'
{{lang-si|කොහොමද}}Sinhala: කොහොමද
{{si icon}}‹See Tfd›(in Sinhala)
To accomplish this, I added si to the override table in Module:Lang/data where all other overrides are. I recognize that this 'fix' doesn't fit with exactly what you want but I see no reason to maintain something that should, ought to, will go away.
When I created the override table in Module:lang/data, I did it, in part, to preserve ~/wp languages as-is for other modules that use ~/wp languages so that whatever those modules were/are doing didn't get broken. Since that time I have come to realize that the modules that use ~/wp languages directly and indirectly are, for the most part, unused so that ~/wp languages can go away. The unique code-to-language mapping in ~/wp languages will transfer to Module:lang/data; code-to-language mapping in ~/data that are there to override ~/wp languages will be removed from the override table.
Trappist the monk (talk) 10:56, 13 April 2019 (UTC)
@Trappist the monk: Thank you for what you have done so far. However, it's only half fixed the problem. {{Lang|si|Sinhala text}} now goes under Category:Articles containing Sinhala-language text but {{Lang-si|Sinhala text}} still goes under Category:Articles containing Sinhalese-language text. Danielklein (talk) 01:37, 15 April 2019 (UTC)
I think that you are mistaken. Try this:
  1. copy this to your clip board: {{code|{{Lang-si|Sinhala text}}}}
  2. paste your clipboard into any mainspace article (must be mainspace because Module:lang does not categorize other namespaces)
  3. click Show preview; you should get this:
    [[Sinhala language|Sinhala]]: <span lang="si">Sinhala text</span>[[Category:Articles containing Sinhala-language text]]
If you get the expected response then the template is working as it should; if you do not, report that here or at Template talk:Lang
Remember that changes to templates are not necessarily instantaneously reflected in mainspace; a null edit is often required.
Trappist the monk (talk) 02:37, 15 April 2019 (UTC)