request la_VA locale

[John D.]

Currently, there is no locale support for Latin (la_VA) in glibc or in ubuntu.

I have been running into a few issues in the past couple years as I develop some applications in PHP that have strings in a number of languages, among which is Latin. However since there is no Latin locale in the system, none of the strings in Latin can be handled through `gettext` or any kind of localization functions in PHP (such as the soon to be deprecated `strftime`, or the newer `IntlDateFormatter`, just to mention a couple examples). Which means I can handle all translatable strings through `gettext` with the exception of Latin, which I have to hardcode into my applications. This makes for an application that becomes hard to maintain, it would be so much easier to just be able to handle any Latin translations just like any other language that is supported by the application.

Here are some of the things I kept in mind in preparing this locale file:

1) since the Vatican doesn't have streets and street numbers, and any mail going to the Vatican needs simply have an indication of a personal name and a department, followed by the zip (00120) and the country name (generally "Città del Vaticano" in Italian is used, so that's what I put as 'country_name' under 'LC_ADDRESS'. Keeping all this in mind I simplified the 'postal_fmt' control characters.

2) Generally anyone being addressed at the Vatican is either the Pope, a Cardinal, a Bishop, a Monsignor, or the head of a department (will often use a title such as "Dottore"), so I formatted 'LC_NAME' with title, name and surname.

3) Yes and No in Latin are expressed as "Sic" and "Non".

4) Monetarily, the Vatican uses the Euro, so this is the same as the Italian locale

5) LC_NUMERIC cannot effectively be defined correctly, because Latin, even ecclesiastical Latin, uses Roman numerals. However, I don't believe any kind of POSIX locale supports anything besides Arabic numerals in ascending order from 0 to 9. So to make this work, I just left it the same as the Italian locale.

6) For the days of the Week, ecclesiastical Latin in fact uses "Feria Secunda" or "Feria II" rather than the classical "Dies Lunae". Seeing that a practical application for this could be formatting Dates to be printed in texts such as the Roman Missal, and considering that in the Roman Missal the days of the week are printed with Roman numerals rather than in word form ("Feria II" rather than "Feria Secunda"), I opted for using the Roman numerals in the names of the days of the week.

7) I'm not sure I fully know the format for the 'LC_CTYPE' section, but I eyeballed the German locale to have an idea. Seeing that Latin has a few ligatures, I'm guessing they need to be defined?

Upstream glibc bug ticket: https://sourceware.org/bugzilla/show_bug.cgi?id=24366

[Nicolalucasbuescher+bugzilla (nicolalucasbuescher+bugzilla)]

Currently, there is no locale support for Latin (la_VA) in glibc. Since, iirc, the address, date, time, etc info for the Vatican is identical to those for Italy, this can easily be created based on the it_IT locale definition, with minor changes for the language and country.

[Nicolalucasbuescher+bugzilla (nicolalucasbuescher+bugzilla)]

Created attachment 11707
Create la_VA locale based on it_IT

And the corresponding patch.

[Jakub Jelinek (jakub-redhat)]

For la_VA, perhaps more appropriate would be to use the names of days used in the church, so Dominica, Feria Secunda, Feria Tertia, Feria Quarta, Feria Quinta, Feria Sexta and Sabbato.

[Digitalfreak (digitalfreak)]

I hesitate to add such locales. My point is that Latin language does not have native speakers. We will not make anybody's life easier supporting Latin language in Linux. I believe people who could be potential users of la_VA would rather choose their native languages in their computers, like Italian, Spanish, German, and so on.

However, I don't want to close this as "WONTFIX" yet because maybe I am missing something.

By the way, the patch contains some technical issues but I don't want to discuss them here because I don't want to suggest that I will accept the patch if they are corrected.

See also: bug 21919.

[Nicolalucasbuescher+bugzilla (nicolalucasbuescher+bugzilla)]

Rafal, that's certainly a fair point. My point of reference is that, while it may or may not be appropriate to compare the feature set of Linux to other OSs, in both macOS and Windows 10 you are able to set your locale/language to Latin (lingua latina). In both cases, the setting applies only to the locale used for applications, and Latin is not provided as a display language (what I would call a second-class locale). But this enables applications to support Latin as a language option.

Technically you can also do this in Linux without explicit glibc support, and I did, but the fact that there is no actual locale file crashes some programs, Dropbox being one. This fact that apps can actually crash is the reason I filed a bug.

I'm not deeply familiar with how locales work in Linux, but my thought was that having an la_VA locale would enable applications to support the language for spell-checking, which I think would be a very good use-case for the locale, considering that many many people still work with Latin, even though there are no native speakers.

[John D. (johnrdorazio)]

I would second the addition of a Latin locale. I have been running into a few issues in the past couple years as I develop some applications in PHP that have strings in a number of languages, among which is Latin. However since there is no Latin locale in the system, none of the strings in Latin can be handled through `gettext` or any kind of localization functions in PHP (such as the soon to be deprecated `strftime`, or the newer `IntlDateFormatter`, just to mention a couple examples). Which means I can handle all translatable strings through `gettext` with the exception of Latin, which I have to hardcode into my applications. This makes for an application that becomes hard to maintain, it would be so much easier to just be able to handle any Latin translations just like any other language that is supported by the application.

[John D. (johnrdorazio)]

Created attachment 13866
latin - Vatican locale definition

Compared to the previously proposed patch, here are some of the choices I made in this proposal:

1) since the Vatican doesn't have streets and street numbers, and any mail going to the Vatican needs simply have an indication of a personal name and a department, followed by the zip (00120) and the country name (generally "Città del Vaticano" in Italian is used, so that's what I put as 'country_name' under 'LC_ADDRESS'. Keeping all this in mind I simplified the 'postal_fmt' control characters.

2) Generally anyone being addressed at the Vatican is either the Pope, a Cardinal, a Bishop, a Monsignor, or the head of a department (will often use a title such as "Dottore"), so I formatted 'LC_NAME' with title, name and surname.

3) Yes and No in Latin are expressed as "Sic" and "Non".

4) Monetarily, the Vatican uses the Euro, so this is the same as the Italian locale

5) LC_NUMERIC cannot effectively be defined correctly, because Latin, even ecclesiastical Latin, uses Roman numerals. However, I don't believe any kind of POSIX locale supports anything besides Arabic numerals in ascending order from 0 to 9. So to make this work, I just left it the same as the Italian locale.

6) For the days of the Week, ecclesiastical Latin in fact uses "Feria Secunda" or "Feria II" rather than the classical "Dies Lunae". Seeing that a practical application for this could be formatting Dates to be printed in texts such as the Roman Missal, and considering that in the Roman Missal the days of the week are printed with Roman numerals rather than in word form ("Feria II" rather than "Feria Secunda"), I opted for using the Roman numerals in the names of the days of the week.

7) I'm not sure I fully know the format for the 'LC_CTYPE' section, but I eyeballed the German locale to have an idea. Seeing that Latin has a few ligatures, I'm guessing they need to be defined?

[Gunnar Hjalmarsson (gunnarhj)]

Hi John!

Thanks for your report. The locales provided by glibc are what Ubuntu use, so let's await the outcome of the upstream ticket.

[John D. (johnrdorazio)]

I'm also seeing that there has been some activity on CLDR for a definition of Latin among supported locales:

https://github.com/unicode-org/cldr/blob/main/seed/main/la.xml

However, I see that Ancient / Classical Latin has been used rather than Ecclesiastical Latin in some cases in the proposed definition. I'm checking in with the Pontifical Academy for the Latin Language ( https://www.pontificiaacademialatinitatis.org/ ) and with other academic resources such as the Pontifical Institute for Higher Latin at the Salesian University in Rome, for a consultation, so as to propose a correct definition that will be useful for modern uses. If I have any updates from this consultation, I will post them here (my understanding is that comments posted here will be synched back upstream?)

[John D. (johnrdorazio)]

I'm also seeing that there has been some activity on CLDR for a definition of Latin among supported locales:

https://github.com/unicode-org/cldr/blob/main/seed/main/la.xml

However, I see that Ancient / Classical Latin has been used rather than Ecclesiastical Latin in some cases in the proposed definition. I'm checking in with the Pontifical Academy for the Latin Language ( https://www.pontificiaacademialatinitatis.org/ ) and with other academic resources such as the Pontifical Institute for Higher Latin at the Salesian University in Rome, for a consultation, so as to propose a correct definition that will be useful for modern uses. If I have any updates from this consultation, I will post them here.

[Gunnar Hjalmarsson (gunnarhj)]

No, regular comments here won't be synced back upstream. It's the other way around.

[Mike FABIAN (mike-fabian)]

@John R. D'Orazio Do you have any updates?

[John D. (johnrdorazio)]

(In reply to Mike FABIAN from comment #8)
> @John R. D'Orazio Do you have any updates?

I've been keeping in touch with the Pontifical Academy for Latin, they've had a few changes internally between one director and another. I'm now in touch with the latest director and we're going to work out a meeting fairly soon, to look into the possibility of defining the Latin locale with the CLDR survey tool. As soon as I have any more news I'll post it here.