Page MenuHomePhabricator

Identify initial target languages for reading and search
Closed, ResolvedPublic

Description

What are the languages that need to be 100% for the initial KaiOS app launch?

Event Timeline

English
Hindi
Assamese
Bangla
Gujarati
Kannada
Kashmiri
Malayalam
Marathi
Nepali
Oriya
Punjabi
Sanskrit
Tamil
Telegu
Bodo
Maithili
Konkani
Manipuri
Dogri
French
Spanish
Italian
Portuguese

According to the current message group stats, Hindi and Marathi are 0% translated. We'll have to find translators for these.

SBisson edited projects, added Inuka-Team (Kanban); removed Inuka-Team.
SBisson moved this task from Backlog to Dev on the Inuka-Team (Kanban) board.

Yes, reaching out to community relations to support us on finding volunteers to assist.

One of the next steps here is to decide which languages we allow to search and to read articles.

Until we can support all languages, if ever, I propose we start a whitelist of languages that respect the following

  1. Input method works on the device. It is possible to type search terms in this language using the device keyboard.
  2. Display of article works well enough. Acceptable fonts available.
  3. Fully, or mostly, translated on TWN. [1]

Pinging @Amire80 for input

[1] As new features are added and new keys are introduced. A language may go from 100% to 90% and back to 100% when the translators catch up. It is possible to dynamically detect if all keys are translated and enable/disable the language accordingly.

One of the next steps here is to decide which languages we allow to search and to read articles.

Until we can support all languages, if ever, I propose we start a whitelist of languages that respect the following

  1. Input method works on the device. It is possible to type search terms in this language using the device keyboard.
  2. Display of article works well enough. Acceptable fonts available.

Both of these make sense to me. Are you familiar with all the possible devices?

  1. Fully, or mostly, translated on TWN. [1]

Makes sense, but it's a really easy thing to achieve. The project doesn't have a lot of strings now, and most of the strings are very easy technically, so it's not supposed to take a long time to translate everything. For someone who can write in the language, a couple of hours is enough, including coffee breaks :)

It's also possible in translatewiki to encourage translations only to some languages. For example, if you know that you are never going to market the app to Italian speakers, you can discourage translation into it.

All that said, I'd say that you shouldn't block features. For example, excluding Gujarati because the device doesn't have a font for it is probably OK, but make sure that you don't forget to add it when it does become available. If this can be automated, it's even better. And don't totally block searching and reading in Italian even if you don't market to Italian speakers—it's totally imaginable that a Hindi speaker would occasionally want to peek at Italian. It's probably OK to show Italian with lower priority, though.

[...] Are you familiar with all the possible devices?

Somewhat, but a lot of them are locked so we can't install and test the app on them. We should receive test devices that are the target of the initial launch soon so we can test on them. We should probably start there.

  1. Fully, or mostly, translated on TWN. [1]

Makes sense, but it's a really easy thing to achieve. The project doesn't have a lot of strings now, and most of the strings are very easy technically, so it's not supposed to take a long time to translate everything. For someone who can write in the language, a couple of hours is enough, including coffee breaks :)

It's also possible in translatewiki to encourage translations only to some languages. For example, if you know that you are never going to market the app to Italian speakers, you can discourage translation into it.

Is there a way to reach out to users or groups on TWN and solicit translations for some languages?

All that said, I'd say that you shouldn't block features. For example, excluding Gujarati because the device doesn't have a font for it is probably OK, but make sure that you don't forget to add it when it does become available. If this can be automated, it's even better. And don't totally block searching and reading in Italian even if you don't market to Italian speakers—it's totally imaginable that a Hindi speaker would occasionally want to peek at Italian. It's probably OK to show Italian with lower priority, though.

For sure, we want it to be available in as many languages as possible even if they are not the official languages of our target regions. We can probably test and whitelist all LTR languages using latin or Cyrillic pretty easily.

  1. Fully, or mostly, translated on TWN. [1]

Makes sense, but it's a really easy thing to achieve. The project doesn't have a lot of strings now, and most of the strings are very easy technically, so it's not supposed to take a long time to translate everything. For someone who can write in the language, a couple of hours is enough, including coffee breaks :)

It's also possible in translatewiki to encourage translations only to some languages. For example, if you know that you are never going to market the app to Italian speakers, you can discourage translation into it.

Is there a way to reach out to users or groups on TWN and solicit translations for some languages?

Yes. Some ways:

  1. Check the portal for the language. For example https://translatewiki.net/wiki/Portal:Mr for Marathi. Contact the translators in the list.
  2. If this doesn't help, check who translates MediaWiki core to the same language. The translatewiki usernames appear at the top of the i18n/*.json files. (Not that sometimes they also include usernames of people who just make technical fixes without knowing the language, like me.)
  3. If the above doesn't help, ask me or community relations specialists for more advice :)

Updating here to add the list of languages we must support for our readers to read articles:
English
Hindi
Assamese
Bangla
Gujarati
Kannada
Kashmiri
Malayalam
Marathi
Nepali
Oriya
Punjabi
Sanskrit
Tamil
Telegu
Bodo
Maithili
Konkani
Manipuri
Dogri
French
Spanish
Italian
Portuguese

Readers should also (ideally) be able to search in the languages listed above.

AMuigai renamed this task from Identify initial target languages to Identify initial target languages for reading and search.Feb 6 2020, 8:55 PM
AMuigai moved this task from Dev to QA on the Inuka-Team (Kanban) board.