CLDR-18155 Order languageData's scripts by number of users #4237
Conversation
conradarcturus
force-pushed
the
CLDR-18155-Order-languageData-scripts-by-usage
branch
from
757c108 to
cbce758
Compare
|
Notice: the branch changed across the force-push!
~ Your Friendly Jira-GitHub PR Checker Bot |
| @@ -1346,7 +1346,7 @@ XXX Code for transations where no currency is involved | |||
| <language type="awa" scripts="Deva"/> | |||
| <language type="awa" territories="IN" alt="secondary"/> | |||
| <language type="ay" scripts="Latn" territories="BO"/> | |||
| <language type="az" scripts="Arab Cyrl Latn" territories="AZ"/> | |||
| <language type="az" scripts="Arab Latn Cyrl" territories="AZ"/> | |||
There was a problem hiding this comment.
likely subtags for az is az_Latn_AZ but I will note there are more Azerbaijani speakers in Iran (az_Arab_IR). It's just that they haven't transitioned to the internet like Azerbaijan has.
There was a problem hiding this comment.
While more speakers, it appears that there is more written in Latin (AZ) that Arabic (IR). So probably better Latn Arab Cyrl.
There was a problem hiding this comment.
This language points out a good flaw in population estimates alone -- and it depends on the usage.
Do people want likely subtags based on the largest corpora, for active internet users, for official government mandates, for any potential user (regardless of online status), for spoken content, for local content v global content (eg. AE's largest internet usage is en_AE not ar_AE).
An interesting idea (for LikelySubtags) could be to add alt="official" and potentially other alt tags when it's different. But yes, let's just stew on this thought and return to it if/when we have a good evidence it would be useful downstream.
There was a problem hiding this comment.
Strong agreement with what Mark said: despite spoken Azerbaijani being quite common in Iran with more speakers than in Azerbaijan, literacy in the writing system is pretty low in Iran, due to it not being formally taught in schools. There are books and magazines published in az-Arab in Iran, but most Iranian Azerbaijani speakers have difficulty reading them, even if they are literate in Persian (which is is theoretically a close writing system). So if anyone runs into written Azerbaijani in the wild these days, there's a much higher chance it's in Latin than in Arabic.
I also believe there's much more spoken "content" (read "media") produced in the language in Azerbaijan than in Iran (there's limited Azerbaijani radio and TV stations in Iran, and I believe less podcasts and songs produced in Iran). But either way, I don't think we want to tag such content as az-Arab or az-Latn. That's just az (or azb/azj if you insist on using ISO 639-3), and not az-Latn or az-Arab.
tools/cldr-code/src/main/java/org/unicode/cldr/util/SupplementalDataInfo.java
Outdated
Show resolved
Hide resolved
| for (String script : scripts) { | ||
| scriptsByPopulation.put(script, 0); | ||
| } | ||
| return setScripts(scriptsByPopulation); |
There was a problem hiding this comment.
Better to build once, make immutable.
There was a problem hiding this comment.
I know generally what you mean but I'm not sure how you want me to do this. Do you want me to avoid calling setScripts and include the code here as well? Personally I prefer DRY code over WET.
conradarcturus
force-pushed
the
CLDR-18155-Order-languageData-scripts-by-usage
branch
from
cbce758 to
f03fd8c
Compare
|
Notice: the branch changed across the force-push!
~ Your Friendly Jira-GitHub PR Checker Bot |
CLDR-18155
This sorts the scripts in the SupplementalData tags such that the first script has the highest population. This helps resolve some of the ambiguities interpreting the data.
ALLOW_MANY_COMMITS=true