About [[MediaWiki:Narayam-ta-99/hi]]
This and other messages containing numerals tend to create a problem for hi translation (and probably other indic languages too). Although indic languages have their own numeral systems, these are not as popular as the arabic numerals.
Most indic language wmf projects use arabic numerals, some use native numerals. It is possible that different projects in a single language use different numerals (e.g hi-wp uses arabic while hi-wikt uses devanagari numerals).
So some people using hi, for example, would want to see devanagari numerals while others would want to see arabic numerals. Wherever the numeral is in the form of a parameter its not a problem (mediawiki handles what to do), but where the numerals come explicitly within the messages, we have to choose one numeral system over the other, which leaves one side or the other annoyed.
Ideally, I'd like it if I had something simple that took care of this, like {{int:1}} or something. Is something like this currently possible/can this be made possible?
Thanks
{{FORMATNUM:$1}} should do it (when wikitext and magic words are allowed, which is not always the case)
Ok, it seems to be working with this message. Is there a way to know if this will work with a particular message without having to ask here for each one? Thanks.
Not really, unless you are not scared of reading the program code. Or you can test if you can see the message in real action here in translatewiki.net (don't trust the preview of translations).
Testing here wouldn't be possible for all extensions (or even for core for that matter), since:
- Most messages don't contain precise documentation for where to find them
- Some messages are used in places with restricted access, which will not be accessible to translators here
I'm not excited about having to learn programming languages for translating.
So there's no solution for this currently? If not, should I go ahead and file a bug for this in BZ? Is TWN covered there?
Even worse. Since this wiki has only one specific one of the various possible general settings, you would only be able to see that your test apears to work for this specific one. This is not really sufficient.
I believe that such tests could be made on LABS, but I am not certain. I am still learning about it.
What about messages like MediaWiki:Narayam-he-standard-2011-extonly/hi ? Using formatnum caused a comma in between and using formatnum with r parameter caused it to go back to arabic numerals.
Solution?
The formatnum magic word does not expose the ability to avoid adding thousand separators (MW itself can do it). There is no fix for this, except the very ugly workaround of using formatnum twice with two numbers each time.
Any potential problems with/opposition to being "very ugly" if required?
Why use formatnum *at all* in these cases? It's static text. Format the number however you like.
It's explained in the first comment. The local numbering system is not universally used and the number transliteration could potentially be turned off in some wikies but not in others. I don't know if that is done in practice.
In practice:
- hi-wikt: devanagari numerals http://hi.wiktionary.org
- hi-wp: arabic numerals http://hi.wikipedia.org
You need more/other examples?
Filed bugzilla:35753 for this issue. Look like 'fa' also wants this.
Thanks. I just hope it gets fixed soon :)