|Translation system problem
||[Sep. 17th, 2002|04:51 pm]
The problemWe are currently having a big problem in the translation system. In English you can easily go
The Russians have it a little more complex:
Similarly, they have an entirely different system again in Tagalog and possibly other languages.
My ideaMy idea is to create a system that allows the translators to enter as many translations as they need for a string that contains a number. We would need:
With this, when a new language is added to the mix, all that needs to be done is either assign it an existing number system, or create a new one. That shouldn’t be much work at all.
- A flag for each translation string to determine if it contains a number or not.
- A property of each language which states what number system this language uses (see below).
- Determine how many translations per string containing a number are needed for each such system. Perhaps a hash in langlib?
- For each such system, create a subroutine that decides which translation string to use for a particular number. Perhaps add subrefs to aforementioned global hash?
- Enhance the translation editor to display several input boxes to translators wherever appropriate. The database could store these strings in one, separated by some special character, perhaps a pipe (|), so we would have something like:
[[number]] hour ago|[[number]] hours ago
- Add the required functionality to BML::ml().
Please comment whether or not you like this and/or whether you think this is feasible. If this is accepted, I will post it to Zilla and start right working on it.
( Description of number systems behind this cutCollapse )