Hi,
Sorry to be asking a series of unpleasing questions, but I'm realizing,
translating language variables, that there is a lot of duplication of terms.
For example "Last supplier's invoices" (and a series of incorrect
variations) appear in several different files (I also count
InvoiceDeletedDolibarr, Customer,
This reveals an issue that we just recently fixed in Chamilo: multiple
language files might not really be as beneficial as they may seem.
These were the issues with multiple files that we think we have solved
my moving everything to one file in Chamilo.
1. opening files (I/O) is expensive (although the expense can be reduced
by opcode caching files, we have to assume that not all Dolibarr setups
use opcode cache)
2. the decision of which files to include is complex for the system
(part of Translate::Load() if I'm not mistaken)
3. the decision of which files to include is complex for the developer
(if he wants *just one* term from another file, he needs to include the
whole file - if this is "admin", then he needs to include 1600 terms)
4. gettext (the only real standard in terms of translations) doesn't
play too well with multiple files
5. there is some level of duplication, because people do not look at all
files, which also means translators have to translate the same thing
several times, with the doubt of having seen that variable someplace else.
On the other side, advantages of multiple language files are only:
1. Reducing the number of variables stored in memory for any given
script loaded
2. Giving meaning to language terms (because they are in such or such
file, then they are obviously meaning something more specific)
I'm not asking for a straight answer but I'd like more information on
that, to know if this is something that is continued from the past, or
if it has been reconsidered recently.
Cheers,
Yannick
_______________________________________________
Dolibarr-dev mailing list
Dolibarr-dev@nongnu.org
https://lists.nongnu.org/mailman/listinfo/dolibarr-dev
