bkml wrote:
> On the system level there may then be another database file which
> would contain definitions of languages and countries/territories, for
> example ...
>
> Languages = {
> ...
> Romantsch = rm;
> Raeto-Romance = rm;
> ...
> };
>
> CountriesAndTerritories = {
> ...
> Japan = JP;
> Nihon = JP;
> Nippon = JP;
> ...
> };
The requirement for these second lookup tables seems to only be caused
by the fact that you don't use "rm_JP" in the first place.
If what you describe is just metadata, and fairly loosely coupled
metadata at that, I look forward to writing short stories in the
language field, and then joining them to an ISO code in the other lookup
table ;-)
Languages = {
...
Drunken-Friday-Nightish = es;
Hung-over-Saturday-Morningish = de;
When-The-Girlfriend-Calls = fr;
...
};
> By looking up the user level and system level dictionaries, the
> system would then be able to translate the above into "rm_JP" and use
> that as a locale code internally.
>
> Just an *example* of how things could be abstracted and meet both
> user level and system level objectives.
Do we need so many abstraction layers? I mean, are we losing the plot
here? We are coding a PBX afterall... you know, that thing that connects
calls so that people can talk to people (not just computers).
If you want to code up this stuff for a hobby, by all means go ahead. I
just think it's over-engineering the solution a bit at this stage. I
think just getting some concept of multiple, regional versions of
english/german/french/whatever is more pressing. This can be done with a
minimum of coding by just enhancing the existing two-letter language
code support to a xx_YY locale code. This would also be a handy
transitional step to the more complex solution you propose.
THEN we can get all fancy pants with XML metadata.
Remember, KISS.
_______________________________________________
Openpbx-dev mailing list
[email protected]
http://lists.openpbx.org/mailman/listinfo/openpbx-dev
