[asterisk-dev] [policy] Language specific prompts

[Tilghman Lesher]

On Friday 06 February 2009 14:10:32 Tzafrir Cohen wrote:
> On Fri, Feb 06, 2009 at 11:57:32AM -0600, Tilghman Lesher wrote:
> > Ever so often, we get requests on the bugtracker, stating that this or
> > that language is not pronounced correctly.  For English, Spanish, and
> > French, where we have standardized those prompts (and documented each),
> > this isn't a problem.  More recently, for someone who brought the Hebrew
> > language to Asterisk, I required the reporter to create an Open Document
> > Spreadsheet (commonly written by OpenOffice.org) detailing every single
> > prompt, the module in which it is required, and the native language
> > script, to be used by a native language speaker in creating a new set of
> > prompts.
>
> Off-topic: Is there a reason that document is not plain text?

Yes, because we want a standard format that can be easily read by just about
anybody.  If it were plaintext, we'd need to worry about things like document
encoding, something that is already taken care of by the OpenDocument
format.

> (Yes, I don't mind converting it myself. And Nir, I know you read this
> list :-)
>
> > I am proposing, as a matter of policy, that anybody wishing to make ANY
> > change whatsoever to multi-language support, that they should have to
> > document every single prompt in that native language, in a fashion that
> > is identical to the Hebrew document (which may be found in doc/lang/).
> >
> > Trying to get every source to agree on the contents of prompts is
> > incredibly difficult without having a set of documentation on which to
> > fall back, and I believe that having this documentation will go a long
> > ways towards making these kinds of issues more readily resolvable.
> >
> > Additionally, note that the Hebrew language prompts are written in a
> > non-Western script, so we have already found a solution which addresses
> > that particular issue.
>
> What would you like to see there? The non-western script seems a must
> for me to avoid ambigiouty.
>
> Do you want something in the lines of:
>
>   filename	meaning (English)	Words (original script)
>
> The text of the English sound files often spans more than one line per
> entry. E.g. 5 lines or so for the "codezone".

I think you should definitely open the sample hebrew.ods and examine it for
yourself.  The format is exactly as I specified, and I believe it fully and
completely meets our needs for documentation.  This isn't a specification
alone.  It has a completely valid sample from which others may propose their
own language documentation.

-- 
Tilghman