[bezema@lat-lon.de: Re: [martin.desruisseaux@geomatys.fr: Re:[MetaCRS] Introduction]]

[Frank Warmerdam]

Rutger Bezema wrote:
> Sorry, I might have been not very clear, what I meant was constructing a set of
> 'Parsers' or 'Providers' which would return a standardized format, with a simple
> but usefull interface. It doesn't matter than in what language such a provider
> would have been implemented.
> 
> This way, all srs-packages would only create a single mapping function which makes a
> call to such a provider and receives the standardized code (may it be xml, text
> or binary, but in any case standardized), so that each srs-project only needs to
> implement one parser to talk to all supported backends.

Rutger,

You make this sound like the backend providers are being invoked at runtime,
in which case I would think that implementation language would matter!

However, this is essentially what I'm proposing - but done in batch.  That
is, we construct a variety of tools which we can use to translate various
dictionaries into some common normalized form, and then we distribute these
and the various package make sure they can either read this normalized form
at runtime, or distribute their own dictionaries derived from this normalized
form.

Some dictionaries (ie. EPSG) will need to be retranslated fairly often.
Others may be fairly static (perhaps a one time utility to generate them).
And some (one?) may be essentially manually created with various special
case coordinate systems we want to support.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo, http://osgeo.org