[mapserver-dev] call for vote on rfc 48

[thomas bonfort]

Daniel,

> I'd suggest rewriting the "Detailed functionality" section to make it
> clearer which operations will be in the initial implementation and which
> ones won't be.

ok. I'll do that, at least for the documentation part.

> Also, in the current doc, the parameters have [geom] in
> square brackets, suggesting that this param is either optional, or the
> square bracket is required, but AFAIK that's not the case and 'geom' should
> always be passed as a reserved keyword to the operators... so I suggest we
> rewrite that section as follows:

the square brackets are required, as they denote an argument that
depends on the data query. The aim would be to stay consistent if some
day we have something like buffer([geom],[distance]). I agree that the
notation is a bit awkward as "geom" has no linkage with the actual
data request.
the initial implementation actually skips this and acts more like the
PROCESSING parsing for raster layers, i.e.
strncmp(expression,"start",5) , as the need for a full blown parser
was out of the scope of this rfc.

[snip]

bbox will be one of the implemented transforms to.

thanks for the feedback,

thomas