[geos-devel] ABI/API status and future development thougths

[strk at refractions.net]

On Thu, Apr 21, 2005 at 12:27:26PM -0400, Ted Macy wrote:
> Can I get a candid assessment here?   I'm new to GEOS.  And you guys are
> scaring me off from wanting to use it.  Here is my situation:

Your scare is an effect of people here being
sincere, as opposite to proprietary software vendors.

> I normally use the Cadcorp or Tatuk COM objects for our production work
> (using the published OGIS Simple Feature COM Interfaces).  However we have
> some public domain stuff that we share with the agriculture industry, and I
> cannot make other users license the components from these vendors.  I needed
> a COM compatible set of objects that that could be plugged in as an
> alternative to these two commercial options, and they would only convey
> geometry data between objects through a "Stable" API :).   Two years ago, I
> contracted a developer to implement a COM object using C++ that would
> implement the OGIS Simple Features COM interface specification.  The
> implementation only knew about WKT parsing and regeneration.  It had no
> other spatial functionality... buffering, intersection, etc.  
> 
> Now, I need my COM objects to implement some true spatial functionality.  I
> thought I might contract a developer to set my COM wrappers on top of the
> GEOS codebase and be good to go with a stand-alone set of COM-based geometry
> objects.  But all of this talk about breaking compatibility, performance,
> etc, etc... it is scaring me.   I don't write C++ code... I'm a C programmer
> and a power VB COM programmer.  I have to contract this project out.  Can I
> expect that the GEOS stuff is in a suitable state to make this project a
> reality?

It surely is, making wrappers is exactly the way I was suggesting to
obtain a stable API. I planned to do similar thing for C/C++ code.
GEOS already has a python interface with same abstraction
charateristics. This would be the COM interface.

> And, by the way... when done, I was planning to contribute the full COM
> implementation back to the community if anyone was interested in it.

This would be the best investiment in stability, which means whatever
will happen to GEOS maintainence of wrappers would be easier if shared.

--strk;

> 
> Thanks.
> Ted Macy 
> 
> 
> _______________________________________________
> geos-devel mailing list
> geos-devel at geos.refractions.net
> http://geos.refractions.net/mailman/listinfo/geos-devel