[postgis-devel] Raster Support

[Paul Ramsey]

Jaoa,
  I had an opportunity to speak with Gilberto on the phone while he was
up here in BC last week. He told me about INPE, and some of the (rather
amazing) stuff you are doing and have done. We would like to re-iterate
our invitation to INPE to be involved in the development of PostGIS and
advanced functionality in PostGIS. I understand the TerraLib project
will make use of PostGIS, and we definately want to be able to support
it!
  The major development you will see discussed on this list primarily is
the "Topological Operators for PostGIS" project being done with the
University of Victoria. We hope to have a basic object model done, and
code into CVS within two weeks. At that point it is possible that we can
parallelize some of the development of various portions of the project,
and INPE developers could help. When we get to that point, it is
possible that we can look over what is to be done and find chunks which
can be done in parallel. (Perhaps, Martin, you could think about which
chunks of the binary predicates are separable.) 
  The other possibility which Gilberto and I discussed was that INPE
take up the raster support in PostGIS and advance it. Right now, the
support consists of the CHIP structure, but there are only a few
functions which support CHIP, and no direct indexing support on it (we
have been indexing with proxy columns for our own internal purposes).
Things like a GDAL->PostGIS(CHIP) importer/exporter, a tie in from CHIP
to TerraLib, some internal functions for handling chip databases (like
mosaic(CHIP)), a tie in to OSSIM, could all the important additions to
raster handling. As a remote sensing organization, you probably have a
better feel for the potential of raster database storage than we do.
  Fundamentally, I hope to get INPE involved in development one way or
another as soon as you want to get involved! 
  I think INPE should (a) decide what it wants to be involved in (b)
outline some development goals. We are open to providing full CVS access
to the code base. Our only stricture is that changes be incremental and
non-destructive (no "I am just tearing it apart for a while and will put
it back together real soon when my new design works" development) so
that everyone can work together in harmony (changing structures without
discussion is a no-no :).

Looking forward to great things,
Paul


-- 
      __
     /
     | Paul Ramsey
     | Refractions Research
     | Email: pramsey at refractions.net
     | Phone: (250) 885-0632
     \_