[Java-collab] First contents in osgeom repository

[Jody Garnett]

Hi Martin:

I hope to compare the proposed interfaces against GeoAPI - since I do
not have a copy of ISO 19107 to check against. I believe you had a
student working with you who we invited to take part? Perhaps this
would be of interest. We can talk on the geoapi list about ways to
help - one of the big ones would be using to osgeo membership in OGC
to get the ISO documents in the hands of more people; or to give up on
the ISO specifications as a broken model for open collaboration.

To continue with feedback:
- I would like to try and define a Factory (or if people think it is
worth the risk we can forgo the use of interfaces)
- I would like to define the "Operation" interface which will lurk
behind the geometry objects packaging up the implementation. This is
an area where I would need to revisit JTS and the geotools geometry
module with my thinking hat on. And it would go much better with
working code/ test cases to refactor into position.

I will wait to here from the collab list before proceeding.

Cheers,
Jody

On Mon, Aug 10, 2009 at 11:39 PM, Martin
Desruisseaux<martin.desruisseaux at geomatys.fr> wrote:
> Hello all
>
> Just to remind that the offer still open. Since I noticed that many of
> Jody's comment about the proposed set of interfaces are inline with what
> GeoAPI does (some examples at the end of this email) - some of those
> resulting from years of experience, working with GeoAPI may still have
> technical interests. The geometry interfaces are in the "pending" directory
> precisely for incresing the freedom to adapt them according experience
> gained by implementations. The whole GeoAPI content is under OGC copyright.
> Since Degree is an OGC member, you can put your weight on the OGC SWG
> according OGC rules if you wish.
>
> This is just an offer - I'm not aiming to start a debate.
>
>        Regards,
>
>                Martin
>
> -------------------------------------------------------------
>
> Annex: examples of Jody's comment inline with what GeoAPI is today:
>
> * Check against ISO 19107 (correspondance is documented for every
>  interfaces/methods, and a summary is produced as part of the build
>  process)
>
> * Sequence of Point objects: PointArray
>
> * No getX() or getY(); instead getOrdinate(int)
>
> * Use of JSR-275 (which still having progress, admitly not very
>  fast but the JSR-275 team is still actively working on it).
>
> * Use of the ISO 19111-compliant "CoordinateReferenceSystem" name
>  (or CRS if abreviation is a must) instead of "CoordinateSystem".
>
> * CodeList as an open enum.
>
> * getSpan(int) instead of getWidth() and getHeight().
>
>