2012/4/9 Jaak Laineste (Nutiteq) <[email protected]>:
> Kuupäeval 9. aprill 2012 22:31 kirjutas Stefan Keller <[email protected]>:
>> 2012/4/9 Tom MacWright <[email protected]>:
>>> The GEOS dependency is not a long-term challenge, nor spatialite's licensing
>>> - you could write a clean-room implementation of both and they would be
>>> much-appreciated. But I don't think anyone _will_ write a clean-room
>>> implementation unless there's a good reason to, and thus the chicken-and-egg
>>> problem of adoption versus implementation. Standardizing abstract file
>>> formats without easily-used implementations is not useful: formats need to
>>> prove their usefulness through real-world success, or everyone will
>>> implement their own.
>>
>> To understand you correctly: I only need to implement a
>> parser/reader/writer of Spatialite's way to put WKB (and the two
>> metadata tables) into the existing sqlite file format.
>> OGR and QGIS don't need to change anything.
>
>  Almost. Just Spatialite uses internally own BLOB for geography, not WKB [1].

Ok, own BLOB (depends who defined what "well defined" means in the future :->)

>  To make bbox based queries in OGR and QGIS faster you would want to
> generate index data also. Again - I doubt if this is documented
> really. Shapefile situation with indexes is also not so good: some
> have added own index for live shapefile usage, as public shapefile
> spec does not reveal this. So a bit different cases and solutions for
> live data usage and bulk loading/conversion cases.
>
> [1] http://www.gaia-gis.it/gaia-sins/BLOB-Geometry.html

Good point I left out up to here: The index file and format.
Before all, I don't consider this as a show stopper since an indexes
can always be derived and recreated from the original tables.
Of course being able to optinally read/write indexes would be a plus
and indeed the OGR (any many other) implementations should avoid being
dependent on Spatialite libraries.

Frank wrote:
> My *ambition* was a well specified schema in Sqlite so that
> apps could depend on it for interoperability with or without having
> spatiallite available.  However the spatialite project didn't seem to
> want to get tied down to something stable and I lost interest.
>
> I still like the idea of sqlite as a geodatabase but I haven't the
> fortitude to advocate strongly for how I think it ought to work.

I did'nt realize that. What do you think should be the next things to
be sorted out?

-S.

_______________________________________________
Geowanking mailing list
[email protected]
http://geowanking.org/mailman/listinfo/geowanking_geowanking.org

Reply via email to