[GRASS5] Re: question about sqlite

[Hamish]

> AFAICT, the only advantage to the DBF driver is that we bundle it with
> GRASS. SQLite does essentially the same thing (i.e. both are
> server-less, zero-configuration DBMSs) but does a better job of it.
> 
> If it's essential to bundle at least one database backend, we could
> always bundle SQLite (the code is public domain).


While I think it is essential that we provide something that "works out
of the box", we do rely on e.g. GDAL as an external package without too
many hickups. If maintaining our own copy of SQLite within the GRASS
source is going to be anything like what maintaining our own copy of
GDAL/PROJ4 within GRASS 5 was like, I'd be a bit worried.

The nice thing about DBF is that it is simple, well known and widely
used, and ain't about to change anytime soon. Is SQLite mature and
stable enough both on its own and as part of grass for us to depend on
with some level of comfort? (as GDAL is)  ie let's avoid moving targets.

Note I am not making technical arguments against SQLite, as I'm not
qualified to comment on that. Just being conservative.



Hamish