GNOME Database Kit Part II

[Adam Keys <adam pingpage com>]

Thank you very much for all the feedback.  I'm very excited about this
project and the response I've gotten from it.  This is a great community
of coders as far as I can see :).

Now for the meat!

ODBC is probably a bit large and unwieldy for our needs as far as I can
see.  It's well thought out, but there's nothing that stops us from
stealing their ideas...if the shoe fits, steal it!  I've been unable to
find info on the FreeODBC project, which looks like the most promising
lead in that direction.  If someone wants to send me a white paper or a
URL, I'd be enlightened to give them a look-see.  Right now my verdict
is start from scratch and steal as neccessary. :)

So far as interoperability between languages, I think I've come up with
a good middle ground.  For me, the most natural way to implement the
gdbk is in C++.  It's simply the only language that gives me the power
and features I need without bending my arms funny ways to get them.  The
issue with other languages, I beleive can be solved through the use of
CORBA.   I'm guessing that the languages available to GNOME programmers
will also have connections to ORBit.  If this is true, a simple
client-server mechanism can be set up, so that when an app needs to talk
to a database, it talks to the CORBA broker that has been written.  Does
this sound right to you guys?

I agree we don't want to tie the widgets specifically to anything.  The
development of the widgets would be more in the scope of gdbk
development than gtk development.  We'd just need to steal a gtk coder
for a week or two once we know what widgets we want and how we want them
to look/feel.

As far as apps, I like the screenshots I've seen of filer, and it's okay
with me for that to go its own direction.  Perhaps if this API becomes
mature filer can use it.  My ultimate goal is to see some kind of
Access/Paradox app written, and to attract application developers who
want database stuff in their apps.  It'd be nice if things like cddb
could have a full-blown mySQL or PostgreSQL backend to them and not some
custom database backend.  These are some of my application oriented
goals for the gdbk.

I'm working on a code demo to release to you chaps so that I can
demonstrate my ideas and the path I'd like to take in implementing this
API.  I'd hope to have it done within the week.  So far as everyone
else's code, perhaps if we could set up a CVS directory and I can
receive and upload people's ideas as they come?  Would this be desirable
at this point in the game?

Keep the responses coming

Adam Keys
So much to think about, so few brain cycles!