Re: [gnome-db] post-GUADEC plans

[Adam Williams <adam morrison-ind com>]

> I've been this week in Dublin for GUADEC (http://www.guadec.org), and I
> come with very good news for us.
> I've been talking with Martin Sevior, from Abiword, about a GNOME Office
> release in August. Because of that, we talked about the following
> things:

Curious, does Abiword support GNOME-Db yet, for things like mail merge,
etc....?

> * libgda/libgnomedb should reach a 1.0 version at that time. And,
> looking at the current status, it seems we can punt some functionality
> originally planned for 1.0 and concentrate on stabilizing the current
> codebase, writing documentation, fixing crashes, etc.

This would be very nice for us primarily-end-user types.  I recently
presented an updated version of my "Data Access" presentation,  but with
much reduced GNOME-Db emphasis since the documentation seems a
out-of-sync with the tools, and the tools really buggy (mostly mergeant,
which is about the only GNOME-Db enabled tool other than Gnumeric).  I
hope to create a GNOME-Db specific presentation later in the year.

> Thinking about this, it seems to me a good idea, since I have to agree
> that the objectives for 1.0 have been large too much (reports, graphics,
> lots of providers, etc). So, freezing now (or in a few weeks time so
> that some new features can be added for a little while) could help us in
> making more widely known our work. Also, it seems to me
> libgda/libgnomedb (5 years old) really need a 1.0 ASAP.

Yep.

> * unfortunately, I don't think Mergeant is ready for 1.0 yet, so all
> this would mean that we would decouple libgda/libgnomedb and mergeant to
> be separated releases. A worst consequence of this could be that

I agree. Current mergeant seems less stable than the final gASQL was. 
It hangs up ALOT, at least for me.

But defining data sources is currently handled via mergeant,  would that
part be broken out into a little app akin to unixODBC's ODBCConfig.

> Mergeant is not included in the GNOME Office release. Or do you think
> with a bit of bugfixing mergeant, with its current features, can be
> released as a 1.0 software?

I'd love to see mergeant stabilized, but I'll be happy with whatever. 
Is there any tool that makes it simple for the layman to submit
debugging information on hangs?  Like Bug Buddy does when an application
crashes.

> * we would need to take advantage before gnumeric/abiword freeze for the
> August release to come up with some new DB-oriented features in those 2
> programs. Both use libgda now, but I guess they could use it for more
> cool things that would allow users to operate with their databases from
> these apps. The only problem is that I'm out of imagination on this

I don't know if you typically use Open Office, but the database features
there (via unixODBC) are really good.  I'd advocate using that as a
model.  The ability to drag data from a source into a text document as a
formatted table, or into a spreadsheet, is quite addicting.

> topic, so I'd appreciate having ideas from gnumeric/abiword users (I'll
> post a message about this to their lists as soon as we agree on the 1.0
> plan) so, what do you think?

I'll build the RPMs. :)  [Still need to test the spec files from CVS,
hopefully tomorrow... hangs head in shame...]

> Another good news :-) The people from igalia (http://www.igalia.com)
> have used libgda for their ERP app (http://www.fisterra.org) after
> having tried with Java, and libgda was 40/50 times quicker than the Java
> technologies being used (SOAP, JDBC IIRC what they said in the talk). So
> yea, we're ready for 1.0 :-)

Cool.