Re: [gnome-db] Re: question: in memory table representation: please help

[Neil Zanella <nzanella gmail com>]

On Mon, 01 Nov 2004 11:57:15 +1300, Dru <andru treshna com> wrote:

> Bond - building object network databases
> http://bond.treshna.com
> Linas started qof after looking at bond a couple of years ago. All these
> projects attempt to bypass
> the horrid task of building forms manually, and taking a glade xml file
> and making its widgets live
> database objects.

First of all, thank you for your replies... :-)

As I said, this is exactly what I'm doing, but I haven't finished it
yet, and I am
sure that when there will be done there will be lots of things in
bond, qof, etc...
that will not be in my API (for lack of time I suppose, and at least won't be in
the original release), but there will be other features unique to my API not
found at all in theirs.

> There is another vaguly similar implimentation been worked on by Daniel
> with gtk-perl
> http://enthalpy.homelinux.org/Gtk2-Ex-DBI/
> And there are dozens of web based solutions like this for doing web
> based forms.

I guess, but web browsers are slow, so I didn't want to do it this way.
I want to give the user maximum speed.

> Mergeant provides the envoriment to develop the databases, relationships
> etc.

As a matter of fact, I have already designed some databases but would like
to use mergeant. It just seems neater. Mergeant does for me what I have been
doing by hand.

Eventually though, I want to make it possible to create my database from
my GTK+ code. I don't know whether mergeant has this already. So here is
my suggestion for mergeant: a feature that does what libglade did to glade.
Allow the user to save the database tables into an XML file and load their
structure at runtime so that the whole database structure can be
instantiated from C code. Perhaps it's already there, not sure.
As soon as I can get the CVS version to compile I will test it.

> DWI is similar to libgda though simpilar and not as many databases
> supported.

The strength of libgda is that it supports so many databases. I am even
wondering whether I can issue SQL statements with libgda without having
any database at all. That would be a plus cause that way I can keep the
installation of a database server optional (i.e., only those users who are
going to anticipate having lots of data can install it, and the others can
do whatever they want if they don't need the speed. But does libgda
really have this? It would mean implementing an SQL parser to
search a database consisting of perhaps just a few files,
designed without efficiency in mind, which is OK for
small applications but not for big ones, but would be
a plus, cause coders can still use SQL).

Thank you so much for your feedback,

It is constructive. :-)