Re: libgda .... Re: GnuCash page on GO site

[Charles Goodwin <charlie xwt org>]

On Sat, 2004-03-06 at 17:07, Linas Vepstas wrote:
> But I thought the goal of libgda was to provide a set of high-level
> abstractions to multiple data sources, including sql and xml and ldap,
> which would imply that libgda is comparable to other high-level
> abstraction libraries.  But maybe that is libmergeant.

That is the goal of libgda.  But you're comparing this to the goal of
QOF which is to provide data-aware widgets, which libgda does not.  This
is our frustration; you don't understand libgda because you're not
prepared to, you're only looking at your own situation subjectively
rather than stepping back and being objective about your requirements.

In the last several days you've said:
- Your software is better than Gnome-DB
- Your only think of Gnome-DB as it was several years ago
- Your software is crap (!?)
- There are major problems with your software (!?)
- You are short on resources
- You will not compromise
- Your users are most important to you...
- ...yet you're release process disregards your users
  (will refer to this one later!)

This is in-amongst a verbose set of explanations of what QOF is,
repeatedly.  Then we say 'we know, but we're saying use libgda for data
source abstraction for QOF' and you then proceed to explain what QOF is
and why it is better than libgda.  But it's _different_ to libgda!?

We appreciate your time, we do not want to waste your time, but you've
got to stop talking with your heart and start talking with your head. 
We know you're proud of GnuCash, QOF and DWI and we're not trying to
detract from that.  We've only ever said that there are places where
libgda will save you time and effort (so you don't have to juggle with
XML and with SQL like you keep complaining about) so you can concentrate
on getting QOF and GnuCash working correctly.

Hell, even Andrew complained about his problems abstracting from data
sources, then writes off libgda (WHICH DOES JUST THAT!?) in the same
paragraph.  How do you expect Rodrigo to respond to that?

Nobody is saying QOF or GnuCash is bad (other than yourself?) or that we
don't care for bond.  We're just trying to take a _practical_ view of
where efforts overlap and can be shared.  It's not black and white. 
libgda, bond, and QOF are _not_ competing libraries.  So stop painting
it that way.

As for my comments on your goals for GnuCash; they're grand and
ambitious and admirable for that.  But when do you plan on releasing
GnuCash?  Why not set a string of short term goals that are achievable
whilst moving toward your long term ambitions?  Users (at least anybody
that I've talked to about it which is a fair number of people) want a
single-user Gnome2 version of GnuCash.  A multi-user SQL-using
super-application is not realistic before the end of this year.

There's no pragmatism about your perspective or any of the other GnuCash
developers whose opinions I've read.  Look at AbiWord and Gnumeric as
prime examples.  They set definite realistic goals and head towards
them, whilst diligently planning for the long term goals.  The result,
AbiWord 2.0.x is awesome but a few features short of Word; but they're
coming in 2.2 and 2.4.  Gnumeric 1.2.x is just an amazing set of
releases that should, frankly, be award winning and is arguably better
than competing products.  GnuCash... is flailing, still in Gtk1 mode
which is coming up to being 2 years off the pace.  Even when your
release it will be untested and problems will be quickly exposed by the
community as happens with all software.  OTOH, both AbiWord and Gnumeric
are well tested stable applications due to constant community review.

Rodrigo said it well when he spoke to me of GNOME Office: "Step by
step."  And that's what we're saying to you for GnuCash; step by step
you should look at ways to offload some of your resource problems by
reusing or working with what we have here (or what somebody has
somewhere else, or even with an independent route by planning what you
can accomplish today as well as tomorrow).  We're _open_ to input and
_open_ to improvement.  Rodrigo openly admits that libgnomedb or
libmergeant (which you should be comparing to QOF instead of libgda)
needs more widgets and has not, at any point, pushed it in front of you.

You are perhaps the least humble developer I've witnessed since I hit
the open source scene.  And that lack of humility is making you an
infuriating person to exchange (or witness exchanges) with because you
just don't listen.  We're not saying you're wrong, but stop assuming
we're wrong just because you are not.  Most of the time we're not even
disagreeing with you!

> I still can't find the libmergeant documentation.  Its not there.
> I looked. Its just a user's guide.

You couldn't have looked hard.

Rodrigo was talking about [wrt libmergeant documentation] the CVS
checkout of mergeant.  I would have thought somebody of your stated
calibre would have been able to figure that one out.  It took me less
than 3 minutes to find and I'm hardly a rocket scientist:

http://cvs.gnome.org/bonsai/rview.cgi?cvsroot=/cvs/gnome&dir=mergeant/doc/libmergeant/C

Although I'm not sure why it's in the 'C' directory here. ;)
-- 
- Charlie

Charles Goodwin <charlie xwt org>
Online @ http://www.charlietech.com