Why we move to libdbi rather than libgda?

[Derek Atkins]

Phil Longstaff <plongstaff at rogers.com> writes:

> On Sat, 2010-05-22 at 22:49 +1000, Tao Wang wrote:
>> Hi,
>> 
>> When I look at the file 'packaging/win32/libdbi-0.8.3.patch', I found
>> the libdbi hasn't been actively developed for a long time, the last
>> release is 2 years ago. On the other hand, 'libgda' is keeping active
>> development, now is version 4. I also found the discussion in the
>> mailing list:
>> 
>> [DB abstraction layer: libdbi vs libgda]
>> https://lists.gnucash.org/pipermail/gnucash-devel/2006-October/018976.html
>> 
>> I looks we were going to move to libgda. However, the current
>> dependency shows GnuCash is using libdbi. So, what happened? What made
>> us move to the 'libdbi' rather than 'libgda'?
>
> There is a thread about libdbi vs libgda as well.  Basically, when I
> started the sql backend, I used libgda v3.  A few bugs, but usable.
> They then started developing libgda v4.  Given a new major release,
> interface changes, internal changes, I was running into more problems.
> In addition, there was the question of what version the various
> distributions would have.  In contrast, libdbi was more stable.
>
> There is still some activity going on with libdbi, but slowly.  I agree
> that a new release would be good.

The reason there hasn't been a release is because libdbi is pretty
stable and has been for a very long time.  It's not like MySQL or PG are
moving targets.  SQL has been around for a very long time.  libDBI
doesn't really need to be a moving target!  And if you look closely
you'll notice that there ARE still commits in their repository, so there
IS active development, but frankly they have everything they need.

On the other hand, GDA has been a mess.  It's been a completely moving
target.  They keep changing their API.  They can't decide what they want
to do.  GDA2 -> GDA3 -> GDA4 were effectively THREE different products
with three different APIs.  I'm sorry, but no, that's NOT acceptable.
Maybe when they get their act together and stabilize, then maybe we
could re-examine the sitaution.  But GDA was broken, buggy, didn't
have the support we required, and wasn't distributed in distributions.

LibDBI solved our problem.

Newer != Better.
Long time between releases != Worse.

> Phil

-derek
-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available