Configuration API

[Sean Middleditch]

On Tue, 2004-04-27 at 17:59 -0300, Avi Alkalay wrote:
> Maybe I get lost in translation, but I think I agree with your "configurable
> is the *wrong* way". It can't make everybody happy. Specially in the level
> discussed here (change notification, etc). And I think non-desktop software
> are also important for this subject.

Configurable is very important.  Mandatory, in fact.  Havoc's point (I
believe, he's probably better suited to clarify ~,^ ) is that having
this new theoretical API have a configurable backend that then layers
onto another API with configurable backend is just silly.  Configurable
backends are needed because no one backend is ever, ever going to work
for everyone.  A simple file backend may work great for your average
desktop or simple server, but something like an LDAP backend may be
necessary for large desktop deployments, and a cluster-oriented backend
may be useful for other specialized tasks.  Small systems may not be
able to fit D-BUS and change notification features in, but apps may wish
to be portable between a hand held and a workstation with a minimal of
fuss.

I also rather disagree with Havoc that GConf is "too complex."  More
just that it has some implementation warts from being a first
attempt.  ;-)  There are some things that GConf _makes_ complex simply
because GConf itself doesn't have enough functionality, such as storing
more complex preference structures.  The API is also a little more
complex than necessary for the common case.  (I'm a strong believer of
one low-level powerful API and where necessary a second layer simpler
API that wraps the former both being standard to account for all needs.)

> 
> I'd like to have the chance to have you explain me better this part, in
> private, if you have time.
> 
> Regards,
> Avi
> 
> 
> ----- Original Message ----- 
> From: "Havoc Pennington" <hp at redhat.com>
> To: "Thomas Leonard" <tal00r at ecs.soton.ac.uk>
> Cc: <xdg at freedesktop.org>
> Sent: Tuesday, April 27, 2004 4:25 PM
> Subject: Re: Configuration API
> 
> 
> > On Tue, 2004-04-27 at 10:59, Thomas Leonard wrote:
> > > If we make an API, then people can get started writing backends for it.
> > > Eg, someone could make a gconf backend and all programs using the API
> > > would get their settings stored using gconf. Someone else could make a
> > > Linux Registry backend, or a D-BUS daemon, etc.
> >
> > If gconf and the D-BUS daemon and so forth already have pluggable
> > backends, why would you layer an additional abstraction API on top?
> >
> > Or put differently, if you had the abstraction API on top why would
> > gconf or the D-BUS thing be pluggable?
> >
> > IMO the lesson from gconf is "slightly too complex" and the lesson from
> > straight to text files is "not quite complex enough" and the happy
> > medium is in between. But refusing to choose one is inherently more
> > complex than either alone. Making it configurable is the *wrong* way to
> > compromise, it's design-by-committee yuck.
> >
> > Havoc
> 
> 
> _______________________________________________
> xdg mailing list
> xdg at freedesktop.org
> http://freedesktop.org/mailman/listinfo/xdg
>