Re: MIME types, again ("Nautilus" ain't the answer)

on Wed, Feb 09, 2005 at 07:44:40AM -0500, Adam Hooper ([email protected]) wrote:
> On Tue, 2005-02-08 at 20:25 -0800, Karsten M. Self wrote:
> > With cause and provocation.  May I advise the always illuminating Jeff
> > Waugh In His Own Words:  An agony in seven fits?
> > 
> >     http://zgp.org/pipermail/linux-elitists/2004-January/008588.html
> > 
> > That post, and the thread to which it's posted, contain several and
> > repeated dismissals and insults from _core_ GNOME devs.  Waugh was (is?)
> > GNOME release manager, a fact I still find hard to believe.
> 
> Totally awesome. That made my day :).
> 
> As a humble side note: Jeff Waugh contributes a hell of a lot to GNOME.
> Flaming him *really* won't get you anywhere, and it does nothing but
> hurt the entire project.

Um.  If you actually _read_ that, you'll find the flamage *starts* with
Waugh.  Who then demands respect and polity.  Being told repeatedly that
I'm "not a GNOME user" when I am, in fact, making use of apps which use
GNOME features, was particularly galling.  This sort of doublespeak I
expect of Microsoft and the tobacco industry.  Not respected Free
Software advocates and developers.
 
> > 1.  ...which itself raises a few interesting points.  Somewhere in the
> >     Waugh fit linux-elitists thread, IIRC, the old "but the gconf
> >     registry *isn't* the same as the legacy MS Windows registry 'coz
> >     it's not a fragile binary structure that shits its pants when you
> >     look at it crossly" argument was trotted out.  While that is *a*
> >     fault of Microsoft's implementation, it's hardly the only:
> > 
> >       - Undocumented settings.
> >       - Indecipherable/obfuscated keys.
> >       - Indecipherable/obfuscated values.
> >       - Limited interface.
> 
> GConf does not share any of these properties with Microsoft's registry.

In the abovementioned Waugh quotes, upthread, there's a specific example
showing Evolution (the thread's long enough that it's not surprising
you'd miss it).  In which the data for its information panel weather and
related settings is base64 encoded for some as yet unknown reason (took
me a few years to find out what the encoding was).

You can claim that this "isn't a fault of gconf", but the system
_facilitates_ the problem.  To a markedly greater degree than typical
value-assignment config files (often with embedded comments) does.  In
large part because human-editied-and-read data tends to be written for
humans to read and understand.

Designing something that's trivial to abuse, noting abuse, and blaming
the user  (in this case, other apps' developers) is pretty disingenious.

> Each undocumented key represents a bug in its respective application's
> schema, not in GConf itself. If Galeon didn't have a manual, would you
> blame Yelp?

I *do* blame the GNU project for deprecating manpages.  GNOME apparently
hews to the GNU standard.  Debian requires manpages, though lack is
considered a non-critical bug:

   http://twiki.iwethey.org/Main/InfoVsManpage

*Also* in the interminable l-e GNOME threads was an exceptionally heated
exchange in which Waugh and Martin Pool berated me for not finding the
gconf editor...and failing to admit for multiple rounds of discussion
that, in fact, there was no man page.  In fact there was little if any
documentation, which I demontrated by looking at several sources,
including the offial project website, etc.  Some of which is here:

    http://zgp.org/pipermail/linux-elitists/2003-October/007510.html

...you'll note a mutually low level of polity.

The start of the thread is here, and while brusque, isn't particularly
slanderous:

    http://zgp.org/pipermail/linux-elitists/2003-October/007505.html

My specific criticism of gconf was:

    It's specifically the lack of generic tools to access the Registry,
    the inability to comment it (or comment out keys while trying
    settings or dealing with a contingency), the masking of
    functionality through intentionally obfuscated keys and values, and
    the fragility of the system as a whole to any file damage, which are
    the primary complaints against the Registry.  Best I can tell, gconf
    is largely bug-for-bug compatible on all points.

None of those claims was refuted in the firestorm that followed, modulo
assertions that gconf is slighly less prone to file corruption issues.

 
> Research before you flame!

Please give me the benefit of the doubt:  the issues I've raise here
_are_ real, observed, and often verified by others or GNOME developers
themselves.
 


Peace.

-- 
Karsten M. Self <[email protected]>        http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
    You know, maybe FSF should just rebrand emacs as the hurd and stick
    a fork() in it...
    - Karsten M. Self, on linux-elitists

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFCCixTefG8443k044RAg+AAJ4sr8TiZWhyU4i4LnUswW5oWN4LcQCffZEW
PxBMtmuFww+VPF7wOsAIb1c=
=QldK
-----END PGP SIGNATURE-----