OFB Community Mailing Lists

[Karl Kleinpaste]

Timothy Butler <tbutler at ofb.biz> writes:
> Here's a wild one: what if we dumped both legacy DEs? I think
> it wouldn't be too bad of idea to build an entirely new DE on top of
> GNUStep and Objective C.

No.

No.

No.

Absolutely not.  Stop that.

Fragmentation killed UNIX.  Real UNIX, that is -- SysV, SunOS/Solaris,
HP-UX, AIX, OSx (remember Pyramid's dual universe?), BSDI, Xenix, ...

BSD became an independently interesting "distribution" because AT&T had
such a severe case of NIH that none of the good stuff of BSD could be
expected ever to get grafted back into SysIII, er, 'scuse me, SysV.  And
that's even *without* considering the effects of the System Purity Hit
Squad of Bell Labs' Area 11.

All that's left of all those is the somewhat marginalized BSD, itself
fragmented into at least 3 subvarieties, and Solaris.  There is so
little HP-UX and AIX left that they're noise in the graph.  Xenix is
barely a memory.  SysV ... who owns that, and what is it?

Everyone and his brother thought that he *really* understood What UNIX
Should Have Been, and darn it everyone and his brother was going to
(Re-)Do It Right and then the whole world would see the Great Shining
Light Of Their Genius so they could Rule The World.

[...crickets chirp into the still night air...]

Or maybe not.

Come on.  GET OVER IT.  Get over this kind of hubris.

Along came Linux and the wonder of a totally (ok, mostly) free system
when that kernel was combined with a free libc and a mountain of GNU
tools.  Yea verily, the Age of Aquarius of Software Freedom was upon us,
and its Kingdom had been established.  SourceForge was to be our Savior,
SlashDot our Prophet, FreshMeat our Comforter.

[...crickets chirp into the still night air...]

Or maybe not.

Because then we found that users needed whole, complete, easy-to-install
distributions.  And verily, along came InfoMagic (went under without a
ripple, but I used it -- 1994), Slackware, Debian, RedHat, Mandrake
(*burp*) oh, excuse me, make that "Mandriva", SuSE, ...

How many Linux distributions are there now?  Hm, let's see, DistroWatch
claims a couple hundred.  And they've got moronic names like CrunchBang
and Xubuntu, the latter being "needed" because of course whatever it was
that Ubuntu provided was just not quite good enough, so rather than
re-engineer the tiny little part of Ubuntu that wasn't adequate, they
forked off a whole new distribution.  Riiiiiiiiiight.  May I regale you
sometime with the nightmares involved in trying to keep Xiphos (née
GnomeSword) build-able in both Debian and Ubuntu?

The only thing saving Linux from death-by-fragmentation is momentum.
Nothing else.  Every time something forks, its momentum gets divided
between the two resulting pieces.  Linux has so much momentum that even
forking its distributions a hundred times hasn't (yet) killed it.

We learn from history that we do not learn from history.

Now we have the claim that The Desktop Is Not Good Enough.  And so what
is the solution to that?  Is it to find what is yet inadequate about
GNOME or KDE, find new solutions within them, re-implement those
portions we have analyzed, and continue their lineage?

No.  The suggestion at the top of this message is to announce that now
we *really* understand What The Desktop Should Be, and darn it we will
(Re-)Do It Right (*this* time, honest!) and then the whole world will
see the Great Shining Light Of Our Genius so we can Rule The World.

[...crickets chirp into the still night air...]

Or maybe not.

GNOME is a fine desktop environment.  A suggestion that it *and* KDE be
"replaced" (ha, phooey...like that's going to happen before the heat
death of the universe) because we can now do a "better" job using -- and
this is truly priceless -- minority players in the areas of *both* the
language infrastructure (ObjC) *and* the toolkit infrastructure
(GNUStep) is beyond absurd.

I'll say it again: Come on.  Get over this hubris.

Fragmentation kills, unless it is fragmentation of the sort that takes
surgery to cut off a gangrenous limb.  That is not the case here.

"An entirely new DE" is not a desirable goal at this stage.  If Linux is
to gain further foothold on the desktop, the idea of scrapping
everything that has gone before because you think you can replace it
with something new&improved is absurd on its face.  The army has been
winning some turf wars, winning a few souls away from the Dark Side,
winning some market share.  So the solution now is to disband the army,
and raise a new one?  What does the word "momentum" mean, anyway?

--karl