[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [olpc-software] AbiWord, HIG

On Thu, 2006-03-16 at 01:05 +0000, Mike Hearn wrote:
> On Wed, 2006-03-15 at 16:34 -0800, Alan Kay wrote:
> > It's worth pondering just why large portions of the computer community 
> > didn't build on the best designs (and still don't) ... This wouldn't be 
> > tolerated in any real science or engineering discipline ...
> Well, ignoring that 'best design' is subjective, the industry is still
> pretty young and has a short memory. The X11 protocol was designed at
> about the same time I was learning to talk :)

Yeah, I know people programming it who weren't born at the time.

> How old was JWZ when he started work on Netscape I wonder?

I have great sympathy for the NCSA folks: Mosaic was a success disaster.
X was a success disaster.  If you haven't lived through such and
experience, it is hard to explain. People get to learn as they go along,
and often/usually don't get the time to do the homework they should; in
our case (Bob Scheifler and I), we were at somewhat pretty familiar with
some of the significant previous work, and had access to some of it to
play with (unfortunately, not Alan's Kay's Smalltalk based systems, at
least more than a few superficial touches; we never got to really *use*
one), but Xerox Altos, Lisp machines, Sun and Apollo boxes were around.
And we were both about 10 years older than most of the NCSA crew.  So I,
for one, am reluctant to throw many stones at the NCSA crew.

There are things in X I hate that I know could/should/would have been
done much better had Bob Scheifler and I had a couple more years to work
on X before it took off.  We didn't have the luxury of time, and did the
best we knew how at the time.  Would that we had known Lyle Ramshaw at
the right instant, for example; our graphics would have been sane
decades earlier (some of the key problems in Cairo were solved by Lyle's
insight).  Our greatest fear was that X10 would become so widespread
we'd have to live with that: now that would have been a disaster.

And sometimes there are deliberate compromises made to make something
acceptable to the culture.  For example, believe it or not, some of the
very early X code (mostly client side stuff Bob wrote) was in CLU.  We
made the deliberate decision to use the vastly inferior C language, as
that was culturally acceptable and portable, and CLU was not.  This was
a decision made with eyes wide open.

When things like Mosaic, or X happen and take off at overwhelming speed
(Linux's takeoff is slow by comparison; just much larger at this
point ;-)), the people are overwhelmed by events, and do the best they
can.  Then you get to clean up your mess (or abandon it; I've done
both...).  Much of my last 4-5 years has been working on cleaning up the
mess, while not breaking people's existing code.

I'm now very proud of X, by far more so than at any previous time in its
history, with what people are doing/can do/will do with the new X
system, a vastly better system than 5 years ago.  

Thank all of you for your efforts...

					- Jim
Jim Gettys
One Laptop Per Child

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]