ideals and practice (Re: View Source question)

Yoshiki Ohshima yoshiki at
Mon May 19 22:00:28 EDT 2008


  I put Reply-To: header (I don't know if it goes through).  Please
send the reply to, if it continues.  For the
middle part of this email, you are just paraphrasing what I wrote, so
I see not much difference.

> This is a great thread. It is bringing out a lot of useful information
> from many points of view. You ask good questions, Yoshiki, which is a
> more important skill than giving good answers. It's a pleasure to work
> with you.

  Thanks! (this wasn't sarcasm, right?^^;)

> >  I'm not sure if Python has that edge over Squeak, but probably it
> > does.
> From which point of view? A significant fraction of the world's
> software infrastructure is now implemented in Python, and Python is
> clearly a more appropriate language for the novice than BASIC, C,
> Pascal, Java, or the other languages usually favored in education and
> commerce.

  I have been always agreeing with that (Python has been used widely
and have a big mind share).  Python is the practical choice, even
though the creator doesn't seem to like the project (Open Source is

  If "access to most of the rest of the GNU/Linux world" means to
access the mind share, documentation, and etc., it is simply true.
But, if it was about viewing the source of the rest of the GNU/Linux
world in the environment, Squeak can just do the same as any other
program written in any other programming language.

  (BTW, I never suggested that Squeak/Smalltalk should have been

> The issue is not technical superiority on some measure.

  It isn't.

  My point was that people who know about computer systems would like
to show a technically better system to kids, if it is somewhat used as
a reference.  (Not that I think most of kids are going to see the
code, nor I claim there is another practical platform exists.)

  Some may say "hey, if you don't know an alternative, just accept the
status quo."  But this project has been going for a while, and the
longer times goes the more interesting the question "if we had started
from something different, where would have we been" become.

> > And, it wasn't conceivable to actually do in that short
> > time, an OS-less system would have made more sense for the target
> > system; as you know, Ivan was thinking the possibility of "full Python
> > machine".)
> Eh, LISP machine, Smalltalk machine, FORTH machine, APL machine. All
> great technical innovations, all market flops, if they were even
> implemented fully. Computers should be general-purpose.

  Currently, dominant architectures are essentially C and Fortran
machines.  And as Jecel wrote, market flop and technical superiority
are just very different things.  My boss sums it up by saying
"evolution doesn't optimize", meaning that the natural selection
process only makes an "ok" situation (unless it is on the dying
branch), but doesn't reach the best possibility.

> Now if somebody wants to design and implement a Parrot machine that
> will support dozens of languages, we can talk. But Parrot isn't
> ready.

  Yeah, timing is always bad.  VPRI's new project is not ready at all.

> >  If we are trying use the OLPC XO as the trojan horse of
> > disseminating a better idea of computer including operating system, it
> > is unfortunate that we needed to use Linux.  It is the most practical
> > system to use in the short term, but basically we are using it because
> > it is the de-facto standard.  (And people are rather thinking it
> > better because it is not Windows.  Strange.  Without real education
> > content, neither is good enough.)
> Linux is, of course, way better than Windows. Not least in supporting
> better education software and content, because we can integrate
> security, collaboration, and fantastically low power consumption into
> the kernel.

  It is probably true for security.  I don't know if it is true for
collaboration and low power consumption.  Especially for
collaboration, what would prevent Windows and apps on it to be

-- Yoshiki

  (My flight was delayed 90 minutes, but I finally arrived Kendall

More information about the Devel mailing list