Re: Let's do some coding :-)

[Jonathan S. Shapiro]

On Tue, 2005-10-25 at 13:58 +0200, Alfred M. Szmidt wrote:
>    This is not for you alone to decide,
> 
> And it sure as hell isn't for you to decide either.  Infact, I have
> more to decide about it than you; you're not even subscribed to the
> list that discusses the design of the GNU system.
> 
>    In other words, keep your side sweeps at home, they are uncalled
>    for and we don't take them from you anyway.
> 
> Only if Jonathan keeps his `side sweeps' at home.
> 
>    If you have troubles understanding the design, or see some problems
>    implementing it, then you are welcome to ask technical questions
>    about it.
> 
> Maybe I understand the design, and dislike Jonathan argogance of
> trying to impose `academia' on the GNU system.

I do not "impose `academia'". I cannot. I have no ability to control
where the Hurd project goes, except by asking questions. If the answers
to these questions are uncomfortable, this may mean that the questions
are bad, or it may mean that the Hurd goals need to be re-evaluated.
This is for the Hurd project and the GNU project to discuss and decide.

What I *do* know is that the right -- and the obligation -- to ask
questions is necessary if "free as in freedom" is to have any meaning at
all. These questions should be polite, and they should be asked
responsibly; not just for the sake of disruption.

However, I also think that I have a well-earned right to my opinions
about POSIX. POSIX was built, in part, on my work (and the work of many
others). POSIX was not an innovation! It was an attempt to codify a
system that had already been in use for almost 20 years at that time:
UNIX. The UNIX system was a very good system, and it still is. It has
been amazingly successful, and supporting the large number of
applications that run on it -- and supporting them well -- is a very
important thing to do.

But here are my opinions:

I believe that POSIX is not the way of the future. Furthermore, my
conversations with Ken Thompson, Dennis Ritchie, Brian Kernighan, Rob
Pike, Dave Korn, and Dave Presotto clearly indicate that they agree.
These people did not stop with UNIX; they went on to try to build better
systems and/or improve on the system that existed. Ultimately, Pike
concluded that the UNIX API was not the best way forward. You may have
noticed that Ken Thompson is one of the authors on the Plan 9 papers.

I believe that if the goal of the GNU effort is merely to rebuild POSIX
then it is doomed and it is a waste. That goal made sense in 1984, when
it was expected that it would be complete by 1990. It does not make
sense today. Given the overwhelming success of Linux, and the incredible
amount of effort that has gone into and continues to go into that
system, I would say that the GNU system is an unquestionable success.
You (and I) may not think it is perfect, but there definitely does exist
a free POSIX system with comprehensive support and a very widespread
user base.

shap