From mboxrd@z Thu Jan  1 00:00:00 1970
To: 9fans@cse.psu.edu
Subject: Re: [9fans] Re: Plan9 Programming languages ! -- The Future !
In-reply-to: Your message of "Thu, 13 Dec 2001 17:26:24 +0000."
             <9vane1$g9s$1@ctb-nnrp2.saix.net>
Message-ID: <27962.1008265896@apnic.net>
From: George Michaelson <ggm@apnic.net>
Date: Fri, 14 Dec 2001 03:51:36 +1000
Topicbox-Message-UUID: 35391914-eaca-11e9-9e20-41e7f4b1d025


> According to my opinion - The success, future etc etc. of any operating
> system (Plan9 in this case) depends on its ability to be programmed by all
> possible programmers (in more languages has possible etc) No Software !! -
> No Users etc !!! :-}

success is very subjective. the Apollo-11 OS was a fantastic success
given the engineering constraints but the UI was maybe a bit tortuous.

language plethora is a good thing in research space and usually a bad
thing in deployment. my experience is that it massively increases the
opportunity for bad behaviour. GC/VM models, IPC, program-system boundarie
turn out not to be completely abstract but reflect the language of
implementation (is that unfair?)

My brother is a lamda calculus expert. I would find it very hard to
survive in his domain, but I also suspect good programmers would achieve
very very good things if that was the 'one true way'

The Pascal discussion reminded me of my first feelings on the apparent
convergeance of the algol-60 and fortran i/o binding mechanisms. How wierd!
two languages so different yet one apparent method to say how to connect
to a punch or a reader. Then to get Pascal, and have to come to terms with
its bindings..

I suspect any language with i/o or IPC a first-class concept in the language
and not buried under methods or procedure call interfaces would blow my
lobes. rendesvous between asynchronous real-world events, or the emulation
of parallelism in a serial instruction machine and how that exposes to
the apparently discrete programs is a wonderful opportunity for language
designers to play the deity and show their view of 'how it is done wrong'

I hope P9 remains a small-set-of-ported-languages space. I think it will be
better for it. One from the lisp family, one from the interpreter space,
one from the C gang, one for old timers. one shell/rc to bind them. one
namespace scheme to find them.

cheers
	-George