Wed, 30 Dec 1998 00:45:56 -0800 (PST)
On Mon, 28 Dec 1998, Matt Miller wrote:
> Ok. I have read dem's specification of the system. I have read Brian's
> Arrow stuff [although I am still trying to parse some of it.]
Same here. I like the basic idea of representing a program as a bunch of
objects connected by arrows. It's easy to see parallelism that way, and
it's more intuitive. Of course, other syntaxes (LISP, RPN, algebraic
notation) are better in some circumstances, so let's support them all.
> Now that you have stated the goals, where does work commence?
> At what level do you want to start building?
> Prototyping something on top of a current system? Low level kernel work?
> I know all of these have "started", but nothing seems to be moving...
> where can i apply my time to have the most impact?
I don't think the distinction between "standalone" and "on top of
Linux/DOS/etc" really matters. We should have both.
One thing we need is a system for managing objects in memory and on disk.
I'm just starting a prototype of a distributed persistent object system.
I'm using a Forth-inspired interpreter (with its core written in assembly)
running under Linux, DOS, and by itself. The idea is to use machine code
where it really counts, while keeping it to a minimum to ease porting to
There's plenty of work to do.. designing new user interfaces and
programming languages, rewriting documentation in plain language,
networking and distributed computing... I say, pick something you're
interested in and go with it.