hehehe

[Tom Novelli]

On Mon, Feb 15, 1999 at 01:12:35PM -0300, Jecel Assumpcao Jr wrote:
> While Logo is more friendly than other Lisps by having a syntax
> that greatly reduces the parenthesis (but is a bit confusing for
> my taste) and nicer error messages (due to, among other things,
> a separation of functions that return values from those that don't),
> it lacks these features:
> 
>   - lambda
>   - closures
>   - continuations
>   - macros

Wow! I never thought of that.. I was looking for a LISP without so many
parenthesis, and it turns out to be the cheesy language I learned in grade
school!

> My point? As Jerry Pournell loves to say "perfect is the enemy of
> good enough". Even a Tunes written in Logo would be way better than
> no Tunes at all. Of course, Scheme is a much better systems programming
> language than Logo (and ML may be even better for Tunes from what
> I have heard, but it is one of the few languages I don't know). The
> open source development style doesn't work unless there is running
> code to be dicussed and improved. If there was something running
> (in Scheme, for example), I'll bet Rice's arrow system could be
> implemented with little effort. Nothing explains ideas as well as
> a working system.

ML? I spent a lot of time designing my ideal language.. but I still hadn't
looked at ML. I had a good feeling about it though. This morning I took a
look at it, and you know what? It's exactly what I had in mind! Well, not
exactly, but it's a great starting point. It's simple and easy to
read. Forget Forth, forget LISP.. I'll write a small ML-like
compiler/interpreter for my prototype OS.

> As for low level stuff, I don't think it makes sense to work on
> that anymore with Utah's OSkit finally out.

There's nothing in OSkit that wasn't already in Linux or BSD. Sure, it's
helpful when I need example code... But personally, I refuse to use
something that occupies so much space and requires a Unix system and a C
compiler to build it. I'm working on a *new* OS, from the ground up. 
Assembly isn't that hard, really.. another week's work (when I have a week
to spare) and I'll be on to high-level stuff. Another concern with OSkit:
I don't think it's very well suited to what we're doing.. it's too Unixy..
of course, if someone else wants to use it, be my guest. Who am I to
dictate?

Tom Novelli