Hello everyone..
David Jeske
jeske@home.chat.net
Thu, 10 Dec 1998 19:06:53 -0800
On Thu, Dec 10, 1998 at 09:19:05AM -0500, Tom Novelli wrote:
> It's about time I introduced myself. I've been checking out your project
> for a few months, and I like what I see, so I decided to join.
Welcome aboard.
> I know Tunes isn't a functioning OS yet, and these things like FLUX,
> OS Toolkit, SOLO, don't really seem to exist.
> How about a nice practical approach to get Tunes going:
>
> Start with a simple Forth kernel, because it's easy to implement with
> assembly language. Add DOS filesystem support. Possibly write an editor
> and an assembler in Forth. Write a Scheme interpreter in Forth &
> Assembly.
If we agree with the presumption to base work on Forth, why wouldn't
we just write our stuff under Forth for Linux or any other existing
OS?
> Once we have a basic operating system and a few tools, we can start
> experimenting with new languages, object/module systems, user
> interfaces, networking, and so on. Some ideas that sounded good in
> theory may not work so well, and new possibilities may present
> themselves. After a while (assuming things work out) it'll all pull
> together.
We can already do all these things with existing systems.
One of the major goals of tunes (IMO) is to be able to figure out a
programming model where we can programs things like you mention _once_
and truly be able to reuse them wherever we need to, and to be able to
abstract out things like the 'symbolic language you use to enter a
program'.
Writing another implementation of network or filesystems dosn't help
us. I'd rather just code up some four line matrix math computation and
figure out a way we can input it so we can connect it to any language,
let the user view it in any language, and input any 'matrix' form we
want into it.
> I should mention a few ideas I had:
> A new language with the simplicity of Forth and LISP, preferably with
> support for a variety of syntaxes, with a really simple module/object
> system.
There are certainly good properties in both Forth and LISP. Can
someone refresh us or point us to a description of Forth's strengths
and limitations for tunes?
> A "smarter" GUI that's quick to use, without a mouse. Also, windows
> should be avoided except when they're actually useful. I fail to see the
> point in using 3/4 of the screen instead of the whole thing ;)
You _always_ use the whole screen. You just might not use it very
well. I'm currently typing in a terminal which is only taking up about
1/5 of the screen. It is exactly the right size. If I made it bigger,
either I would have too many rows or columns, or the text would be too
big. The rest of the screen is conveying information to me about
'where' I am (i.e. what virtual desktop I'm on) and is free for when I
open a web browser to view a URL someone emailed me.
> Since Tunes will be modular, have a web form where you say what kind of
> hardware you have, and it sends you a custom boot image containing only
> what you need. I hate downloading 50 megs of Linux. (This is a long-term
> thing, when there are hundreds of drivers and crap)
This sounds like a Linux oriented project that has nothing to do with
Tunes. It is true that we don't have a Tunes implementation currently,
but when we do, I guarantee that the distribution model for installs
and boot images will be radically different from Linux.
--
David Jeske (N9LCA) + http://www.chat.net/~jeske/ + jeske@chat.net