ok, in my experience I am far more of an os coder than a lisp
compiler/interpretter coder, so the lisp based stuff is going slow.
my newest source contains a basic interpretter (which I have been working
on/extending/fixing since I had posted it...).
(I have since redesigned calling/functions a bit, and had discovered that
the interpretter was dynamically scoped...).
the code is mostly in c, as that is what I am most fammiliar with. spread
around on my site is some info that tries to explain what I am trying to do.
ok, the interpretter is a little "schemish", but really that looks like it
would be easier to implement right off.
ok, I lack any real specific problems right now, but mostly I am just
interested if anyone has comments...
it can be found at the site at the bottom of the post.
I don't really mean to top post, but I haven't really seen an appropriate
thread to place this in.
--
<·······@hotmail.com> <http://bgb1.hypermart.net/>
possibly worth mentioning (though I forgot in last post). I do have a basic
kernel, and my lisp stuff sits on top of that. the images floating around
in the source should be usable (though getting it to recompile on a
computer other than mine might take a little fiddling...).
I just wanted to make sure people don't think that *all* I have is an
interpreter...
once lisp stuff generally exists and stabilizes then I will try to
integrate it with other parts of the kernel. then maybe I can go and start
working a little more on the rest of the kernel...
I am thinking: will lexical scope require binding/resolution to be done at
parse time; or at "define" time (possibly by making and alternate copy of
functions); or is there some better way to do it.
my interpreter currently just parses the s-expressions (which is good
enough for my early practice...). the variables are resolved at
interpretation time, and the scope is generated by calls. possibly for a
lexical scope a function will have to point to its own bindings list.
sorry I don't know much.
--
<·······@hotmail.com> <http://bgb1.hypermart.net/>
Hi,
I don't know if you have ever tried to search for LispOS on google but
if you do try you'll find a mail archive maybe. I think you would
appreciate to read some of the interesting posts...
Best regards,
Paulo
In article <··············@corp.supernews.com>, "cr88192"
<·······@hotmail.com> wrote:
> possibly worth mentioning (though I forgot in last post). I do have a
> basic kernel, and my lisp stuff sits on top of that. the images floating
> around in the source should be usable (though getting it to recompile on
> a computer other than mine might take a little fiddling...). I just
> wanted to make sure people don't think that *all* I have is an
> interpreter...
>
> once lisp stuff generally exists and stabilizes then I will try to
> integrate it with other parts of the kernel. then maybe I can go and
> start working a little more on the rest of the kernel...
>
> I am thinking: will lexical scope require binding/resolution to be done
> at parse time; or at "define" time (possibly by making and alternate
> copy of functions); or is there some better way to do it. my interpreter
> currently just parses the s-expressions (which is good enough for my
> early practice...). the variables are resolved at interpretation time,
> and the scope is generated by calls. possibly for a lexical scope a
> function will have to point to its own bindings list.
>
> sorry I don't know much.
>
--
Paulo J. Matos : pocm(_at_)rnl.ist.utl.pt
Instituto Superior Tecnico - Lisbon
Software & Computer Engineering - A.I.
- > http://www.rnl.ist.utl.pt/~pocm
---
Yes, God had a deadline...
So, He wrote it all in Lisp!
> Hi,
> I don't know if you have ever tried to search for LispOS on google but
> if you do try you'll find a mail archive maybe. I think you would
> appreciate to read some of the interesting posts...
>
worthwhile tip...
since my post I have made things lexically scoped and started tying things
together with other parts of the os...
have been imagining something similar to packages/sets for organization
purposes, I have started to design things now. the concern: have them
referred to by name, or treat them as objects and possibly allow predicates
to select them. the second would be more difficult I would think for now,
thus I will go with the first.
I was thinking of organizing the packages in a tree (much like a
conventional fs, except that other structures could be built on).
a shell would bind certain packages to certain locations in the local space
for convienience.
ie: "temp-user", "static-user", "temp-system", and "static-system", maybe
"system-root".
each package could contain other packages and so forth...
or maybe they could be more like hash tables. it might be useful just to
define a new system though.
for now there are just variables...
I am still planning the tag based access control stuff (which will probably
have to be a property of variables...).
--
<·······@hotmail.com> <http://bgb1.hypermart.net/>