Forum unknown
· SoftwareDevelopment
MacWorks for the Amiga
15 messages in this thread
Yes, I've read the Dispossessed. Definately a great work. By the way,
are we likely to see a version of BFORTH for Amy (trying to get back on the
track of software…). I'm translating my LISP interpreter from PASCAL to C,
and at the same time porting it to Amy. No easy task, especially since I don't
have an Amiga of my own yet (I hope none of Rand's followers will now condem me
for using an Amiga for free that doesn't belong to be.). Anyway, as soon as
its ready you'll see it here.
Hazy, I'll probably do BFORTH a some point for Amy, but not real certain that
Forth is really suited for it. I like forth more on smaller systems,
personally. I've been fiddling with writing a smalltalk for the thing, but
it's slow going (trying to learn the system, figure out how tto do smalltalk,
etc.) There are other free sorts of things I have in mind also. I would love
a good Lisp – are you going to any trouble to make Common -able Lisp? That
might be nice, if it's not to much work. Best, SDB
CthiJust what I was thinking the other day: "The amiga could make a great
Smalltalk system if it could somehow be gotten onto the amiga intact." I
wonder at the speed though, I am porting Icon to the Amiga and it is no great
speeddemon, although a lot of that seems to be Lattice C. Now that Icon is
almost done, I am taking the C code and transferring it back to the SUN
workstation, compiling it into assembler there and running it through a 68000
assembler optimizer and then transferring it back to Amy to be assembled.
Seems to make a noticeable difference in speed. I assume you have the
smalltalk book? Gary
Gary, Yes, I have the smalltalk book. The speed is indeed going to be a
problem, but I usually do most of this kind of thing in assembler, (grinding
it out by hand…since I don't own a Sun – and won't be getting one anytime
soon 90k bucks – Phew! you're a lucky guy. I knew you were doing Icon, and
I've been on the lookout for it – are you going to make the source available
to – or is it too big for here or there? Best SDB
Source will probably be available "on request", source now is hovering at
around 750K of code, really too much to upload. Even the executable is going
to be about 200K+, I have been talking with john foust about turning it over
to Amicus and letting them distribute it on their disks. Smalltalk would
really be wonderful! I've wanted it since they had the article in Byte
magazine years ago.
Even playing with the object oriented programming in XLISP is neat. My
biggest problems with porting Icon have been memory management, i.e. the
amiga doesn't provide you with contiguous data areas for Icon's workspace as
it would on any computer with an MMU, and the fact that Lattice C cannot do
call-by-value with structures,unions etc.
The manual says they can, but it is a very strange way of doing it, they
don't actually push the structure word for word onto the stack as other C's
do, they push a pointer to the structure and the compiler generates code
inside your function to take the pointer and make a local copy. Dumb!
The entire Icon interpreter is based on passing structures called descriptors
around to do all its work. Snobol would have similar problems too, maybe
smalltalk too but I don't know the internals, but a lot of the so-called big
languages like these at least assume a contiguous workspace which you can't
get on the Amiga. From my battles with Lattice's C compiler and MetacomCo's
excuse for an assembler and linker you'll probably be in for a very long
grind, but don't despair! I want smalltalk! <grin> Bye – Gary
Source will probably be available "on request", source now is hovering at
around 750K of code, really too much to upload. Even the executable is going
to be about 200K+, I have been talking with john foust about turning it over
to Amicus and letting them distribute it on their disks. Smalltalk would
really be wonderful! I've wanted it since they had the article in Byte
magazine years ago.
Even playing with the object oriented programming in XLISP is neat. My
biggest problems with porting Icon have been memory management, i.e. the
amiga doesn't provide you with contiguous data areas for Icon's workspace as
it would on any computer with an MMU, and the fact that Lattice C cannot do
call-by-value with structures,unions etc.
The manual says they can, but it is a very strange way of doing it, they
don't actually push the structure word for word onto the stack as other C's
do, they push a pointer to the structure and the compiler generates code
inside your function to take the pointer and make a local copy. Dumb!
The entire Icon interpreter is based on passing structures called descriptors
around to do all its work. Snobol would have similar problems too, maybe
smalltalk too but I don't know the internals, but a lot of the so-called big
languages like these at least assume a contiguous workspace which you can't
get on the Amiga. From my battles with Lattice's C compiler and MetacomCo's
excuse for an assembler and linker you'll probably be in for a very long
grind, but don't despair! I want smalltalk! <grin> Bye – Gary
Gary, Yes, I have the smalltalk book. The speed is indeed going to be a
problem, but I usually do most of this kind of thing in assembler, (grinding
it out by hand…since I don't own a Sun – and won't be getting one anytime
soon 90k bucks – Phew! you're a lucky guy. I knew you were doing Icon, and
I've been on the lookout for it – are you going to make the source available
to – or is it too big for here or there? Best SDB
CthiJust what I was thinking the other day: "The amiga could make a great
Smalltalk system if it could somehow be gotten onto the amiga intact." I
wonder at the speed though, I am porting Icon to the Amiga and it is no great
speeddemon, although a lot of that seems to be Lattice C. Now that Icon is
almost done, I am taking the C code and transferring it back to the SUN
workstation, compiling it into assembler there and running it through a 68000
assembler optimizer and then transferring it back to Amy to be assembled.
Seems to make a noticeable difference in speed. I assume you have the
smalltalk book? Gary
Scott-
My LISP isn't anywhere Common at the moment; then again, its not that close
to running on Amy yet either. I imagine I'll try to get some or muc8e~
Common compatibility, as that's the standard of today (I haven't worked in
LISP since '83). Maybe I'll go for Golden Common Lisp. I don't know much
about the standard, but Guy Steele Jr., the GCL Guy, is an old prof. Anyway,
it may be some time in coming, but I'll definately do it.
Hazy, There is a book out (I think by Steele) that gives the dope on the
standard. I think Golden Common Lisp is a trademark (??) not sure. Have you
seen it? Aside from being very solid, it's also very witty. I have two copies
here….. Best SDB
Scott-
Steele did the GCL book after I left school, and as yet I don't have a
copy, but this is a good an excuse as any to get one. I've also heard that an
addition of the famous Patrick Henry Winston/Berthold Klaus Paul Horn "LISP"
book is out, in Common LISP flavor, but that may not really describe the
standard. In any case, I've got about half of the original interpreter
compiled on Amy, and I've already added several things that should speed it up.
I'll really have to check on GCL, though, to see if the implementation details
(i.e. where do you hang your functions, atom values, etc.) is part of the
standard.
-Hazy Davy
Scott-
Steele did the GCL book after I left school, and as yet I don't have a
copy, but this is a good an excuse as any to get one. I've also heard that an
addition of the famous Patrick Henry Winston/Berthold Klaus Paul Horn "LISP"
book is out, in Common LISP flavor, but that may not really describe the
standard. In any case, I've got about half of the original interpreter
compiled on Amy, and I've already added several things that should speed it up.
I'll really have to check on GCL, though, to see if the implementation details
(i.e. where do you hang your functions, atom values, etc.) is part of the
standard.
-Hazy Davy
Hazy, There is a book out (I think by Steele) that gives the dope on the
standard. I think Golden Common Lisp is a trademark (??) not sure. Have you
seen it? Aside from being very solid, it's also very witty. I have two copies
here….. Best SDB
Scott-
My LISP isn't anywhere Common at the moment; then again, its not that close
to running on Amy yet either. I imagine I'll try to get some or muc8e~
Common compatibility, as that's the standard of today (I haven't worked in
LISP since '83). Maybe I'll go for Golden Common Lisp. I don't know much
about the standard, but Guy Steele Jr., the GCL Guy, is an old prof. Anyway,
it may be some time in coming, but I'll definately do it.
Hazy, I'll probably do BFORTH a some point for Amy, but not real certain that
Forth is really suited for it. I like forth more on smaller systems,
personally. I've been fiddling with writing a smalltalk for the thing, but
it's slow going (trying to learn the system, figure out how tto do smalltalk,
etc.) There are other free sorts of things I have in mind also. I would love
a good Lisp – are you going to any trouble to make Common -able Lisp? That
might be nice, if it's not to much work. Best, SDB