Re: [Funge] question + a new weird thing..

[vsync <vs...@bi...>]

"Joshua J. Belsky" <jjb...@uc...> writes:

> You've kept us in the dark about this Temporal Funge thing. Are you going to

Yeah, I apologize for that.  I've been meaning to get to it for quite
some time now, but school and work kept getting in the way.  Darn
useful stuff!  =)

> store a record of the funge space at every tick, so that time travel will be
> possible within the program?

That's what I'm looking at at this point...  I plan to be able to send 
values back in time, with something analagous to the "p" command, and
also spawn new threads in the past...  I'm planning on implementing
Befunge-93, with enough of the Funge-98 threading stuff to be
consistent.  I might add more later.

Of course, this is going to hog memory something fierce, in theory,
but since Funge-space generally isn't that big, it probably won't be
too bad.  Later I can play with optimization.

Actually, I'm starting work on my prototype right this minute...  I
don't know how far I'll get, but I should be able to get a basic
Befunge interpreter done quickly tonight and go from there.

> Here's what I've been thinking about:
> I'm trying to cut down the Funge language even more, perhaps limiting the
> stack to three levels and other such things, so that I can build an embedded
> funge interpreter. I'm thinking something along the line of the BASICStamp.
> (if you don't know what a BASICStamp is, look it up; it will change your
> life.)

Three levels?  That's a bit small...  I also had the neat thought of
porting Befunge to my HP calculator(s).

> I'm also interested in producing a new standard of the language from the
> Befunge-93 specification. There are some elements of Funge-98 that I'd like
> to implement (file I/O), some I don't really like (the stack-stack), and
> some new ideas of my own (floating point operations.). Now, of course,
> floating point operations are unnecessary, as demonstrated in the mandelbrot
> generation program, but I'd like to work out some compromise half-way.
> 
> Oh, and along the lines of subroutines, Gregory, I have some thoughts. While
> I was sitting in a class today, I worked out a relatively nice way to
> implement subroutines using only the original Befunge-93. It involves self
> modifying code, and seems to be the most character-efficient way to do it.
> I'll write a demonstration program and post it to the list in the near
> future.

Self-modifying code would be the best way, of course...  Just curious:
Would you be able to do recursion at all?

> Speaking of which, someone (vsync?) should set up a Funge source archive to
> replace the old Cat's Eye one. In the index file, it would be good to have
> the version of the interpreter it is intended for, along with other
> interpreters that it works or has been tested with. That way, when we
> develop new stuff, we can share it with the world.

Take a look at:

  http://vsync.cjb.net/funge/downloads/bef93src/

These are only Befunge-93, since anything later could only be gotten
to by means of the CGI scripts which don't work anymore.  I wouldn't
mind setting up an archive, but I'd need some help on that...
Basically, if people could do a quick version/compatibility/etc check
on it, I could archive it a lot more easily...

I'll try to archive anything that appears on this list, too.

-- 
vsync
Beware.