On 6/8/05 6:54 AM +0900, Joel Rees wrote on [OT] Re: FORTH
>> I was just wondering what the magic was that you saw in FORTH. My
>> understanding is that it is a very low level language.
Gee. I must have missed this one!
One upon a time, I worked with Forth quite a bit, even developing a few
implementations of my own. The one thing that's always struck me about
Forth is that it is much less a "language" than it is a technique for
programming. To me, that's always been a part of its "magic".
>Think of FORTH as LISP without parenthesis underneath everything,
Using Forth in what I'd call a "surface" level, the above might be
reasonable but the "magic" starts when you get beneath the hood.
Fortunately, this is not all that difficult in most Forth systems.
>FORTH needed a lot of work, and the current standard misses a lot of
>points, leaves you stuck with a reverse polish C and not-quite-unix
>libraries, and still no standard object format.
Statements like this are among the main reasons why more people don't use
Forth. They get scared off by the lack of all the dross computer languages
carry with them. I don't know what "points" the current standard misses
(FWIW: I rather objected to its creation in the first place - people doing
Forth work seldom pay it much mind anyway). I personally love being "stuck
with reverse Polish" - postfix notation actually makes operational sense,
unlike infix notation. But the truth is that if you must write arithmetic
expressions in infix notation, it is not a big problem to add the ability
to Forth. As for Forth libraries, there a lots of them - they're called
source code. Forth compiles almost as fast as its source can be read from a
hard disk on a well written implementation, anyway. There's very little
reason to have an "standard object format" in Forth as one does not link
modules to build programs as a rule (such things can be and have been done
but are hardly worth the bother).
>Java tries to do what FORTH could have done ...
I assume this is about transportable code. Actually Forth did this a VERY
long time ago. It's just that hardly anyone bothered to notice. For
example, I once wrote a version of Forth, dubbed Wump4TH, for the Apple ][.
It used what is known as token-code, something I did not invent, so it
precedes my use. Shortly after Mitch Bradley and I spent the night talking
about how this was implemented, he invented Open Firmware. The idea behind
OF is that a computer need only have the basic token interpreter in its ROM
to handle dealing with cards developed independently of the computer's CPU.
Things like OF might be a bit more well known if it weren't for the
plethora of PC BIOS vendors who aren't interested in that sort of
standardization - it'd put them out of business by obviating what they do
now.
>Whether FORTH could have answered the problems that you run into when
>you start trying to implement true context sensitive grammars...
According to Alan Turing and Chuck Moore, if you can't do it in Forth, you
can't do it. ;-)
--
Walter M. Pawley <[EMAIL PROTECTED]>
Wump Research & Company
676 River Bend Road, Roseburg, OR 97470
541-672-8975
