Förster vom Silberwald wrote:
> Surely, one should have
> a better knowledge of Forth than what I call my knowledge of Forth.
But
> have you ever seen numerical algorithms implemented in standard Forth
> (study the Forth scientific library)? I mean what are all the
chemical
> substances which have wrecked havoc in all the brains of Forth users?
> Nobody on earth can tell me that a human beeing voluntarily choses
> Forth for writing numerical code. Although, the bad side: such humans
> exist - really.
I strongly suggest that you stay away from Hewlett-Packard RPN
calculators. You know, the kind that are very popular with engineers
and scientists. You wouldn't like them at all.
Btw, carefully read Professor Noble's comments about a FORmula
TRANslator extension for Forth. It addresses your concerns rather
well, in my opinion.
Also, there is a lot more to writing programs than crunching scientific
formulas.
> The aformentioned does not weight heavy compared to one fact:
Forthers
> really believe that they are in the upper 10 per cent out of a
> population of programmers. That is not a joke: just google in
> comp.lang.forth.
Please. That is a gross generalization that is simply not supported.
> Sorry for that impudent rant; but I have often wondered when reading
> clf what the hell is making them to believe what they actually are
> believing. I am no psychologist and I do not have any answers and
> solutions.
I'm curious. What kind of reaction to your post were you expecting?
One might get suspicious that you are a troll.
Doug Hoffman Mar 4, 2:27 pm wrote:
==
I'm curious. What kind of reaction to your post were you expecting?
One might get suspicious that you are a troll.
==
Hello:
Oh yes, I am troll. Do you feel better know?
When I was replying there were no indication that the reply will also
march through comp.lang.forth. I post stuff from Google and they have
changed some options in the past few weeks.
However, people may use whatever they want in their life: being it
called assembler or C or Fortran or Lisp or Scheme, etc.
But I will always open my mouth and bark: Forth is really not for me
either do I think that a Forth programmer is any better than the
average population of programmers. Forth programmers share typically a
very good knowledge of that hard wired machine instructions.
But the latter means nothing.
I use Scheme nearly every day. However, it would be far fetched from me
to think that a Scheme or CommonLisp programmer is above the average
programmer.
I am really not after speed but using Forth in scietific computing is
pure lunacy. I sometimes get the feeling that some Forther have never
seen anything else than Forth. Believe me they will miss a lot!
PS: I haven't written in my last (call it troll post if you like)
posting that rpn methods come into my way. There are so many things in
Scheme or CommonLisp which Forth will never! deliver me.
That said (I have no clue whether the post will also appear in
comp.lang.forth; I post from comp.lang.lisp),
Förster vom Silberwald
On 7 Mar 2005 02:59:27 -0800, F�rster vom Silberwald
<··········@hotmail.com> wrote:
>
> PS: I haven't written in my last (call it troll post if you like)
> posting that rpn methods come into my way. There are so many things in
> Scheme or CommonLisp which Forth will never! deliver me.
>
> That said (I have no clue whether the post will also appear in
> comp.lang.forth; I post from comp.lang.lisp),
> F�rster vom Silberwald
>
Don't worry. Your message actually go me interested in forth and I have
spent the
last 4 days beginning to learn the basics.
As a old C programmer I like the idea of a powerful language
*close to the metal*. It is taking me a lot of time and trial and error to
wrap my head around postfix notation and stack operations but then
I rather expected that it would. The unforgiving do ... loop and friends
has
made me pull my hair and scream more than a few times..
But I'm smart enough. With time and effort Forth will flow swiftly
from my fingers in a couple of weeks.
What attracts me is the intrinsic simplicity of the language.
Once you have the basics down it should hold few surprises
(unlike, say, C++) and I suspect that this is one of the reasons
for the alleged productivity. Well, I'll try and see.
(Also posting from comp.lang.lisp)
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
F�rster vom Silberwald wrote:
> When I was replying there were no indication that the reply will also
> march through comp.lang.forth. I post stuff from Google and they have
> changed some options in the past few weeks.
Google groups sucks, IMHO. Have you had a look at
http://news.individual.net/? They charge �10/year (they used to be
free...), but at least I enjoy the convenience of having a real NNTP
connection for three cents per day.
> But I will always open my mouth and bark: Forth is really not for me
> either do I think that a Forth programmer is any better than the
> average population of programmers. Forth programmers share typically a
> very good knowledge of that hard wired machine instructions.
You mean of Forth instructions that use the stacks? I think that the
main difference between Forth and Lisp is memory layout. Forthers (have
to) think about it, but that makes programs very efficient and also
reveals a nice program structure in the end. In Lisp you always have GC
and more abstract structures (named parameter lists being just one), so
you _can_ program more sloppily.
I'm more on the Lisp side myself I guess, but Forth is interesting and
very nice, especially for low-level (system, device drivers) stuff.
> I use Scheme nearly every day. However, it would be far fetched from me
> to think that a Scheme or CommonLisp programmer is above the average
> programmer.
Not? I think luckily most morons are scared away by the parentheses,
just like they are scared away by words like : over does> or ,
> I am really not after speed but using Forth in scietific computing is
> pure lunacy. I sometimes get the feeling that some Forther have never
> seen anything else than Forth. Believe me they will miss a lot!
I believe that most Forthers have some knowledge of assembly and
probably C. Some might have tried the Lisps or Java.
If you can't accept that Forth is nice in some situations (say, 64k
RAM), then just stop posting :)
You are free to use whatever language you like.
> PS: I haven't written in my last (call it troll post if you like)
> posting that rpn methods come into my way. There are so many things in
> Scheme or CommonLisp which Forth will never! deliver me.
Agreed. A Lisp macro is much higher level than a Forth word in that it
manipulates a syntax tree. But there's a lot you can do with simple
Forth words (such as define other words, together with data structures).
There's also a lot you can learn from Forth's minimalism, in terms of
programming style. But again, you don't have to.
From: Darin Johnson
Subject: Re: what kind on programming is lisp?
Date:
Message-ID: <psyao3st.fsf@usa.net>
Ulrich Hobelmann <···········@web.de> writes:
> You mean of Forth instructions that use the stacks? I think that the
> main difference between Forth and Lisp is memory layout.
I've always thought a big distinguishing characteristic of Lisp is
that it's designed to use abstract symbols. You could build this
in Forth of course, but "out of the box" Lisp has an abstract feel
to it whereas Forth feels concrete. But they are both low level
toolboxes for building higher level programs.
> Agreed. A Lisp macro is much higher level than a Forth word in that
> it manipulates a syntax tree.
A Lisp macro is really a lot like Forth's EVALUATE. It doesn't
manipulate a syntax tree per se, it just builds up a list which
then gets evaluated. (that's another distinguishing characteristic
of Lisp, that the dividng line between data and programs is blurred)
--
Darin Johnson
Support your right to own gnus.
Ulrich Hobelmann wrote:
> Agreed. A Lisp macro is much higher level than a Forth word in that it
> manipulates a syntax tree.
Since there is no syntax tree in Forth, there is no need for a Forth macro
to manipulate it. The most "complex" metaprogramming stuff in Forth I've
done is a regexp compiler, and it's less than 200 lines long (almost
everything is macros). It's not just a recursive backtracking regexp
matcher, it's a compiler that generates highly efficient code that performs
the pattern matching (about 10 times faster as the PCRE library).
--
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/
Bernd Paysan wrote:
> Ulrich Hobelmann wrote:
>
>>Agreed. A Lisp macro is much higher level than a Forth word in that it
>>manipulates a syntax tree.
>
>
> Since there is no syntax tree in Forth, there is no need for a Forth macro
> to manipulate it. The most "complex" metaprogramming stuff in Forth I've
> done is a regexp compiler, and it's less than 200 lines long (almost
> everything is macros). It's not just a recursive backtracking regexp
> matcher, it's a compiler that generates highly efficient code that performs
> the pattern matching (about 10 times faster as the PCRE library).
>
True.
Probably I have too little Forth experience to see how you would
implement something like Lisp macros in Forth.
Is there info available on your regexp matcher (docementation, code)?
How does the result match the string?
Ten times faster than PCRE sounds really impressive, especially since I
always assumed that PCRE only builds an automaton and traverses that,
which amounts to the same as threaded Forth code. And you *have* to
perform one state transition per character read, don't you?
Bernd Paysan wrote:
> Ulrich Hobelmann wrote:
>> Agreed. A Lisp macro is much higher level than a Forth word in that it
>> manipulates a syntax tree.
>
> Since there is no syntax tree in Forth, there is no need for a Forth macro
> to manipulate it. The most "complex" metaprogramming stuff in Forth I've
> done is a regexp compiler, and it's less than 200 lines long (almost
> everything is macros). It's not just a recursive backtracking regexp
> matcher, it's a compiler that generates highly efficient code that performs
> the pattern matching (about 10 times faster as the PCRE library).
Nice! Is the code available?
--
Gareth McCaughan
.sig under construc
Gareth McCaughan wrote:
> Bernd Paysan wrote:
>
>
>>Ulrich Hobelmann wrote:
>>
>>>Agreed. A Lisp macro is much higher level than a Forth word in that it
>>>manipulates a syntax tree.
>>
>>Since there is no syntax tree in Forth, there is no need for a Forth macro
>>to manipulate it. The most "complex" metaprogramming stuff in Forth I've
>>done is a regexp compiler, and it's less than 200 lines long (almost
>>everything is macros). It's not just a recursive backtracking regexp
>>matcher, it's a compiler that generates highly efficient code that performs
>>the pattern matching (about 10 times faster as the PCRE library).
>
>
> Nice! Is the code available?
>
Since Mr Paysan set a followup to c.l.l. only (I suspect he doesn't read
this NG), he's not likely to read your message. However I posted my
reply separately to c.l.f also.
Ulrich Hobelmann wrote:
[Bernd Paysan:]
>>> Since there is no syntax tree in Forth, there is no need for a Forth macro
>>> to manipulate it. The most "complex" metaprogramming stuff in Forth I've
>>> done is a regexp compiler, and it's less than 200 lines long (almost
>>> everything is macros). It's not just a recursive backtracking regexp
>>> matcher, it's a compiler that generates highly efficient code that performs
>>> the pattern matching (about 10 times faster as the PCRE library).
[me:]
>> Nice! Is the code available?
[Ulrich:]
> Since Mr Paysan set a followup to c.l.l. only (I suspect he doesn't
> read this NG), he's not likely to read your message. However I posted
> my reply separately to c.l.f also.
I assumed that if someone writes an article only about Forth
and requests followups only to comp.lang.lisp then it
probably means that he reads comp.lang.lisp. Maybe that
was naive.
--
Gareth McCaughan
.sig under construc
=?iso-8859-1?q?F=F6rster_vom_Silberwald?= wrote:
.
.
.
> I am really not after speed but using Forth in scietific computing is
> pure lunacy. I sometimes get the feeling that some Forther have never
> seen anything else than Forth. Believe me they will miss a lot!
Me, I have used a lot of things. For me Forth has a hobby interest now, but I
first started using it when outside requirements stopped me from having the
right tool for the job so I had to find something to work round them.
>
> PS: I haven't written in my last (call it troll post if you like)
> posting that rpn methods come into my way. There are so many things in
> Scheme or CommonLisp which Forth will never! deliver me.
Not directly, but you can easily put on the extensions that will deliver them
to you. There is a functional language, Joy by Manfred von Thun, that ended
up using a very Forth-like approach. You might want to have a look at the
thread on the subject at news:comp.lang.functional which I have started cross
posting here on news:comp.lang.forth. Ulrich Hobelmann has also replied to
that thread, so that might show you it is worth looking at.
P.S. That sort of google groups problem makes it hard for me to spot all
interesting threads, even when they are only on one newsgroup to begin with.
I just find it easier to use google than my newsreader to lurk. PML.
--
GST+NPT=JOBS
I.e., a Goods and Services Tax (or almost any other broad based production
tax), with a Negative Payroll Tax, promotes employment.
See http://member.netlink.com.au/~peterl/publicns.html#AFRLET2 and the other
items on that page for some reasons why.
Doug Hoffman Mar 4, 2:27 pm wrote:
As an appendix: sorry there is indeed an indication where the posts
will go.
Typically when I post from comp.lang.lisp I expect that postings will
appear only in that forum.
When I post from Mozilla I always have the chance to remove additional
newsgroups, though.
Förster vom Silberwald