"Andrew Haley" <andrew29@[EMAIL PROTECTED]
> wrote in message
news:13k2nqv56t3757a@[EMAIL PROTECTED]
> In comp.lang.forth cr88192 <cr88192@[EMAIL PROTECTED]
> wrote:
>
>> "Andrew Haley" <andrew29@[EMAIL PROTECTED]
> wrote in message
>> news:13k0uqca84ruue1@[EMAIL PROTECTED]
>>
>>> So, cr88192, we have now discovered that as well as being unfamiliar
>>> with Forth you are also unfamiliar with PostScript. I'm guessing that
>>> you are not really familiar with *any* reverse polish programming
>>> langauge. However, you don't let a mere lack of knowledge get in the
>>> way of having an opinion.
>
>> well, I have used them before, but mostly I have created my own
>> (used for things like an IL), in the past I have also used them for
>> other things (such as geometric calculations, parametic geometry,
>> ...).
>
>> now, here is a difference:
>> I almost always ordanize stack args differently '<value> <name>
>> <whatever>'', because this is more useful in my experience. this
>> managed to interfere with my thoughts of postscript as well, since I
>> haven't really used it much, in itself, but more just sort of partly
>> used it as a syntactic template for some of my own languages, which
>> have drifed pretty far away.
>
>> now, my opinion is based on my own personal experiences with said
>> languages, which generally lead me to believe that such languages
>> are not particularly readable...
>
>> I can complain similarly about s-expressions, though I have a lot
>> more experience with those...
>
>> so, mostly, I am a fan of C-like systax, as IMO, it is actually
>> about the only real kind of syntax that both looks nice and I find
>> readable...
>
>
>> here is an example of my RPNIL language (a piece of the IL code dumped
>> from
>> my compiler):
>
> Well, hold on a moment. Those of us who know Forth well (and
> presumably also users of other langauges that use RPN) already know
> that automatic translation from other languages almost always results
> in programs that are execrably bad. You have a spectacularly awful
> example here: it is very unlikely that any programmer would write
> anything like this.
>
yes, this is true. autogen code is particularly terrible looking.
but, for non-autogen code, the question is between tolerable and nice
looking or readable.
I will claim, expressions like:
x x * y y * +
or:
x x times y y times add
are at a natural disadvantage to:
x*x+y*y
for many any varied reasons (at least wrt normal humans).
one can assert, well, maybe it is because that is what people know already
(most people being exposed to algebra and C family languages), or maybe
there is something more, intrinsic...
now, I don't say we need full on C (or that C proper is, syntactically,
the
one language to rule them all).
but, IMO:
pyth(x,y) sqrt(x*x+y*y);
does have at least some niceness difficult to capture in most different
forms of syntax...
> Andrew.
|
