Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!crdgw1!sunbelt!eaker
From: eaker@sunbelt.crd.ge.com (Charles E Eaker)
Newsgroups: comp.lang.forth
Subject: Re: Random comments
Message-ID: <1483@crdgw1.crd.ge.com>
Date: 3 Aug 89 19:20:57 GMT
References: <10425@dasys1.UUCP>
Sender: news@crdgw1.crd.ge.com
Reply-To: eaker@sunbelt.crd.ge.com (Charles E Eaker)
Organization: General Electric Corp. R&D, Schenectady, NY
Lines: 50

In article <10425@dasys1.UUCP> aj-mberg@dasys1.UUCP (Micha Berger) writes:
>As someone new (less than 1 month, only have one working program - life)
>to FORTH, 2 things really bug me. Why is it that we have 2!, 2DUP, 2CONSTANT,
>but D. ? It seems inconsistant. You can't use 2 as a prefix across the board
>(2+ and 2* are taken) but why not use D?

Here's why. The prefix 2 generally means operate on 2 things of the
usual size.  The "usual size" in Forth has traditionally been a 16-bit
"cell".  Note that the operations 2!, 2DUP, and 2CONSTANT could care
less what the 2 cells are interpreted to be. They could be two single
numbers or 1 double number or 4 characters or one or more bit strings
representing a set or representing a packed record structure or
representing a date.  But the dot operation which prints its operand
must interpret the operand's contents.  D. interprets two cells as
being one double number. Another operation, say .TIME might interpret
the two cells as representing seconds since some specified moment. I
suppose a reasonable action for 2. would be to print the two cells as
two consecutive single numbers.

   [What bugs ME is the traditional asymetry between D. and ." and the
    inclination to use the dot as a prefix in other print words such as
    .TIME . Instead, I think that the dot should uniformly be a suffix
    so that to print something right adjusted in a field whose width is
    on the stack we would use D.R and ".R and TIME.R but I'm afraid
    that using dot as a prefix to mean print is now too deeply
    ingrained.]

>                                         I also mind that DO loops up to,
>not including the limit - it's not intuitive.
>

On the contrary, I find DO loop limits in Forth to be quite intuitive.
I always get it right in Forth, I never get it right in Pascal, I
often get it right in C. A limit, after all, is easily thought of
as something which is never quite reached.
If you want to initialize SIZE cells of an array, for example, 
starting at address ARRAY with the pattern PAT, you would use:
     ARRAY SIZE + ARRAY DO I PAT ! LOOP
otherwise you would constantly have to subtract one:
     ARRAY SIZE + 1- ARRAY DO I PAT ! LOOP
which strikes me as being an unnecessary and counter-intuitive nuisance
in a piece of code that becomes a cliche in any array processing
application. (Or would you rather the value of SIZE be one less than
the actual size of the array? That's even more counter-intuitive.)

For being new to Forth, you're asking good questions.
Chuck Eaker                                |  eaker@sunbelt.crd.ge.com
Software Engineering Program               |  eaker@crdgw1.crd.ge.com
GE Corporate Research & Development Center |  eaker@crdgw1.UUCP
P.O. Box 8, K-1 3C12 Schenectady, NY 12301 |  518 387-5964