Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!julius.cs.uiuc.edu!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!pitt!willett!ForthNet
From: ForthNet@willett.pgh.pa.us (ForthNet articles from GEnie)
Newsgroups: comp.lang.forth
Subject: ANS FORTH TECHNICAL COMMITTEE
Message-ID: <2300.UUL1.3#5129@willett.pgh.pa.us>
Date: 4 Feb 91 12:35:31 GMT
Organization: (n.) to be organized.  But that's not important right now.
Lines: 87

Category 10,  Topic 2
Message 307       Sat Feb 02, 1991
F.SERGEANT [Frank]           at 18:03 CST
 
     Elizabeth Rather wrote:
 >I'd really like to see a "show of hands" ...
  ...
 >Once again, stand up and be counted:  how do you vote?
  ...
 >OK, FOLKS, PUT UP OR SHUT UP.  I HAVE LISTED SOME ISSUES THAT ARE 
 >OPEN FOR DISCUSSION AT THE NEXT MEETING AND I PERSONALLY PLEDGE TO 
 >REPRESENT MAJORITY VOTES ON THE ABOVE ISSUES.  TAKE A POSITION AND 
 >TELL ME WHAT YOU THINK AND WHY!
 .
     I replied to it with some comments (because, naturally, it pleased  me to
do so).  I was willing to state my opinion (my "vote").  I did  not
necessarily go into detail as to the WHY.  I was not attempting to  convert
others to my views, merely to state what my views were -- more  or less in
answer to Elizabeth's request.  
 .
     Then Jan Stout commented on some of my comments.
 .
 JS>How can you expect people to go along with your ...
 .
     I really don't expect them to.  
 .
 JS>How can you expect people to go along with your ( row col -- ) 
 JS>proposal if you don't give some example of the benefits of this 
 JS>factoring.
 .
     Mainly, I was just saying that I use and prefer ( row col -) to (  col
row -).  I am not saying anyone else needs to prefer it.  I think  my primary
reason is it is easier for me to visualize going down and  then across that to
visualize going across and then down.  What with  +  to modify one and +UNDER
to modify the other and the possibility of AT  being a CODE definition, I
don't see machine efficiency as the  overwhelming criterion for the order of
the arguments.  Elizabeth &  PolyForth probably prefer ( row col -) for
"better" reasons than I  offer.  I'd like to hear them.  Meanwhile she's got
one more "vote" for  her preferred method.  I often do something like 
STARTING-POINT 2@ ( y  x) dX + AT  but how can I really object to STARTING-
POINT 2@ ( x y) dX  +UNDER AT ?  And, while I said what I used & preferred, I
also  suggested it be omitted from the standard all together.
 .
 .
 JS>Again I don't understand your enthousiasm for FOR/NEXT over 
 JS>DO/LOOP.
 .
     Well, I don't understand it either.  FOR/NEXT just resonated with  me
somehow.  I liked it from the start and never use DO/LOOP.  I don't  even keep
DO/LOOP loaded in my system.  As I mentioned, I don't object  to others using
it.  And, with the proposed standard allowing its  requirement of DO/LOOP to
be met merely by including the source code  for it, I'm looking upon the
proposed standard more favorably than I  used to.
 .
     Furthermore, you offer efficiency arguments with the examples
 .
 : TYPE   SWAP OVER +  ?DO  I C@ EMIT LOOP ;
 instead of
 : TYPE   FOR C@+ EMIT NEXT ;   with : C@+   1+ DUP 1- C@ ;
 .
     might I offer clarity, simplicity, readability arguments in favor  of
FOR/NEXT?  In addition there are the many loops that do not need  access to
the index.  FOR/NEXT's index is down counting.  In many  situations its index
I can be used as conveniently as FOR/NEXT's I.   And, since implementations
vary, I don't see that it is clear that  DO/LOOP is more efficient even in
those cases that do use the  upcounting index.  Depending on the
implementation, it may be that  either computing the index is less efficient
or that testing the loop  termination condition is less efficient (or both)
than using FOR/NEXT  either with or without its index.  I believe that Robert
Berkey has  offered implementation examples of both FOR/NEXT and DO/LOOP in
terms  of the other.  All in all I am of the opinion that neither can be 
declared the "best" loop based only in terms of efficiency of its 
implementation.  All I know is I like FOR/NEXT and find it much more  pleasing
than DO/LOOP.  I think it is because it is simpler to use.  I  use it (as I
think I've mentioned) with the loop done u times rather  than u+1 times.  I
have even been speculating lately whether I should  convert to an *up-
counting* index in FOR/NEXT.  One of the arguments in  favor of FOR/NEXT has
been its simplicity of keeping only one item on  the return stack and the ease
of testing it for loop termination (<is  it zero yet?>).  I would give that up
to get my up-counting index.  I  have come to no conclusion yet about this.
 .
    -- Frank            ---- continued ----
-----
This message came from GEnie via willett.  You cannot Reply to the author
using email.  Please post a follow-up article, or use any instructions
the author may have included (USMail addresses, telephone #, whatever).
Report problems to: dwp@willett.pgh.pa.us or uunet!willett!dwp