Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!gatech!gitpyr!gus
From: gus@pyr.gatech.EDU (gus Baird)
Newsgroups: comp.lang.pascal
Subject: Standard Pascal
Message-ID: <8736@pyr.gatech.EDU>
Date: 12 Jul 89 00:42:07 GMT
Distribution: usa
Organization: Georgia Institute of Technology
Lines: 64


>Article 2169 of comp.lang.pascal:
>>From: milne@ics.uci.edu (Alastair Milne)
>Newsgroups: comp.lang.pascal
>Subject: Re: Standard Pascal
>
>schwartz@shire.cs.psu.edu (Scott Schwartz) writes:
>>Actually GET and PUT are much better than READ and WRITE.  One reason is
>>that you can interrogate the file variable without committing to read it.
>>For example, to strip leading whitespace from the input stream, you 
>>do the following:
>
>>	while (input^ in [space, tab, newline, cr, formfeed]) do
>>		get (input)
>
>    How does this constitute not commiting to read the file variable?  You are
>    doing exactly that when you dereference the file variable to get its
>    window.  I don't see how this differs at all from
>
>	repeat
>	  read( input, ch);
>	until not (ch in [space, tab, newline, cr, formfeed]);
>
>    except that you first copy what would have been the file variable
>    reference into another variable.  Saving one variable seems to me a
>    trivial difference in transport.  Replacing the references so that all
>    "input^"'s become "ch"'s is worse, but not much.
>
>>This works without side effects, so that it can be bundled up in a
>>black box and called by anyone.  

No! There's a TREMENDOUS difference between
                                                     
    repeat 
	read(FileVar,ch)
    until ch {is whatever};
                                                     
and
                                                     
    while FileVar^ {is not whatever} do
	get(input);
                                                     
far beyond the trivial savings of a local variable.
                                                   
The point is that that the version with read consumes the character 
that stops the loop.  This means that the routine after the skiptill 
routine will have to be written to get its first character from the 
skiptill code, and get its others from FileVar.
This constitutes an unconscionable tight coupling between the routines,
needlessly complicating the program and adversely affecting reliability
and re-usability.  
We want our procedures to be as loosely coupled as possible.

Turbo "Pascal" is a good language and development system, but it 
ain't Pascal.  That's why I won't use it in my courses, where I'm
training up computer scientists.  Pascal's frequently-damned I/O model
is actually one of its very best features - budding computer scientists
who understand that model are well equipped to deal with real-world
I/O systems.

-- 
gus Baird
School of ICS, Georgia Institute of Technology, Atlanta, Georgia, 30332
...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!gitpyr!gus