Re: Just an observation

[Simon Coulter <shc@xxxxxxxxxxxxxxxxx>]

On 17/04/2008, at 2:46 AM, James Lampert wrote:
> Personally, I've always felt that both Niklaus Wirth and Dennis  
> Ritchie
> were full of equine scat on that one: if a language as advanced as  
> PL/I,
> and one as simple as all but the most rudimentary dialects of  
> BASIC, can
> both recognize whether "=" was being used in an assignment context  
> or an
> equality context, then why SHOULD there be separate operators, at  
> least
> for general use?

Can't recall about Pascal but C requires two operators because it  
allows assignment IN an equality context (i.e., if ( a = b) which  
doesn't behave in the obvious manner). The only reason C can do this  
is because it also has the behaviour of 0 is false and anything else  
is true. Any language that allows both assignment and equality in the  
same statement requires two operators in order to distinguish between  
both operations.

I doubt Pascal allowed assignment in a conditional expression and I  
think it had two operators simply to "force clarity of intent" on the  
programmer.

However, I'm in agreement with you about using separate operators for  
assignment and equivalence. Interestingly Rexx uses doubled operators  
for 'strict' comparisons. This to me makes sense. Comparing two  
values (ignoring leading zeroes, and leading or trailing blanks) is a  
different operation from comparing two values EXACTLY. Different  
operations require different operators.

There are certain conditions in systems coding where being able to  
assign a value in a conditional statement MAY result in code  
efficiencies but these are in the minority--and often only in the  
mind of the coder. Seems pointless to me to provide language support  
for a construct that:
	a) is fraught with peril
	b) is considered poor practice
	c) is rarely used

So says Simon currently involved in getting C code that compiled  
cleanly under the 440 C compiler to compile under the new PASE C/C++  
compiler. I thought the old C compiler was strict but the new one  
seems over the top. The dreaded "CZM0280-Function argument assignment  
not allowed" error. It's not my code by the way--all MY stuff  
compiles cleanly under both compilers--but more 'open-source' crap  
that follows both old-fashioned and poor coding practices "for  
portability reasons". Forgive me for appearing cantankerous ...

Regards,
Simon Coulter.
--------------------------------------------------------------------
   FlyByNight Software         OS/400, i5/OS Technical Specialists

   http://www.flybynight.com.au/
   Phone: +61 2 6657 8251   Mobile: +61 0411 091 400        /"\
   Fax:   +61 2 6657 8251                                   \ /
                                                             X
                 ASCII Ribbon campaign against HTML E-Mail  / \
--------------------------------------------------------------------