Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!pdn!oz!alan
From: alan@oz.nm.paradyne.com (Alan Lovejoy)
Newsgroups: comp.lang.modula2
Subject: Re: type conversion question
Message-ID: <6187@pdn.paradyne.com>
Date: 14 Jun 89 14:00:09 GMT
References: <14226@watdragon.waterloo.edu> <7465@xenna.Encore.COM>
Sender: news@pdn.paradyne.com
Reply-To: alan@oz.paradyne.com (Alan Lovejoy)
Organization: AT&T Paradyne, Largo, Florida
Lines: 40

In article <7465@xenna.Encore.COM> soper@xenna.UUCP (Pete Soper,,,) writes:
>From article <14226@watdragon.waterloo.edu>,
> by djsalomon@watdragon.waterloo.edu (Daniel J. Salomon):
>> 
>> Use the standard procedure VAL.  E.g.
>
>  Which standard? Many implementations do not have VAL as it came along with
>the 3rd edition of the Wirth book and the single pass compiler. Also, I
>suggest Wiener and Ford as an alternative to the Wirth book.

VAL has always been in the language.  Where did you get the idea that it first
appeared in PIM2v3?

The problem with both VAL and "type transfer functions" (e.g., TYPENAME(x))
is that many compiler writers seem unable or unwilling to implement them
correctly.  VAL is supposed to perform MEANING-PRESERVING conversions that
may require changing the number and/or value of bits.  Type-transfer functions
are supposed to perfrom BIT-PATTERN-PRESERVING conversions which may change
the meaning of a value (e.g., CARDINAL(-1)).  But many implementations do
not conform to this specification.  And PIM2v3 never clearly stated the
compatibility rules of the LONG types, so of course every implementation
adopts its own policy as it sees fit (if it even HAS long types).

The other favorite misinterpretation of the definition of the language is
the nutty idea that providing set types  which can contain more elements 
than the number of bits in a WORD is "nonstandard."  It is NOT!!!!  PIM2
says only that implementations are not REQUIRED to support set types which
can contain more elements than the number of bits in a WORD.  But they are NOT
FORBIDDEN from doing so.  Any implementation that does not support SET OF CHAR
is useless for most purposes.  SETs are  almost never used for anything except
SET OF CHAR.

One of the best things about the coming ISO standard for Modula-2 is that
it promises to fix this intolerable situation.


Alan Lovejoy; alan@pdn; 813-530-2211; AT&T Paradyne: 8550 Ulmerton, Largo, FL.
Disclaimer: I do not speak for AT&T Paradyne.  They do not speak for me. 
______________________________Down with Li Peng!________________________________
Motto: If nanomachines will be able to reconstruct you, YOU AREN'T DEAD YET.