NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c++-part1/9742

Internet Message Format | 1996-08-05 | 3.8 KB

Path: galaxy.ucr.edu!not-for-mail
From: [e-mail] (Tom Payne)
Newsgroups: comp.lang.ada,comp.lang.c,comp.lang.c++
Subject: Re: C/C++ knocks the crap out of Ada
Followup-To: comp.lang.ada,comp.lang.c,comp.lang.c++
Date: 4 Mar 1996 04:55:38 GMT
Organization: University of California, Riverside
Message-ID: <4hdt4a$[e-mail]>
References: <[e-mail]> <4etcmm$[e-mail]>   <[e-mail]> <4gad29$[e-mail]> <4gl72q$[e-mail]> <4h1vt6$[e-mail]>
NNTP-Posting-Host: corvette.ucr.edu
X-Newsreader: TIN [UNIX 1.3 950824BETA PL0]


Pete Becker writes:
>
> In article <4gl72q$[e-mail]>, [e-mail] says...
> [...]
> >We are talking about a language whose
> >syntax is so convoluted that, Dan Saks devoted over a hundred column
> >inches of the January issue of C/C++ Users Journal to telling
> >professional C/C++ programmers how to parse declarations.  In the same
> >issue, Pete Becker devoted space to explaining the meaning of
> >sizeof(Sample&), which reads, "the number of bytes in the object
> >representation for the type reference-to-Sample."  That value turns
> >out to be (get this!) the number of bytes used to encode objects of
> >type Sample, regardless of the number of bytes used to encode
> >references --- a design decision so obscure and irregular that even
> >the authors of the April Draft of the C++ standard missed it.
>
> Yes, languages need to be explained, and sometimes have subtle points.

I take some exception to this use of the word "subtle."  If wRef is a
Widget reference whose referent is w, then the subexpression wRef in
the expression sizeof(wRef) denotes w, which is a Widget.  So,
technically, it is impossible to apply sizeof to an expression of type
Widget reference, since such an expression spontaneously converts to
type Widget when it is the arguement of sizeof, e.g., sizeof(wRef) IS
sizeof(w).  But, in the expression sizeof(Widget&), the subexpression
Widget& does not denote Widget, so it is astonishing that
sizeof(Widget&) should turn out to be sizeof(Widget) even when the
object representations for the two types are quite different.

Generally, the most natural extension of a concept is the one that
requires the fewest changes to the wording of its definition,
preferrably none.  In this case, C++ violates that principle, yielding
an anomalie that is bizarre, rather than "subtle."

> That does not make any particular language "write only".

The "write-only" epithet is obvious hyperbole; with sufficient
practice, intelligent people can read almost anything (e.g., Hollerith
codes and van Wjingaarden grammars) --- the issue is how much practice
it takes.  Some unfortunate design decisions have made C/C++
considerably more difficult to read than necessary and, specifically,
more difficult than Pascal, Modula, and Ada.  The C/C++ Users Journal
article by Dan Saks was both well-written and helpful.  A similar
article reviewing the basic syntax of type-forming expressions for a
language with less convolute syntax would almost surely be viewed as
unhelpful by similarly experienced programmers of that language.

> >P.S. My students say that C is an in-joke that everyone now knows,
> >while C++ is a shaggy dog whose punch line is STL.
>
> I would hope that you are teaching them critical analysis rather than analysis
> by one-liners. Unfortunately, context-free assertions about the complexity of
> C++ are not a useful base for a serious analysis. What are the alternatives,
> and what, objectively, are the critical needs that those alternatives do not
> answer?

We try!  Actually, the one-liners are a glib way of expressing
judgments, not of arriving at them.  C and C++ are the languages of
choice for these students.  Their exposure to Ada is mostly through
its VHDL derivative, which is used heavily in our hardware courses.

Tom Payne ([e-mail])