NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/lang/c/18691

Internet Message Format | 1992-12-21 | 5.3 KB

Path: sparky!uunet!spool.mu.edu!sdd.hp.com!cs.utexas.edu!usc!srhqla!quest!kdq
From: [e-mail] (Kevin D. Quitt)
Newsgroups: comp.lang.c
Subject: Re: Questions about token merging and trigraphs
Message-ID: <[e-mail]>
Date: Sat, 19 Dec 92 11:51:08 PST
References: <[e-mail]>
Reply-To: srhqla!quest!kdq
Organization: Job quest  [phone],  So Cal: [phone]
Lines: 118

[e-mail] (Dr. T. Andrews) writes:

> Underneath a bogus ``reply-to'' header, [e-mail] writes..

    My reply-to line goes out of this site as "srhqla!quest!kdq"; if
another site is munging that, there's little I can do.

> First, necessity: a trivial proof exists that there is no machine
> on which the characters can not be represented in some other way.

    The necessity is to provide a *single* representation that is
accpeted on *all* machines.  Certainly, one could map the APL character
set onto ASCII, and program that way; the fact that something can be
done does not make it useful or meaningful.


> Second, utility: the most commonly-cited character set for  which
> trigraphs are ``justified'', EBCDIC, certainly has the characters
> in question.  Disagreement on  print-train  convention  may  mean
> that  some  people  see  different  characters  in place of their
> expected, on some of their printers.  A  similar  problem  exists
> for  people  whose  terminals  replace  certain  ASCII characters
> according to local preference.

    From the rationale:

"C derived its repetoire from the ASCII codeset.  Unfortunately the
ASCII repetoire is not a subset of all other commonly used character
sets, and widespread practice in Europe is not to implement all of
ASCII either, but use some parts of its collating sequence for special
national characters.

"The solution is an internationally agreed-upon repetoire, in terms of
which an international representation of C can be defined.  The ISO has
defined such a standard: ISO 646 describes an _invariant subset_ of
ASCII.

"The characters in the ASCII repetoire used by C and absent from the
ISO 646 repetoire are: # [ ] { } \ | ~ ^  <<hash mark, left and right
bracket, left and right brace, back-slash, vertical bar, tilde and
caret>>.

"...The obvious idea of defining two-character escape sequences fails
because C uses all the characters which _are_ in the ISO 646 repetoire:
no single escape character is available.  The best that can be done is
to use a _trigraph_ - an escape digraph followed by a distinguishing
character."


> This means that, without trigraphs, some  people  would  have  to
> accept  what  their printers give; others would have to configure
> terminals to show the ASCII characters when writing C.  It is not
> clear that trigraphs are preferable.

    This means that some people would have alphabetic characters in
place of those glyphs.  How'd you like to have code that looked like:

for L i=0; i<max; i++ M J
   aNiO = i;
K


> Had the utility been clear, there would have been a large body of
> prior  art  from  which  to  draw.  The lack of such a large body
> suggests that the problem is insufficiently bothersome to justify
> the development of the art.

    Either that, or those other systems had to put up with the ugliness
above, or they couldn't run C, or programs written on them couldn't be
ported to other unlike systems.


> Third, harm: it is by no means  unlikely  that  ``can't  happen''
> error  messages  may  include  text such as ``zero quark-count??!
> contact support''.  The harm of having the error message  mangled
> is small, but the benefit of this replacement in a string appears
> to be zero.  If you have to use a trigraph for the character,  it
> seems unlikely that you will benefit by putting it in a string.

"?? was selected as the escape digraph because it is not used anywhere
else in C (except as noted below); it suggests that something unusual
is going on.  The third character was chosen with an eye to graphical
similarity to the character being represented.

"The sequence ?? cannot currently occur anywhere in a legal C program
except in strings, character constants, comments, or header names.  The
escape character sequence '\?' ...was introduced to allow two adjacent
question-marks in such contexts to be represented as ?\?, a form
distinct from the trigraph."



> Reduced program legibility is also harmful.

"The committee makes no claims that a program written using trigraphs
looks attractive.  AS a matter of style, it may be wise to surround
trigraphs with whitespace, so they stand out better in program text.
Some users may wish to use preprocessing macros for some or all of the
trigraph sequences".


> Thus, because trigraphs are unnecessary,  insufficiently  useful,
> and  occasionally harmful, I judge that they were an invention of
> questionable virtue.

    I believe I've shown that your first adjective is incorrect, and
your second is questionable.  As to harm, no code was broken; the only
thing that changed was the appearance of some output - and you've
yourself indicated that is of no great harm.

    After all, if some port of C is willing to use alphabetics that
correspond to the ASCII codes for the 9 glyphs, they're free to do so.
It's just that it's possible now, as it was not before, to write C code
that is guaranteed to work on all implementations.


 _
Kevin D. Quitt      96.37% of all statistics are made up.     srhqla!quest!kdq