Newsgroups: comp.lang.c
Path: utzoo!henry
From: henry@utzoo.uucp (Henry Spencer)
Subject: formal language descriptions
Message-ID: <1988Aug7.004203.7178@utzoo.uucp>
Organization: U of Toronto Zoology
References: <1112@garth.UUCP> <1988Aug2.233758.25939@utzoo.uucp> <1151@garth.UUCP>
Date: Sun, 7 Aug 88 00:42:03 GMT

In article <1151@garth.UUCP> smryan@garth.UUCP (Steven Ryan) writes:
>>                                   The problem with formal definitions like
>>the PL/I definition and the Revised Algol 68 Report is that they are
>>hideously unreadable...
>
>As far being `unreadable' or less `accessible,' that's too easy a target. As
>long as we refuse to correct the problems that exists, they will continue.

Quite true.  As long as we refuse to admit that unreadability is a major
problem that greatly reduces the usefulness of formal standards, they will
continue to see little use.  Even if the authors of a formal definition don't
fall into the trap of "lapidary" writing (paring out every non-essential
word that might help the reader make sense out of the formalisms -- a style
which is endemic in mathematics), the result is seldom something that an
uninitiated reader can tackle without help.  Compare this to the appendix
of K&R, or even the X3J11 drafts, which can be heavy going at times but
*can* be understood without formal background.

The attitude that "anyone but an illiterate peasant ought to be able to
read formal definitions" is part of the problem, not part of the solution.
-- 
MSDOS is not dead, it just     |     Henry Spencer at U of Toronto Zoology
smells that way.               | uunet!attcan!utzoo!henry henry@zoo.toronto.edu