Article about C
utzoo!decvax!ucbvax!npois!harpo!zeppo!wheps!ihnss!houxi!u1100a!rick
utzoo!decvax!ucbvax!npois!harpo!zeppo!wheps!ihnss!houxi!u1100a!rick
Mon Jan 25 13:04:19 AEST 1982
The article about 'C' in the December Sigplan states
''It is the purpose of this paper to to begin a syntactic
description of C that will reduce the number of semantic "hooks"
required by incorporating many of the semantic idioms into the syntax,
producing a clearer, more complete definition of C.
''The syntactic description was specifically scanned in an effort
to maximize the correctness of the syntax, while minimizing the number of
semantics associated with the grammar. The study focused on three areas of
concern in Bell Labs' version of C, whose syntax description is listed in
[Kernighan and Ritchie]:
1. Function definitions and declarations are very incomplete,
requiring a large number of semantic check conditions
implementable in the syntax.
2. Statement descriptions are confusing, and in one case, in error.
3. Data and function types are used haphazardly in the grammar,
with no thought to trap the utilization and scope of storage
class specifiers and types.''
The article contained a proposed reworked BNF grammar for the language.
Unfortunately, it was unreadable because it had been printed on a line-printer
with a weak ribbon. The results when the ACM people tried to offset
print from that as a master were predictable.
So I wrote to the authors and requested a copy with a readable BNF,
and they sent me their original Technical Note, with the following letter:
Dear Rick:
Enclosed find a copy of our paper as you
requested. In answer to your question, we have had
no replies from anyone else at Bell Labs.
Since the grammer is basically "untested", we
would appreciate any irregularities or improvements/
suggestions you may have.
Sincerely,
/signed/
Patrick Fitzhorn
How about it, people!
Rick Thomas -- houxi!u1100a!rick
PS: The authors' addresses are:
Patrick A. Fitzhorn and Gerold R. Johnson
Department of Computer Science
Colorado State University
Ft. Collins, Co. 80523
More information about the Comp.unix.wizards
mailing list