Re: How to write this RE? (Barton C. Massey)
16 Aug 1996 11:33:24 -0400

          From comp.compilers

Related articles
How to write this RE? (Johnny Wong) (1996-08-11)
Re: How to write this RE? (1996-08-13)
Re: How to write this RE? (1996-08-13)
Re: How to write this RE? rwatson@CAM.ORG (Richard Edward Watson) (1996-08-15)
Re: How to write this RE? (1996-08-16)
Re: How to write this RE? (1996-08-16)
Re: How to write this RE? (1996-08-19)
| List of all articles for this month |

From: (Barton C. Massey)
Newsgroups: comp.compilers
Date: 16 Aug 1996 11:33:24 -0400
Organization: University of Oregon
References: 96-08-034 96-08-042
Keywords: lex, DFA, comment

There's two things I find amusing about this thread, on some
sort of metalevel:

1) Our esteemed moderator posted somebody's request to solve
their homework problem for them. Many very smart people posted
their solutions to this newsgroup. Bad precedent?

2) The question as originally posted, seems very
straightforward, but it can be read to specify at least two
different obvious problems. I think it makes a very succinct
case for the use of formal methods in compiler construction:
Even if the specifier didn't know how to write the regex
describing the problem (since that was the problem :-), he
could have said something like:
        I need a RE for the language L over alphabet A such that
L = { s in A* | for all a in A . <a,a> is not a subsequence of S }
L = { s in A* | for all a in A . a does not occur twice in S }
The latter meaning is surely more like the natural language
description, but after reading the responses I strongly suspect
that the former is what is intended...

Bart Massey
[Your moderator is not unaware of the temptation for some people to use the
net as a homework-o-matic, but this sounded too hard to be a homework
question. (He rejects lots of them every week.) Or if it is, the professor
likely as not doesn't know the correct answer, either. -John]

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.