|LL(1) vs LL(k) email@example.com (1992-05-09)|
|Re: LL(1) vs LL(k) firstname.lastname@example.org (1992-05-09)|
|Re: LL(1) vs LL(k) email@example.com (1992-05-09)|
|Re: LL(1) vs LL(k) firstname.lastname@example.org (1992-05-13)|
|Organization:||UIUC Center for Supercomputing Research and Development|
|Date:||Sat, 9 May 1992 15:54:49 GMT|
email@example.com (Terence J Parr) writes:
>Given any deterministic language L, L$ (L appended with EOF) can always
>be described with an LR(0) grammar [Knuth65].
IMHO, this result was of limited value without a usable, deterministic
parser associated with the grammar... obviously, the canonical parser for
the LR(0) grammar cited by this result would almost always be
Years of hard work determined how to efficiently augment a
non-deterministic LR(0) parser with look-ahead sets and operator
precedence (especially DeRemer and Pennello's elegant formulation of LALR)
to make it deterministic for most languages... then (and only then) could
parsers based on LR(0) grammars be really useful.
John R. Grout
University of Illinois, Urbana-Champaign
Center for Supercomputing Research and Development
Return to the
Search the comp.compilers archives again.