Ubuntu Manpage: Marpa::R2::Advanced::Bibliography

Provided by: libmarpa-r2-perl_2.086000~dfsg-8build5_amd64

NAME

       Marpa::R2::Advanced::Bibliography - A Marpa bibliography

History of the Marpa algorithm

       1970
           Jay Earley invents the algorithm that now bears his name.

       1991
           Joop Leo describes a way to modify Earley's algorithm so that it runs in O(n) time for all LR-regular
           grammars.   LR-regular  is  a  vast class of grammars, including all the LR(k) grammars, all grammars
           parseable with recursive descent, and regular expressions.  LR-regular can safely be  thought  of  as
           including all grammars in practical use today, and then some.

       2002
           Aycock  and  Horspool describe a way to do LR(0) precomputation for Earley's algorithm.  Their method
           makes Earley's faster in most practical situations, but  not  all.   In  particular,  right-recursion
           remains  quadratic in the Aycock and Horspool algorithm.  Worst case is no better than Earley's.  Leo
           is unaware of Aycock and Horspool's work and Aycock and Horspool seem unaware of Leo.

       2010
           Marpa combines the Leo and Aycock-Horspool algorithms, in the process making significant  changes  to
           both of them.  The result preserves the best features of both.  Marpa also tackles the many remaining
           implementation issues.

Bibliography

Aho and Ullman 1972
The Theory of Parsing, Translation and Compiling, Volume I: Parsing by Alfred Aho and Jeffrey Ullman
(Prentice-Hall: Englewood Cliffs, New Jersey, 1972). I think this was the standard source for Earley's
algorithm for decades. It certainly was my standard source. The account of Earley's algorithm is on
pages 320-330.

Aycock and Horspool 2002
Marpa is based on ideas from John Aycock and R. Nigel Horspool's "Practical Earley Parsing", The
Computer Journal, Vol. 45, No. 6, 2002, pp. 620-630. The idea of doing LR(0) precomputation for Earley's
general parsing algorithm, and Marpa's approach to handling nullable symbols and rules, both came from
this article.

The Aycock and Horspool paper summarizes Earley's very nicely and is available on the web:
<http://www.cs.uvic.ca/~nigelh/Publications/PracticalEarleyParsing.pdf>. Unlike "Earley 1970", Aycock
and Horspool 2002 is not easy reading. I have been following this particular topic on and off for years
and nonetheless found this paper very heavy going.

Dominus 2005
Although my approach to parsing is not influenced by Mark Jason Dominus's Higher Order Perl, Mark's
treatment of parsing is an excellent introduction to parsing, especially in a Perl context. His focus on
just about every other technique except general BNF parsing is pretty much standard, and will help a
beginner understand how unconventional Marpa's approach is.

Mark's book opened my eyes to many new ideas. Both Mark's Perl and his English are examples of good
writing, and the book is dense with insights. Mark's discussion on memoization in Chapter 3 is the best
I've seen. I wish I'd bought his book earlier in my coding.

Mark's book is available on-line. You can download chapter-by-chapter or the whole thing at once, and
you can take your pick of his original sources or PDF, at <http://hop.perl.plover.com/book/>. A PDF of
the parsing chapter is at <http://hop.perl.plover.com/book/pdf/08Parsing.pdf>.

Earley 1970
Of Jay Earley's papers on his general parsing algorithm, the most readily available is "An efficient
context-free parsing algorithm", Communications of the Association for Computing Machinery, 13:2:94-102,
1970.

Ordinarily, I'd not bother pointing out 35-year old nits in a brilliant and historically important
article. But more than a few people treat this article as not just the first word in Earley parsing, but
the last as well. Many implementations of Earley's algorithm come, directly and unaltered, from his
paper. These implementers and their users need to be aware of two issues.

First, the recognition engine itself, as described, has a serious bug. There's an easy fix, but one that
greatly slows down an algorithm whose main problem, in its original form, was speed. This issue is well
laid out by Aycock and Horspool in their article.

Second, according to Tomita there is a mistake in the parse tree representation. See page 153 of "Grune
and Jacobs 1990", page 210 of "Grune and Jacobs 2008", and the bibliography entry for Earley 1970 in
"Grune and Jacobs 2008". In the printed edition of the 2008 bibliography, the entry is on page 578, and
on the web (<ftp://ftp.cs.vu.nl/pub/dick/PTAPG_2nd_Edition/CompleteList.pdf>), it's on pp. 583-584. My
methods for producing parse results from Earley sets do not come from Earley 1970, so I am taking
Tomita's word on this one.

Grune and Jacobs 1990
Parsing Techniques: A Practical Guide, by Dick Grune and Ceriel Jacobs, (Ellis Horwood Limited:
Chichester, West Sussex, England, 1990). This book is available on the Web:
<http://www.cs.vu.nl/~dick/PTAPG.html>

Grune and Jacobs 2008
Parsing Techniques: A Practical Guide, by Dick Grune and Ceriel Jacobs, 2nd Edition. (Springer: New York
NY, 2008). This is the most authoritative and comprehensive introduction to parsing I know of. In
theory it requires no mathematics, only a programming background, but even so it is moderately difficult
reading.

This is "Grune and Jacobs 1990" updated. The bibliography for this book is available in enlarged form on
the web: <ftp://ftp.cs.vu.nl/pub/dick/PTAPG_2nd_Edition/CompleteList.pdf>.

Kegler 2013
My writeup of the theory behind Marpa, with proofs of correctness and of my complexity claims is
available online: <https://docs.google.com/file/d/0B9_mR_M2zOc4Ni1zSW5IYzk3TGc/edit?usp=sharing>.

Leo 1991
Marpa's handling of right-recursion uses the ideas in Joop M.I.M. Leo's "A General Context-Free Parsing
Algorithm Running in Linear Time on Every LR(k) Grammar Without Using Lookahead", Theoretical Computer
Science, Vol. 82, No. 1, 1991, pp 165-176. This is a difficult paper. Unfortunately, there is no copy
of it on-line.

Wikipedia
Wikipedia's article on Backus-Naur form is <http://en.wikipedia.org/wiki/Backus-Naur_form>. It's a great
place to start if you don't know the basics of grammars and parsing. As Wikipedia points out, BNF might
better be called Panini-Backus Form. The grammarian Panini gave a precise description of Sanskrit more
than 23 centuries earlier in India using a similar notation.

Copyright and License

         Copyright 2014 Jeffrey Kegler
         This file is part of Marpa::R2.  Marpa::R2 is free software: you can
         redistribute it and/or modify it under the terms of the GNU Lesser
         General Public License as published by the Free Software Foundation,
         either version 3 of the License, or (at your option) any later version.

         Marpa::R2 is distributed in the hope that it will be useful,
         but WITHOUT ANY WARRANTY; without even the implied warranty of
         MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
         Lesser General Public License for more details.

         You should have received a copy of the GNU Lesser
         General Public License along with Marpa::R2.  If not, see
         http://www.gnu.org/licenses/.

POD ERRORS

       Hey! The above document had some coding errors, which are explained below:

       Around line 29:
           Expected text after =item, not a number

       Around line 40:
           Expected text after =item, not a number

       Around line 53:
           Expected text after =item, not a number

perl v5.38.2                                       2024-03-31             Marpa::R2::Advanced::Bibliography(3pm)