|1970||Jay Earley invents the algorithm that now bears his name.|
|1991||Joop Leo describes a way to modify Earleys 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 Earleys algorithm. Their method makes Earleys 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 Earleys. Leo is unaware of Aycock and Horspools 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.|
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 Earleys algorithm for decades. It certainly was <B>myB> standard source. The account of Earleys algorithm is on pages 320-330.
Marpa is based on ideas from John Aycock and R. Nigel Horspools Practical Earley Parsing, The Computer Journal, Vol. 45, No. 6, 2002, pp. 620-630. The idea of doing LR(0) precomputation for Earleys general parsing algorithm, and Marpas approach to handling nullable symbols and rules, both came from this article.
The Aycock and Horspool paper summarizes Earleys 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 <B>notB> easy reading. I have been following this particular topic on and off for years and nonetheless found this paper very heavy going.
Although my approach to parsing is not influenced by Mark Jason Dominuss Higher Order Perl, Marks treatment of parsing is an excellent introduction to parsing, especially in a Perl context. His focus on just about every other technique <B>exceptB> general BNF parsing is pretty much standard, and will help a beginner understand how unconventional Marpas approach is.
Marks book opened my eyes to many new ideas. Both Marks Perl and his English are examples of good writing, and the book is dense with insights. Marks discussion on memoization in Chapter 3 is the best Ive seen. I wish Id bought his book earlier in my coding.
Marks 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>.
Of Jay Earleys 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, Id 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 Earleys 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. Theres 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>), its on pp. 583-584. My methods for producing parse results from Earley sets do not come from Earley 1970, so I am taking Tomitas word on this one.
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>
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>.
Marpas handling of right-recursion uses the ideas in Joop M.I.M. Leos 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.
Wikipedias article on Backus-Naur form is <http://en.wikipedia.org/wiki/Backus-Naur_form>. Its a great place to start if you dont 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 Sanskirt more than 23 centuries earlier in India using a similar notation.
Copyright 2012 Jeffrey Kegler This file is part of Marpa::XS. Marpa::XS 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::XS 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::XS. If not, see http://www.gnu.org/licenses/.
Hey! <B>The above document had some coding errors, which are explained below:B>
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 v5.20.3||MARPA::XS::ADVANCED::BIBLIOGRAPHY (3)||2016-04-05|