Talk:Left corner parser

AFAIK, a left corner parser is not chart parser

AFAIK, left corner parsers are not chart parsers in the sense that they would keep a chart of already-processed items like CKY-(Younger-Cocke-Kasami) or Earley-Parsers do.

IIRC, they merely keep stacks as the usual Top-Down and Bottom-Up Parsers do.

There exist different algorithms for achieving backtracking and/or look-ahead in Left Corner parsing, though.

I will expand and correct when I find time to do so -- for now, I'll leave the comment here. Please be gentle if this is wrong, this is my first time here.

92.226.109.153 (talk) 11:11, 2 November 2012 (UTC)

Update to the previous point

As I understand it, one would have to say that left corner parsers can use a table driven approach to achieve a sort of a look-ahead, in that it can be extended to using a (pre-calculated) table which holds information on the possible left corners of each rule. These again represent the plausible predictions in the top-down steps. It strongly seems to me that it would be wrong to categorize a left corner parser as a chart parser per se like the CKY and Earley parsers.

92.226.109.153 (talk) 12:21, 2 November 2012 (UTC)