> CYK and Early parsers are both actually simpler than parsing with derivatives, once you've thrown this paper's optimizations in. Yeah, GLR is very elegant as

> CYK and Early parsers are both actually simpler than parsing with derivatives, once you've thrown this paper's optimizations in. Yeah, GLR is very elegant as well; I'm a little puzzled by this paper's initial claim (and similar claims in the Yacc is Dead paper): > Current algorithms for context-free parsing inflict a trade-off between ease of understanding, ease of implementation, theoretical complexity, and practical performance. No algorithm achieves all of these properties simultaneously. GLR is has good performance and isn't much harder to understand than LR, which is a very elegant theory. So I assume that 'ease of implementation' is the property the authors think GLR lacks. But ease is a little hard to measure, until you have an equally fast and production-ready implementation of both algorithms.