Re: predictive parsing and non-recursive predictive parsing

unix.sh@gmail.com writes:


> I have one question regarding the difference between those two:
>
> I can use recursive predictive parsing, which is very straightforward.
> So what's the advantage of non-recursive predictive parsing. To
> perform non-recursive parsing, I need to construct FIRST, FOLLOW sets
> and use explicit stack. On the other hand, recursive predictive
> parsing is very easy to understand. I understand non-recursive calls
> have a better performance than recursive one. Is this the only reason?


The only reasons that I can think of to distinguish between
recursive-descent predictive parsers and explicit-stack predictive
parsers are esthetics and efficiency. Efficiency includes efficiency
of parser construction, whether manual or using a parser generator, as
well as of parsing.


Your comments about FIRST and FOLLOW sets are misleading. You need to
find the FIRST sets in order to make either kind of predictive parser,
even recursive descent. And you don't absolutely need to find the
FOLLOW sets to make either kind of predictive parser, even the
explicit-stack kind. However, if you don't know the FOLLOW sets, you
won't be able to check at parser construction time that your grammar
is really ammenable to predictive parsing, and you won't be able to
produce as high quality of syntax error messages at parse time. Those
motivations for finding the FOLLOW sets apply equally well to both
kinds of predictive parsers.

Return to the comp.compilers page.
Search the comp.compilers archives again.