Re: RelaxNG derivatives vs Earley parsing

> Sounds an awful lot like the description of a derivative:

Yes. There’s definitely similarity here. And, in as much as James is
describing a kind of regular expression matching across trees, it’s
possible that it’s a workable idea for writing a recognizer.

My first thought is, can you make it into a parser? The RELAX NG
validator says “yes, this matches” or “no, this doesn’t”. It (sometimes
very frustratingly) doesn’t tell you anything about *how* it matched the
input.

An Invisible XML processor has to be able to assert not only that the
sentence was valid against the grammar but what parse(s) lead to a
match.

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica