On 27/01/2011 13:26, Dimitre Novatchev wrote:
>> Anyway, it is an interesting question. I think the same question could be
>> asked, rephrased, as "what is the smallest class of formal grammars that
>> every Xpath (evaluating to boolean) belongs to?"
> It isn't a regular expression -- XPath expressions may have an
> unlimited nestedness.
>
> It is definitely a context-free (CF) grammar and even more
> specifically, it is an LR(1) grammar.
>
> Three years ago I defined the full XPath 2.0 grammar (with one
> exception
Ah, you have interpreted the question quite differently from me (and
perhaps correctly?) I thought it was looking at XPath as a grammar in
which the sentences being described are document instances. You are
talking about the grammar in which XPath expressions are the sentences.
Michael Kay
Saxonica
