Re: p:input doppelgangers

In some private correspondence, Alessandro pointed out that I was
wrong when I said that JAXP allowed an XPath expression without a
context. (I'm sure one of the drafts did, but that's neither here nor
there...)

Instead, JAXP says that the default context is an empty document node.

Further, he argued convincingly that the absence of a context node
really is an XPath 2.0 feature.

I wonder if it wouldn't be best (and certainly more reasonable for
implementors) if we adopted the JAXP strategy and said that the
default context is an empty document node instead of an error.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Wherever they burn books they will
http://nwalsh.com/            | also, in the end, burn human beings.--
                              | Heine