XOM API for Processing XML with Java

[John Cowan <jcowan AT reutershealth.com>]

Elliotte Rusty Harold scripsit:

> One of the reasons decision on this point was so 
> tough to reach and so weak in the end was that different groups were 
> going in different directions on both the meaning and use of relative 
> URIs for namespace names. 

On the contrary, I think everyone is now striving to avoid them.  The main
reason they weren't just made illegal is that Microsoft had deployed tools
that generated them (in order to create document-local namespaces), and
some unknown number of documents already contained them.  I very much doubt
whether this practice is spreading, though: people are switching to urn:uuid:
or other schemes that make uniqueness straightforward.

> The second more practical and probably solvable question is how one 
> identifies an absolute URI.

A URI is absolute iff it matches this regular expression:

        ^[A-Za-z][A-Za-z0-9+.-]*:

See the productions "absoluteURI" and "scheme" in RFC 2396.

> The stumbling block is I can't decide if URIs containing multiple 
> colons are legal; e.g. is urn:publicid:foo::bar OK?  

It is entirely legal.  According to RFC 3151, it is equivalent to the
public identifier "foo////bar", which is not a valid FPI but is a legal XML
public id.

See the productions "absoluteURI", "opaque_part", and "uric_no_slash" in
RFC 2396, which justify it.

(Writing RFC 3151 was a bitch.  It involved carefully reconciling the
constraints of RFC 2141 (URN Syntax), RFC 2396, and ISO 9070 on FPIs.)

-- 
John Cowan <jcowan AT reutershealth.com>     http://www.reutershealth.com
I amar prestar aen, han mathon ne nen,    http://www.ccil.org/~cowan
han mathon ne chae, a han noston ne 'wilith.  --Galadriel, _LOTR:FOTR_