Lists Home |
Date Index |
John Cowan wrote:
> What do you mean by "derivation"? Do you mean "subsetting"? If so,
> what subset do you have in mind? RELAX NG is fairly orthogonal, so if
> you decide you aren't going to use some feature, just don't use it.
> Obvious candidates would be the list element, datatypes, separate
> files -- depending on your needs.
No, I meant type derivation. Sorry, should have been more explicit about
this. I'd be very interested in your view.
> The current big players are committed to XSD, without doubt
> (including my own
> employer, alas). However, the forthcoming ISO status of RNG will help
> it in certain other markets.
ISO status is obviously a plus but doesn't do that much on its own to solve
the "public perception" issue.
> I think this is true. I'm certainly not a chief evangelist, even if
> I sometimes play one on xml-dev; because of 2), there isn't likely to
> be a full-time c.e.
That's what I suspected. Having witnessed plenty of doomed battles in the
SGML/XML standards realm (HyTime, RDF, Topic Maps... sorry, couldn't resist
ruffing a few feathers :-), I'd tend to be quite skeptical myself about the
prospect that RNG (John, you're strengthening my case by abbreviating the
name!) could gain widespread adoption. What it needs is a killer app, and
that has to be more than unioning and intersecting schemas, which to be fair
is not a use case I've much heard about.
I think this killer app might be related to type derivation. There are a lot
of problems with XSD derivation that are coming to light as real-world cases
like UBL are addressed.
I'm a typical user of this stuff, and I don't really care if XSD or RELAX NG
"wins". But I would like to have better standard scheme language than the
current version of XSD. If I were a RELAX NG stakeholder I would try to do
1) Piggyback on an effort like UBL by adding type derivation mechanisms that
support real-world business cases (not always the case with XSD).
2) Build support for basic RELAX NG features and the more advanced type
derivation features into Xerces. I understand Xerces doesn't support RELAX
NG at all right now, is that correct? If so, this is probably more of a
barrier to widespread adoption than any other factor.
3) Launch a media blitz to publicize the ISO standardization, support for
UBL (say), Xerces support, etc. <flogging><horse type="dead">Ideally this
effort could be leveraged to announce a new name as well.</horse></flogging>
I guess I'm as attracted to Quixotic endeavors as the next guy...