Re: [xml-dev] ten years later, time to repeat it?

["Simon St.Laurent" <simonstl@simonstl.com>]

Eric van der Vlist wrote:
> I don't think that subsetting only XML 1.0 (or even
> only XML 1.0 + namespaces) would be very useful.

I think it would be the right place to start, however.  It's unfortunate 
that so much effort has been put into burying the 'XML' core under 
specifications that boggle users and implementers alike.   Lots of 
applications and users, however, either don't bother with the crap on 
top, or decide on their own subset in those layers, and do just fine.

> That means that you should probably cleanup the most basic pieces (XML
> 1.0 + namespaces in XML + XML Base + xml:id) and provide a kind of
> "specifications profiles" explaining how the upper pieces can safely and
> sanely be selected and used together.
> 
> This also means that you'd have to debate over highly controversial
> stuff such as namespaces and schema languages.

Schema languages (except DTDs, for now) aren't actually part of XML. 
Namespaces, though completely broken in theory, don't cause that much 
trouble in practice, once you learn that thinking about the theory only 
causes unnecessary pain.

XML Base, XML Include, and (to a lesser degree) xml:id aren't my 
favorite specs, but they do operate at the foundation level and at this 
point should probably be wrapped in, yes.

Making it a principle that the subset's documents have to work with 
existing XML 1.0 processors probably leaves all of the original specs 
(NS, XI, XB, xml:id) outside of XML 1.0 itself in a conformance gray 
area.  Over the very long term, though, wrapping them together should 
actually make it easier to deploy them.

Anyway, we'll see what happens.  The thought experiment has already 
generated interesting conversations, so I'd call it a success so far.

Thanks,
Simon St.Laurent
Retired XML troublemaker
http://simonstl.com/