Re: breaking up?

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

On 06 Aug 2001 10:13:44 +0100, Guy Murphy wrote:
> Who oversees and administers such a fragmentation?

I'm not proposing a fragmentation of organizations or specifications.
The fragmentation in the community already exists - it has for a long
time - and fragmentation in use continues to grow.  Rather than
overseeing and administering what appears to be a natural process, I'd
suggest that specification-creating organizations recognize this process
and consider their work in the light of two very different basic
priorities.

Should things continue as they are, I expect to see XML-based and
PSVI-based processing go their separate ways over time.  XML-based will
still be readable if not especially pleasing to PSVI-oriented
processing, while PSVI will become less and less accessible.

> Also, simply by way of exploring your proposal, not seeking to be
> argumentative, how would you propose to fragment things more than they
> already are where xml schema, transformation, querying etc. are already
> discrete specs?

Again, I'm not proposing the creation of micro-specs by specification
creators.  I'm proposing that these folks take into consideration that
"XML environments" are substantially more diverse than HTML
environments, with a full range from ultra-simplistic parsing to complex
developments using W3C XML Schema, XSLT, XLink, and XPointer in unison.
There seems to be a contingent of developers morally opposed to any
subsetting processes, who don't seem to have noticed that many people
perform a 100% subset of specs by not using them at all.
 
> It is still posible to have a simple XML parser with no validation support,
> and none of the new fangled flashing lights... how might you propose things
> would be simpler?

By expecting that the simplest model is a likely possibility in all
transactions using XML, rather than making claims like XML+XSD=XML 2.0.

> We have a really Good Thing(tm) with our ability to communicate data in a
> common format, it's easy to forget how it was before... are we going to tear
> the tower down ourselves and go back to speaking in a babel?

We're already in babel.  We have a couple of choices: babel where XML
syntax is the unifying glue, or babel where expectations well beyond XML
syntax are the unifying glue.  Some of us would like the former, some
the latter.

> I'm a big fan of simplicity over complexity, but normaly before the item
> under consideration is set in stone. After the fact one tends to leave lots
> of fracture marks.... not opposing your suggestions, simply saying that at
> this juncture one would need to be very careful how one went about it, and I
> have reservations that the resulting split wouldn't introduce more problems
> that we already have.

I think you're looking for some kind of formal split, which I'm not
proposing.  I'm just saying that we already have an informal split in
the XML universe, and that we need to take a look around as we pile
stones on top of stones.