[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Infoset chewing gum (was Re: linking, 80/20)

• To: xml-dev@l...
• Subject: Re: Infoset chewing gum (was Re: linking, 80/20)
• From: "Simon St.Laurent" <simonstl@s...>
• Date: Mon, 19 Aug 2002 14:38:00 -0400
• In-reply-to: <3D6138CA.8040801@d...>

Erik Wilde writes:
> maybe that's the point. i don't have a markup problem. the problem is 
> that some useful semantics (hyperlinking) are being viewed as if they 
> were tied to markup (which they currently are because xlink 1.0 only 
> defines markup).

Funny, I thought the problem we were discussing was the difficulty of
representing a particular set of semantics in markup.  To me, that's a
markup problem, not a reason to march off into abstraction.

> however, the semantics should be defined somewhere else, and then
> people could choose whatever markup they like, xlink 1.0 markup for
> the namespace-tolerant people, and xhtml markup for the html-legacy
> providers. and if the svg people decide that they don't want any of
> these, then they can invent svg-specific markup for the xlink data
> model.

I think there is already an answer for this, and it's called
transformation.  You can implement whatever semantic processing you like
by defining a transform from XYZ markup into your markup.  That's
sometimes tricky, but it's far more concrete than creating Infoset
Information Items that live out in the ether.

> so my point is: define semantics and a associated data model, and
> then let people decide which markup they would like to have. tying
> every discussion to markup and namespaces and the problem that people
> have or may have with colonized names simply misses the point.

And I think you've forgotten about the toolset we already have in your
rush to leave behind the concrete markup provides for the abstraction of
link semantics.

> and again (even though this is dangerous): xml schema did it exactly
> that way, because there were some important semantics to be captured,
> and they did so using infoset extensions. i know that the psvi
> contributions are not everybodies darling, but so far i haven't seen
> any reasonable way of solving the problem differently.

I think some people might suggest that the reason W3C XML Schema created
the PSVI mess was connected to an effort to solve problems for which
markup is probably not the right solution, compounded with a lot of
misspent energy expended on making attributes and elements more similar
in theory than they are in practice.

The PSVI is a much larger and uglier chewing-gum mess than what you
appear to be proposing, but that doesn't make your Infoset proposal any
more attractive.


-------------
Simon St.Laurent - SSL is my TLA
http://simonstl.com may be my URI
http://monasticxml.org may be my ascetic URI
urn:oid:1.3.6.1.4.1.6320 is another possibility altogether

• References:
  • Re: Infoset chewing gum (was Re: linking, 80/20)
    • From: Erik Wilde <net.dret@d...>

• Prev by Date: Re: Infoset chewing gum (was Re: linking, 80/20)
• Next by Date: Re: CDATA section
• Previous by thread: Re: Infoset chewing gum (was Re: linking, 80/20)
• Next by thread: Re: Infoset chewing gum (was Re: linking, 80/20)
• Index(es):
  • Date
  • Thread