[
Lists Home |
Date Index |
Thread Index
]
Or that one set of types suits all?
Or that one set of metatypes suits all?
At one point in the saga that was HyTime, links
were broken up into types of links so that the
problem of how an application structured its
data could be accounted for. The classic
conundrum was CGM. One wanted to pass it
the linking information, but leave it to
figure out how to translate that in terms
of its own structures and behaviors. Not
oddly, over time, the IETM CGM advocates
began to make its support more SGML-like,
introducing notions such as type definitions.
SVG was the admission that one could do
vector graphics in markup (long overdue)
and get the advantages of a common metastructure
that make interoperation with the rest of
the framework more tractable.
len
-----Original Message-----
From: Elliotte Rusty Harold [mailto:elharo@metalab.unc.edu]
XLink can provide the syntax
and very basic linking semantics on top of which this can be built.
It only becomes hard if you insist that XLink must define all
possible linking semantics and behavior for all applications that use
it. This is a mini version of the common markup fallacy that one
vocabulary suits all uses.
|
