At 03:49 AM 2006-09-28 -0400, Ben Trafford wrote:
> 2) Links need to be declared in generic XML, so no forced syntax
> like XLink 1.0. This is so that all the various dialects people have used
> to describe linking can get along without breaking (backwards and
> forwards compatibility).
Why would it necessarily be in generic XML? XPath isn't, though it
inherits the "X" prefix.
> 3) XLink is -conceptually- on the right side of the 80/20. Forget
> the syntax, and focus on the actual ideas -- do they cover what needs to
> be covered? Especially if it were possible to easily extend them in the future.
I'd think links can be interpreted as a separate class of "data about
relationships".
From this angle, its use and arguments about its importance and non-importance
(which defines the "right" in your 80/20) would be different from just
considering
links' contribution to styles. Relationships (like the arcs of a graph)
can exist without
the present instantiation of objects (the nodes of the same graph). The
instantiation
could be deferred, implied, virtual, theoretical, temporary, aliased, etc.
But the
fact that if a graph's relationships are stable and need to be described,
links will
come in handy here.
cheers.
