The problem may be that linking is not a general
purpose functionality. That is, linking, the term,
is defined to mean too many slightly overlapping
but otherwise distinct categories of functionality.
Again: should the declaration to traverse be the
same as the declaration of a relationship? In
the middle of that is the declaration of location,
but it isn't essential except to traversal. Hytime
separated the notion of linking (here, declaring
a relationship) from location (here, declaring
an address). URIs weld those together. If one
steps back and looks at both approaches, what
falls out of each? I think it important to
understand that without all of the other standard
definitions (eg, http headers), the URI approach
falls apart. That is, it takes more than a URI
to traverse a network.
Standards based simply on momentum can be
the same as clearing land by rolling parked
automobiles at the trees. It works but it
creates its own mess.
len
From: Micah Dubinko [mailto:]
Hi Len,
>Do you have a comfort
>level for having done that or do you think that
>simplification (eg, just two common attributes)
>will support growth?
For general-purpose XML linking, I don't think anyone has a comfort level
for what works and what doesn't, otherwise, this discussion wouldn't be
necessary. :-)
The goal of picking only two attributes is to get the thing off the ground.
Adding all the whiz-bang features in SkunkLink 2.0 or some
application-specific language would certainly take several more attributes
(and not necessarily in the xml:namespace, as mentioned in the Q & A
section)
Growth of standards is achieved through momentum. The situation where
everybody is sick to death of even talking about a subject is a terrible way
to make progress. SkunkLink aims to change that.
