Re: [xml-dev] Xlink Isn't Dead

["Peter Hunsberger" <peter.hunsberger@gmail.com>]

On 9/22/06, Ben Trafford <ben@prodigal.ca> wrote:
>
> At 12:59 PM 9/22/2006, Len Bullard wrote:
> >And if the purpose of a given link is NOT to send a resource to a rendering
> >application, what then?
>
>          Then it's irrelevant to the process of what to do with it
> after its modelled. I don't believe you can reasonably standardize
> the behavior of generic applications that operate below user
> visibility, at least in terms of how they deal with modelled
> relationships. Each specific user domain will develop their own
> standards and tools based on what they need the unseen links to do.
>
>          The primary purpose of XLink was to display links in a
> rendering context. That's my point of interest in this discussion.
> I'm sure other people have thoughts on links that operate like
> plumbing, behind the curtains. I really don't, by and large. To me,
> trying to get the concepts behind XLink working at that level is wrong-headed.
>
>          What are your thoughts on the topic?

Just to play the devils advocate for a moment; if all you want is
links at the rendering level why do you even need xLink?

Much of my interest comes at this from the knowledge interchange
aspect: automated ontology exchange and that ilk. As Michael
more-or-less once put it [1], a tag named "postcode" already has
semantics, why does it need to be rewritten to be useful as a link?
(And yes, his example shows why simple links aren't sufficient).

Drilling down from that a level, I'd also ask, why can't the same
relationship traversal mechanism that figures out that it can link a
map to the postcode also figure out that postcode is equivalent to
zipcode in certain contexts?  IOW, if you've got an application that
can understand generalized tag context well enough to do link
rendering then you've got an application that is already doing
relationship traversal that can be used for other purposes...

-- 
Peter Hunsberger

[1] - http://lists.xml.org/archives/xml-dev/200410/msg00300.html