Re: [xml-dev] xlink 1.1

["Simon St.Laurent" <simonstl@simonstl.com>]

rjelliffe@allette.com.au wrote:
> ?? I don't know that XLink was ever intended to build a community in that
> sense: xlink is utility conventions. You don't talk of an XML Namespaces
> community either: both are something that people use or don't use to get
> some other job done. MathML, that is a community. RDF, HTML, SVG. You
> wouldn't have expected an XLink conference, for example.

There are plenty of communities around various forms of hypertext - and 
yes, I'd have been happy to go to an XLink conference.  I delivered a 
fair number of XLink talks to Web developers and even Microsoft 
developers who were eager to have those tools - but never got them.

> I don't know, for example, that it is even possible to have a complete
> XLink library (e.g. in Java), except perhaps a set of interfaces or
> abstract classes. Not specific enough semantics.

I think that may depend on how you interpret complete.  Working in every 
context, probably not.  Working within a specific context, a particular 
interface - sure.

> But I do agree that XLink could have been bigger. We ended up with a
> linking system with no semantics and usable syntax in XLink and a linking
> system with strong semantics and unusable syntax in (initial) RDF. RDFa
> still is not officially defined over XML (neither officially nor workable
> in practice), so there I think there is still good potential for people to
> use XLink+GRDDL.

While RDF is certainly related to hypertext, I've never had a strong 
sense that the RDF community was that interested in creating concrete 
hypertexts.  GRDDL seems to be in a similar place.

My general sense is that simple links might better be implemented by 
giving CSS or schemas a means of designating what content in a given 
vocabulary is a link, and of what kind.  It's just an annotation.

External links are still an open field, though.  I expect we'll see them 
implemented through JavaScript eventually, though per-site sandboxing 
makes them difficult.  (That's always been a problem, though.)

XML did well. XSL did pretty well.  XLL didn't.

Thanks,
-- 
Simon St.Laurent
http://simonstl.com/