Re: In Search of XML Interoperability: XLink + XML Schema =Interoperability?

> In this search for Interoperability, I want to advocate for another way of linking things than Xlink: it's RDF.
> 
> What is RDF's advantage compared to Xlink ?
> 
> Instead of inventing new xlink attributes, e.g. this had been suggested:
> 
> isA | hasA | equiv | isLike | partOf
> 
> RDF allows to link two resources WITH ANY DOMAIN-DEFINED SEMANTICS.
> I'll give examples, and list afterwards the architectural implications.

I think RDF certainly is generic and powerful enough for this purpose
since it essentially is an agreement on how to describe some resource. 
RDF says nothing about what we are actually describing; that is
completely up to the domains to figure out.

So, extending your line of thinking even further, basically any
"language extension/proposal" than wants to describe a property of any
given resource (be the property something like relation-to, linked-to,
encoded-in, etc) can be recast in RDF.  I agree.  Your four examples are
excellent illustrations (well, the first two were more concrete; latter
two were more hand-waving. ;-) ).

In that case, one can probably argue that XLink is simply a particular
instance of RDF, so why bother progressing on XLink or any other
language/extension/proposal that describes some attribute of a resource.

In fact, this point has been acknowledged in the XLink Working Draft at
http://www.w3.org/TR/xlink/#rdf-examples .  They admit it!  XLink can be
written in RDF!

So, I think that it's true that XLink can be recast into RDF.  But that
does not make the problem XLink is trying to solve any easier.  Ok, so
XLink is syntactic sugar for RDFs, but there is real value in defining
and figuring out how to relate and link resources.  And syntactic sugar
is not necessarily bad. :-)  It certainly helps to focus the problem up
front.  You'd agree that having all the power of RDF without a
well-defined way of wielding it can be sorta... dangerous, not to
mention less than useful.

/David

Received on Thursday, 17 February 2000 15:40:40 UTC