hello subbu.
Subbu Allamaraju wrote:
I suspect that this is heading towards XLink.
not really. the important difference is that XLink links instances (it
specifies links for concrete URIs), whereas LDX specifies how to
discover links in media types. i think this is an important
difference, and my post definitely failed to be very specific about
this. the reason why i think it's important is that REST centers
around media types and following links, and LDX should be a god match
with that.
Ah I see. Just read your post again. I don't think link discovery on its
own solves anything, unless the author of the representation or the
media type says something about a link appearing at any given position
an an XML document.
that's exactly what he's doing wih LDX: an XPath can be as specific as
required by the media type and can specifically only select one link
somewhere, or all links anywhere (think about //html:img/@src, for
example, selecting all links to images in XHTML). then you associate
information with that link in the form of the link relation, the
resulting media type, or the expected URI template. this kind of
description makes it much easier for consumers of RESTful services to
discover and understand the links they'll find in XML resources. i am a
bit confused why you think that this does not solve anything. it
definitely does not solve everything in the sense that you can readily
generate code from it to automatically implement an application, but it
certainly would be good enough to for example drive a "link discovery"
component of a REST toolbox, or maybe to just serve as documentation for
RESTful services independent of the schema language used for the media
types exposed by the service.
cheers,
erik wilde tel:+1-510-6432253 - fax:+1-510-6425814
[email protected] - http://dret.net/netdret
UC Berkeley - School of Information (ISchool)
