- From: Erik Wilde <dret@berkeley.edu>
- Date: Fri, 16 Nov 2012 18:40:43 -0800
- To: David Booth <david@dbooth.org>
- CC: Kjetil Kjernsmo <kjetil@kjernsmo.net>, public-ldp@w3.org, Mark Baker <distobj@acm.org>, mike amundsen <mamund@yahoo.com>
hello david.
> Ah-ha! Now I see what Erik meant. Just to be clear on terminology,
> AFAIK Amundsen's requirements go *far* beyond the conventional
> definition of "hypermedia"
> http://en.wikipedia.org/wiki/Hypermedia
> as Amundsen's requirements are all about supporting hypermedia
> *applications*, which of course is what we want to do with LDP and
> Hypermedia As The Engine Of Application State (HATEOAS)
when you talk about any kind of information system providing services,
an important question always is what kind of agent is using it. in
traditional hypermedia, humans are making the interaction choices, so
human-readable labels suffice. in REST as a SOA approach, it's machines,
so we need a little different approach. in this case, the affordances
need to be more user-friendly from a machine-point of view, mike and i
have talked quite a bit about "API UX", i.e. how easy is it for you as a
developer to conveniently consume a service. RESTful design these days
has mostly come to the conclusion that any RESTful service needs a
description that clearly specifies both data models and interaction
models. there are different ways how you can do that (and many different
frameworks have been proposed), but no matter what your preference is,
you somewhere somehow have to address those two issues. RDF currently
gives you a framework for issue #1, and issue #2 is left an exercise for
groups such as us. we can do it informally or formally, and we can try
to do it so that it's reusable or not, but no matter how we approach the
problem, we have to solve it.
cheers,
dret.
--
erik wilde | mailto:dret@berkeley.edu - tel:+1-510-2061079 |
| UC Berkeley - School of Information (ISchool) |
| http://dret.net/netdret http://twitter.com/dret |
Received on Saturday, 17 November 2012 02:41:15 UTC