Re: ldp-ISSUE-38 (filtering, inlining): filtered representations and inlining [Linked Data Platform core]

hello henry.

On 2012-11-19 16:27 , "Henry Story" <henry.story@bblfish.net> wrote:
>You can map anything to RDF, so there being an RDF interface makes
>absolutely no requirement on the backend storage. It's all about the API,
>and at that level we are doing things semantically, i.e.: with RDF here.

of course you can map anything to anything, but that's not all that
relevant here. and yes, we're defining RDF, nobody every questioned that.
but we're *also* defining a service (see charter), and that means that we
are designing just this: the service surface where client and server
interact. we specify the concepts we want to handle, and the ways in which
they can be interacted with. and that's all we do. specifically, we're not
making any assumptions about how client or server are doing their jobs
behind the service surface. unless we give up on the REST part and assume
a homogeneous RDF world and just define vocabularies and no interaction
protocol, there's no way around this.

>That is Atom(Pub) this is LDP. For those who want Atom(Pub) atom(Pub)
>already exists. We are hering mapping this to RDF, which essentially will
>make it possible to work with any syntax.

the problems that need to be addressed and solved for building some
container/member managing service are always the same, irrespective of
your preference of syntax, and metamodel behind it. we can learn from
atompub, and we'll end up specifying something close to it in the end, so
maybe it's worth taking a look.

cheers,

dret.

Received on Tuesday, 20 November 2012 00:54:25 UTC