Re: Hydra: similarities and differences

Hey Ruben,

I'm aware of your work.

The similiarity of Graphity with both Hydra and the LDP is the concept
of resources and collections/containers, which is common to many
LDP-like systems.

Also similarly to Hydra, we use URI templates extensively, both to
match request URIs and to build new instance URIs. Unlike Hydra
however, we do not attempt to specify allowed operations, as we feel
this is covered by HTTP and by optional access control rules [1].
Likewise we do not specify representation constraints in the main
vocabulary and leave it to optional SPIN constraints [2].

As far as I can see, the clearest overlap with LDF is mapping Linked
Data access to SPARQL queries (we assume a triplestore-based system):
http://www.hydra-cg.com/spec/latest/linked-data-fragments/#sparql-query-results
I tried to cover the idea of URI to SPARQL mapping in more detail in
the post on GitHub. We use SPIN extensively to (meta-)model SPARQL.

You are right, there are predefined rules in the specification, about
how the processor should behave based on the application description.
The description itself is however fully declarative, RDF-based and
compatible with OWL. I gave an example of it in my reply to Markus:
http://lists.w3.org/Archives/Public/public-declarative-apps/2014Jun/0003.html

I hope this clears some things. We can bounce some emails back and
forth to identify the overlaps more clearly.

Martynas

[1] http://www.w3.org/wiki/WebAccessControl
[2] http://spinrdf.org/spin.html#spin-constraints

On Fri, Jun 20, 2014 at 4:13 PM, Ruben Verborgh <ruben.verborgh@ugent.be> wrote:
> Dear all,
>
> The previous post asked to introduce ourselves:
> I'm Ruben, semantic hypermedia researcher at Ghent University – iMinds.
> My interest goes out to the development of intelligent clients,
> and the APIs that enable those clients.
>
> I've just become aware of this group and immediately joined,
> but I'm not entirely sure about its context.
>
> In particular, could you shed some light on how it relates to:
> - the Hydra Core Vocabulary [1], which introduces hypermedia controls?
> - Linked Data Fragments [2,3], which defines APIs that allow flexible querying?
> Both initiatives are discussed in the Hydra Community Group [4].
>
> As far as I understand, your goals seem similar,
> but the ideas are executed differently.
> For instance, similar to Linked Data Platform,
> you seem to rely on pre-determined rules [5]
> rather than in-band information in representations.
> But I might be misinterpreting, so this is why I ask.
>
> The similarities with the Hydra Core Vocabulary seem
> that you also have similar concepts (resources, collections).
> The similarities with Linked Data Fragments seem
> that you also incorporate existing interfaces, such as SPARQL endpoints.
>
> Thanks for telling us about this and best of luck with this group!
>
> Ruben
>
> [1] http://www.hydra-cg.com/spec/latest/core/
> [2] http://www.hydra-cg.com/spec/latest/linked-data-fragments/
> [3] http://www.hydra-cg.com/spec/latest/triple-pattern-fragments/
> [4] http://www.w3.org/community/hydra/
> [5] https://github.com/Graphity/graphity-browser/wiki/Linked-Data-Processor-specification

Received on Saturday, 21 June 2014 12:30:24 UTC