W3C home > Mailing lists > Public > semantic-web@w3.org > June 2017

Re: [ANN] Linked Data Templates: First draft and Call for participation

From: Martynas Jusevičius <martynas@atomgraph.com>
Date: Fri, 30 Jun 2017 19:16:53 +0200
Message-ID: <CAE35VmzC13k3vcBAxXbQUqPt4ZF+OK0MiNmQXnJe2jOae=ZsQA@mail.gmail.com>
To: Alfredo Serafini <seralf@gmail.com>
Cc: Kingsley Idehen <kidehen@openlinksw.com>, "Gray, Alasdair J G" <A.J.G.Gray@hw.ac.uk>, public-lod <public-lod@w3.org>, Semantic Web <semantic-web@w3.org>, Hydra <public-hydra@w3.org>, "public-rww@w3.org" <public-rww@w3.org>
Kingsley, Alfredo,

I could mention many reasons why we haven't reused Swagger or a different
API description format: 1) it's not RDF 2) it's not an ontology 3) no
standard way to embed SPARQL 4) depends explicity on HTTP etc.

Most importantly, Swagger is only a description, while LDT not only
provides a description, but also a method to evaluate that description in
order to produce a response [1]. So LDT ontology is more like a definition
of an API, rather than a description. I'm not aware of any other spec that
provides this kind of evaluation. You could compare it to SPARQL algebra

It's not that we decided one day that the world needs one more API
description format. No, LDT has evolved over many years of building SPARQL-
and Linked Data-native systems. Since it's useful to us, we think that by
extension it could be useful for other people working on such software.

[1] https://atomgraph.github.io/Linked-Data-Templates/#http-valuation
[2] https://www.w3.org/TR/sparql11-query/#sparqlAlgebra

On Fri, Jun 30, 2017 at 6:54 PM, Alfredo Serafini <seralf@gmail.com> wrote:

> Hi Kingsely and all
> I agree with this:
>> Have you considered using OpenAPI (nee Swagger) to document your APIs?
>> Doing that would provide another point of intersection between "Web
>> Developers" and "Semantic Web Developers" .
>> [1] https://medium.com/openlink-software-blog/swagger-the-api-ec
>> onomy-rest-linked-data-and-a-semantic-web-9d6839dae65a -- Swagger, the
>> API Economy, REST, Linked Data, and a Semantic Web
> moreover I would guess that approaching to API design as a collaborative
> effort (even before providing the actual underling implementation) could be
> a way to deepen the discussion.
> The community could be interested in merging approaches and best
> practices, as Kingsley said, avoiding choices that could be problematic on
> one of the side (for example opaque URI,, and so on...).
> Given the current status of stadards, with LDP, LDF, it would be useful to
> underline in a practical manner the benifits of a "new" standard around
> resources. For example for some reason I imagine some similarities with the
> ideas behind the "old" Fresnel approach in some way (at least when
> reading/navigating a resource) here (am I completely wrong?), but maybe I
> didn't understand at all the context :-) Engaging people around playable
> wip examples could be a way to focus on the crucial parts.
> my
Received on Friday, 30 June 2017 17:17:27 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:41:56 UTC