Re: Web Services on LLD

In message 
<3A59BB6451C972429019B12996F92DAD048F9FEB@frodo.zbw-nett.zbw-kiel.de>, 
Neubert Joachim <J.Neubert@zbw.eu> writes
>
>We'd like to mark completion of our action item, recorded in
>http://www.w3.org/2005/Incubator/lld/minutes/2011/02/10-lld-minutes.html#action13
>, to look into and say more about LLD Web Services.
>
>We've put together a little text that we hope outlines the *why*, gives a few
>examples of such services, and proposes a possible path ahead (or two)
>for LLD web services. Should the premise be accepted by the group, we
>hope this format will make the work of integrating it into the larger
>deliverable smoother and easier.
>
>Naturally, your comments, thoughts, and suggestions are most welcome.
>If the group wishes to more forcefully endorse the second "proposal" - a
>codification of a core set of library linked data web services - we'd want to
>tweak the last sentence (providing of course, the group would like the
>proposal to even remain).
>It is available at:
>
>http://www.w3.org/2005/Incubator/lld/wiki/Web_services_on_LLD

I would suggest that you mention the Linked Data API [1] as one 
potential model for such web services. While it was developed in order 
to provide an alternative, developer-friendly interface to triple stores 
which would often already have a SPARQL end-point, it is also of value 
in its own right as a generic RESTful API design.

 From a user's perspective, it is of no interest what lies behind the API 
interface - whether a triple store or an entirely different database 
technology - so long as asking the same question causes the same answer 
to be returned.  So it could be a mechanism for harmonising access to 
triple stores and other data being presented as Linked Data.

 From the user's PoV, the approach of having to deal with a unique 
interface, and data format, for each Web Service (as per the examples 
you cite) doesn't really scale beyond accessing a handful of such 
services.

Richard

[1] http://code.google.com/p/linked-data-api/
-- 
Richard Light

Received on Thursday, 10 March 2011 10:27:00 UTC