Re: Semantic Web pneumonia and the Linked Data flu (was: Can we lower the LD entry cost please (part 1)?)

On Mon, Feb 9, 2009 at 9:45 AM, Yves Raimond
> That's exactly the approach we took in http://www.bbc.co.uk/programmes
> and on http://ww.bbc.co.uk/music/beta. We  publish linked data by
> "just" writing a new view in a Ruby-on-Rails-like framework. The cost
> of doing that is really, really extremely low if your web application
> is well designed (well designed = one URI per thing). I think most web
> developers getting the "the website is the API" slogan know exactly
> what to do.

I heartily agree with Rob and Yyves. I think will see (and are seeing)
more linked-data on the web as it becomes easier for web developers
(like the ones at the BBC) to implement data views using their MVC
framework-du-jour with an RDBMs backend. The sweet spot for
linked-data evangelism seems to be web developers who:

- routinely do both data modeling and resource publishing on the web
- have real business needs for providing programmatic access to the
data behind their web applications
- enjoy reading a bit of documentation to guide an implementation

Also important to the process are tools like schemaweb and Sindice
that enable discovery of what vocabularies are in use on the web; as
well as tools like OpenVocab that make it easy to grow vocabularies
collaboratively.

The incentive that I find attractive about linked-data is that it
seems to jive well with WebArch, the self-describing-web [1] and REST
to provide a pattern for API development that goes *with* the natural
grain of the web. It makes me feel like I'm in good company, and in a
growing/vibrant community, which is fun.

I don't want to sound overly negative, because I sincerely believe
that there is real value in some of these technologies. But I think it
will be perceived as a barrier to entry if web developers believe that
they MUST implement a SPARQL service, link out to dbpedia, use some
kind of triple store, implement a Semantic SiteMap or VoID
descriptions to be *really* doing linked-data.

I will admit it seems like a (not insurmountable) barrier to me at least.

//Ed

[1] http://www.w3.org/2001/tag/doc/selfDescribingDocuments.html

Received on Tuesday, 10 February 2009 03:15:54 UTC