Re: linked data mashups

All praise to the work, but..

Although the Semantic Web Client library and middleware that uses it are
exciting, and undoubtedly part of the correct way to go, I am worried about
overclaiming; please can we avoid it.

On 24/11/2008 15:09, "Juan Sequeda" <juanfederico@gmail.com> wrote:

<snip/>

> Now just imagine the Semantic Web Client library on your server, next to the
> rest of the LAMP stack. Your server is the sparql endpoint for the whole
> Semantic Web. No need of querying individual sparql endpoints. This is SQUIN
>
> [1] http://www4.wiwiss.fu-berlin.de/bizer/ng4j/semwebclient/
>
>
"the whole Semantic Web"?
And I think that the Semantic Web Client library does not make such a claim.
This is an excellent aspiration, perhaps between adults in private, but in
public it will always be hard to substantiate, for many reasons that would
make an interesting paper.
In fact it is so easy to refute, it undermines our other claims.

I confess to being a bit sensitive about this; when you are publishing a lot
of data in different datasets, as we are, it can be hard work trying to
ensure that Sindice, SWSE, Swoogle, Falcon-S and anyone else are able to
find them.
Another paper mentioned in this respect on this thread is
http://sw-app.org/pub/isemantics08-sotsw.pdf (Michael Hausenblas et al. I
believe, not Yves Raimond et al.) and which was the start of the wiki page
table.
Of necessity this paper on linkage used what it was able to (as they say in
the text), but as far as I can tell, for example, it seems to have only used
one of the three extant dblp RDF sources, and omitted various others.
Unfortunately the statistics such as the table in the paper then get used
elsewhere without appropriate caveats.

We all do our best to cover everything, but in the world of the web it is
not usually a good thing to claim we have the whole (or even a large part?)
of anything. I can't even find Google making such a claim, but maybe they do
somewhere, if I could work out the text-based search.

As I said, all praise to the work...

Best
Hugh

Received on Tuesday, 25 November 2008 13:43:06 UTC