Forgot to cc the list. -Kei
Kei Cheung wrote:
Hi Don,
Donald Doherty wrote:
Hi Kei,
Great to hear from you and the podcast idea is excellent. I wonder,
is there any effort to have semantics attached to the podcasts?
That'd be cool to have triples indicating content.
Check out the Neuropod feeds at:
http://feeds.nature.com/nature/podcast/neuropod?format=xml
These feeds (XML format) describe different neuroscience podcasts. One
can choose to listen to podcasts relating to dopamine neuron,
Alzheimer disesae, etc.
The semantic web and podcasts can make neuroscience "music to our ears".
Have fun!
-Kei
The main difficulty with making an iPhone app work with multiple
SPARQL endpoints, from my perspective, is due to 1) differing
coverage and content, 2) differing structures, and 3) differing
URIs. Number 3 is the easiest to deal with.
An application that goes straight to the endpoints without using a
server to do translation, aggregation, etc. across endpoints must
either do all of this itself (too much for an iPhone app) or get its
full content from the endpoint it uses.
Detailed PubMed information, for instance, is a very important
feature to scientists. Only the Bio2RDF endpoints provide details
including abstracts in RDF. All endpoints seem to have the authors
associated with a particular paper but a big limitation of every
endpoint is that it is impossible to reconstruct the order of the
authors. (At least I haven't found a way.)
Don
On Dec 17, 2009, at 3:26 PM, Kei Cheung wrote:
Hi Don,
Donald Doherty wrote:
Scott Marshall suggested that people on the HCLS list would have
fun with my SemanticWb iPhone and iPod touch app
(www.actionpotential.com ).
Mobile neuroscience sounds an interesting concept. You may also
want to use it to play the neuroscience podcasts of interest.
It's one of many technologies I play around with to test the
potential (and current limits) of the Semantic Web and Semantic
Web technologies (for instance, see my site using MIT Simile
technology at www.historicshadyside.org).
Currently SemanticWb is using Bio2RDF endpoints. I'm working to
make it possible for the user to select from a number of SPARQL
endpoints but it's unfortunately non-trivial.
Is it because of the bandwidth issue? I think it'd be nice if it
can also access other endpoints like the HCLS KB's. That reminds me
that Scott had mentioned to me recently that the HCLS KB endpoint
at Berlin could not be accessible. I wonder if it has been fixed yet.
Cheers,
-Kei
Enjoy! And I'll be happy to talk about lessons learned.
Donald Doherty, Ph.D.
donald.dohe...@actionpotential.com