[Hplusroadmap] [OWW-Discuss] Tapping into open source / open access and doingslightly more

[Antoon Goderis]

Hi Bryan,

> It is my understanding that we have a hard time representing
> *functionality* in any formal way, other than sending the program
> itself. Think about Google Code Search, or the apt-get interface. You
> can only search for tags, not functionality. If you wanted to search
> for functionality, you'd have to write the code. Oh, and I guess you're
> done then. :)

My own PhD was about discovery of functionality of workflows. Let's
just say it's a challenging problem with room for more research :-)

Have a look at work from the semantic web services community. It's
certainly their goal. I'm not convinced about the current solutions, I
have to add. They typically come with a steep learning curve. For an intro
look at chapters 1 and 2 of
http://www.cisa.inf.ed.ac.uk/ssp/pubs/goderis.pdf
It's old; Duncan Hull's thesis should be more relevant - he's putting in
final corrections at the moment so not yet available.

A good example of the current state of the art in terms of a Semantic Web
language based representation that captures functionality (in this case
based on the relationship between input and output) can be found in
http://www.cs.man.ac.uk/~hulld/papers/HZBH+06.pdf

A simple version of this exists for bioinformatics services
www.springerlink.com/index/XF2BWXQPFB0ANB0W.pdf

It is currently being operationalized by the EBI in BioCatalogue, a
project developing a registry for services.

Antoon