[Techtalk] The Internet is my functional programming language

[Wim De Smet]

On Thu, Nov 27, 2008 at 2:52 AM, Andrea Schweer <andrea at andaka.org> wrote:
>> I know much of this information is available in human-readable format
>> (Horizons, TIGER map, MathWorld, Weather Underground, OandA, etc), but is
>> it
>> available in consistent, machine-readable format?
>
> Consistent, maybe not.
> Machine-readable, have a look here: http://www.programmableweb.com/apis --
> there is a big bunch of services out there. You might want to read up on
> web services (eg start at http://en.wikipedia.org/wiki/Web_service); have
> a look at WSDL, SOAP and UDDI.
>
> I think it's not so much the case of "nobody has thought of doing this"
> but rather, you need to know which terms to use to look for it.

The wikipedia article on REST[1] and a couple of its external links
are also pretty enlightening on some of the challenges involved (in
the case of SOAP, complexity and interoperability issues). I'm
spending some time with taverna[2] at the moment which tries to
incorporate web services straight into the programming language
(bioinformatics often relies on querying data from a multitude of
places, and transforming it in remote places e.g. clusters).

The Semantic Web approach actually takes this one step further. The
typical web service doesn't have much semantic information attached,
i.e. they return a string or an int or an array of ints, which might
be referenced as a type in an xsd somewhere but is often not linked to
any ontologies. If they were linked to ontologies, a lot of the
reasoning that humans have to do (like, how do I transform the output
of this webservice into input for this one) could be done by machines.
There's a good article by Berners-Lee about it somewhere on the web, I
can dig it up if you like.

greets,
Wim

[1]: http://en.wikipedia.org/wiki/Representational_State_Transfer
[2]: http://taverna.sourceforge.net/