Re: Reinventing the wheel ( ...again ); this time with RDF In/Out Services

yes, its reinventing the wheel.

the buzzword name for the idea is most probably "semantic web services"
and there is a gazillion publications on how to describe them, discover 
them, automatically connect them, map them, exchange them, ....

a "down-to-earth-simple" approach is, for example, this one:

@inproceedings{DBLP <>:conf/pakm/TakabayashiNTFY06,
  author    = {Yuya Takabayashi and
               Harutaka Niwa and
               Mitsuharu Taneda and
               Naoki Fukuta and
               Takahira Yamaguchi},
  title     = {Managing Many Web Service Compositions by Task Decomposition
               and Service Quality Evaluation.},
  booktitle = {PAKM},
  year      = {2006},
  pages     = {291-302},
  ee        = {},
  crossref  = {DBLP:conf/pakm/2006},
  bibsource = {DBLP,}


Es begab sich aber da Aldo Bucchi zur rechten Zeit 05.12.2006 10:08 
folgendes schrieb:
> Hi all,
> Don't you hate it ( or love it in a subtly twisted way ) when you run
> across the very same "I bet no one's ever done this" feature you've
> been working heads down for last two nights?
> Well, this time its about this RDFIO thingy... a declarative way to
> wrap services.
> I had two names for it... the first was RDF BAGs ( before/after graphs
> ) or RDFio, but liked the former because it emphasizes the "wrapping"
> action and the latter is just too dumb.
> Anyway,
> Is there any further work around this?
> say... RIROs & REST for example..
> ( noticed how rest has become a little unrestful these days? )
> I've had a lot of fun trying to create some simple algorithms to
> undertsand at a higher level the cost and amount of arc traversals
> that could be made over a set of BAG-wrapped services ( ok... RIRO
> services ). Would make for a nice paper, server yourselves.
> Another neraby issue... does anyone know if there's a "query service
> description" taxonomy /vocabulary?
> When using this wrappers there is a recurrent need to specify further
> "capabilities" of the service.
> I need a way to state that
> paramA, fulltext search filter
> paramB, mask filter ( full match )
> sort on B
> sort on C
> the service provides a count() service ( taking the same params but on
> another URL, or passing a flag, or using OPTIONS, who knows )
> get the point?
> I believe that with this new REST momentum there is pretty nice chance
> to leap this evident gap and push rdf closer to mainstream. To me,
> it's clear that one fundamental mistake in the marketing of the sw has
> been neglecting the need ( and value ) of these simple "recipes".
> We are missing the glue... the transport.
> I'm now leading a big (10-20 devs) project and I'm willing to give RDF
> a try as the framework for integrating part of the data ( mashup )...
> again, I twistedly love the pain found in early adoption.
> Any pointers appreciated!

DI Leo Sauermann 
P.O. Box 2080          Fon:   +49 631 205-3503
67608 Kaiserslautern   Fax:   +49 631 205-3472
Germany                Mail:

Received on Monday, 11 December 2006 08:59:00 UTC