Re: Discussion: OWL-S and Industry Adoption

I generally agree with Evan's comments.  And I would add that there are
some organizational and temporal factors here.  Creating a base language
for the Semantic Web (DAML+OIL which became OWL) was the initial focus
of the DAML program, and work towards that goal significantly predated
the work on OWL-S.  Indeed, DAML+OIL was already mature when the work on
OWL-S started.  Also, as the initial focus of the program, the work on
DAML+OIL benefitted from a lot of attention and fresh energy from
several key people in the program and at W3C.  Also, the technical space
associated with services is broader and there is a broader array of
possible approaches, with less common understanding about the tradeoffs
associated with the different approaches.  Also, a good deal of
attention is currently focused on the commercial Web services
technologies, and there's already a lot there for people in commercial
environments to grapple with for the time being.

In any case, I think it's still "early days" with respect to OWL-S and
other such SWS technologies.  I think the recent W3C Workshop on
Constraints and Capabilities was extremely interesting in revealing the
real-world needs for standards that will support more expressive service
descriptions.  These needs are being felt quite urgently in some quarters.

As someone already noted, OWL-S has only just been submitted to W3C.  I
wonder if you looked at the point in time at which DAML+OIL was
submitted, if you would find there had already been significant
transition into the commercial realm at that time.  (I think not.)

In any case, the W3C team, in their response to the OWL-S submission,
has stated that they expect to have a workshop on SWS in the first half
of 2005.  Hopefully that will be a significant catalyst for things to
start moving along more rapidly.

Regards,
David Martin
SRI International


ewallace@cme.nist.gov wrote:

> Josh Grob started a discussion on OWL-S with the following:
> 
>>Last week I attended a semantic web seminar hosted by Eric Miller, who is a 
>>Semantic Web Activity Lead for the W3C, and we started to discuss the future 
>>of OWL-S and why it seemed that the industry (chiefly commercial interests) 
>>have been slow to adopt semantic web services.  By "slow" we were comparing 
>>how OWL-S does not seem to have the same transition from more of a research/
>>academic initiative to more commercial implemenations as seen with RDF and 
>>OWL.  As such we figured it would be best to open up a discussion as to why 
>>this is, and how to spur the transition as well as to allow people to 
>>comment freely on OWL-S.  Here is a list a questions and statements that 
>>may help jumpstart the conversation:
> 
> 
> As one who is watching more than participating, I can provide one perspective
> on this.
> 
> 
>>This OWL-S standard is still a W3C submission.  Is it still to early to discuss 
>>the viability of OWL-S before it becomes a recommendation?  Perhaps many are still 
>>trying to digest the specifications?
> 
> 
> OWL-S is not a standard.  It is a specification that has been submitted to W3C
> and parts of it are still changing.  OWL and RDF *are* standards which are
> essentially refinements of previous work.  In the case of OWL, the creation
> of a standard by an internationally recognized organization was an important
> catalyst to commercialization and convergence of previous work.
> 
> 
>>Are there not enough concrete examples/documentation for users to follow, and 
>>help expose the benefit of semantically describing a web service?
> 
> 
> There could be more examples, but that's a minor issue.
> 
> 
>>Are the good examples that do exist not given enough publicity, and a convenient 
>>way to search for them?
>>
>>Are there not enough tools to help automate the process of semantically describing 
>>a web service?
> 
> 
> Tools are needed for creating, visualizing, reasoning about, and controlling Semantic
> Web Services.  Some tools have emerged very recently for editing OWL-S based SWS
> specs, but tool support appears to be significantly less mature than OWL and RDF.
> 
> 
>>Are there other standards or emerging technologies that overlap with OWL-S, and 
>>lessen its importance?
> 
> 
> There are quite a few specifications out there for process description, execution,
> and/or planning.  A number of these are standards such as XPDL, BPML, BPE4WS, 
> ebBPSS, BPRI, WMF, and UML2 Action Semantics.  Others are more formal: PSL and 
> SWSL. For me, OWL-S doesn't stand out as much from these as OWL did from conceptual
> modeling languages.  It doesn't seem to have the mindshare, nor does it offer 
> a fairly lossless transition path to comparable expressiveness supported by formal 
> semantics.
> 
> There is another reason for a comparably weak interest level in OWL-S.  Many of
> us who might otherwise be using it are busy with OWL projects! ;/  
> 
> -Evan
> 
> Evan K. Wallace
> Manufacturing Systems Integration Division
> NIST
>

Received on Wednesday, 1 December 2004 23:04:25 UTC