[4]                             home                             [6]

 

Monday, April 10, 2006

Anticipatory Web discussionà [home]

Anticipatory Web Challenge Problem à [link]

SOA CoP wiki

 

 

Community Centric Service Methodology Glass Bead Games

 

"SOA does not stand alone, nor can it be managed that way.  If treated in isolation, Web services will only exacerbate organizational complexity.  Any attempt to manage Web services in isolation is a grave mistake. From (URL)

 

 

Communication to SOA CoP eforum

 

 

The OASIS SOA-RM (reference model) is only 28 pages and really does a nice job of laying out an "abstract framework for understanding significant entities and relationships between them within a service oriented environment."

 

The demo for simple SOA transactions is being developed, but that project will soon come to a completion.  At that point it would be nice to have a forum where the participants were committed to the (often perplexing) discussion about the context of SOA.

 

One view of this context is given in Section 2 of the SOA-RM...  The SOA-RM is very interesting work and has nice use of language. 

 

However, it is said:

 

"Visibility, interaction and effect are key concepts for describing the SOA paradigm." 

 

Of course, this is a true statement for any SOA paradigm, from very simple, seller, broker, buyer SOA (as in the demo for the April 23 - 24 meeting in DC) to complex SOA deployments such as those described in the TopQuadrant presentation.

 

 

 

 

The "however" is a lead into how semantic extraction techniques might be applied to "measuring" the transactional space where the "services" are occurring.

 

Suppose that one is able to measure the communicative acts between those business partners involved in something like a response to the bird flu threat? 

 

There is a knowledge elicitation process involved in a simple SOA demo, or in the types of SOA deployments being attempted by some industries (some parts of the manufacturing industry - for example).  This elicitation process is generally under the control of software designers or architects.

 

But if one moves to a more complex setting, the "origin of design" needs to shift to the stakeholders.  In the case of the national response to the bird flu threat, these stakeholders are more likely than not medical doctors and research scientists who do not feel comfortable with the simple SOA frameworks.   and for good reason.

 

How do I make visible my intuitions about what I might want to do research on once the virus make the jump to human to human transmission?  BioPAX (OWL Full) experiences with bioinformatics is a good way to study the problems of

 

1) elicitation of deep knowledge

2) human induction of information structure

 

but requiring visibility and structured interaction between the core inductive processes involved in crisis management may not be a perfect solution.

 

"Measurement" is an alternative to

"visibility over hard structured interaction".

 

Measurement of a transaction space, like the social discourse, what semantic extraction techniques have been developed for... and some of these techniques involve linguistic techniques ... and some involve types of hybrid techniques involving RDF type ontology or n-ary ontology.

 

Joe, I am pretty sure that the work at DERI and else where on the SEE (Semantic Extraction Environment) specification does not address this alternative... but rather continues to strengthen the standard approach to web services - but with notions that SEE's notions about machine orchestration and service discovery is sufficient in all cases.  The specification is currently long and difficult and strongly misuses the term "semantic", does not even use the term pragmatic... 

 

The notion of choice point is not present (unless I missed it).   So the basic concepts of the development of more than one blueprint for human decision makers (not IT professional) is simply missing in the work being done on SEE (again - please correct me if I am not getting this correct).   Mis-impressions about the nature and reality of "inference" based on description logic is ubiquitous in the preliminary work.  (Again please correct me if I am incorrect.)

 

The draft spec can not be found at

 

http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=semantic-ex

 

but I have a private copy....  for those interested.

 

the following are related specifications

 

WSML

WSMX

 

which can be accessed via

 

http://www.wsmo.org/

 

I feel that the WSMO work is important and valuable - but the basic notions expressed in the SOA-RM might not be addressable with the approach taken by WSMO participants. 

 

I could change my mind, however.