Saturday, April 08, 2006
Community Centric Service Methodology Glass Bead Games
Communication from Paul S Prueitt to the Federal (US)
CIO Council’s SOA CoP forum
Over the past two months, there has been a vetting of the issues, in a most complete fashion in the CIO Council's SOA CoP e-forum.
http://colab.cim3.net/cgi-bin/wiki.pl?AnnouncementofSOACoP
The core assumption that we, forum members, make is that the marketplace already has SOA deployment examples, from which the SOA CoP would like to stand up a transparent and simple example, so as to make clear SOA (Service Oriented Architecture) principles.
But in real life, SOA principles vary and the variations lead into technology dependant implementations. Once dependant on a specific technology/standard or a specific specification then it is often the case that government interest in a specific technology/standard will have, de facto, selected a winner. Historical evidence may be that RDF/OWL was so selected over Topic Maps (in the period 1999 - 2001).
Our forum (SOA CoP) has properly vetted the issue of fairness to alternatives. But there is no action to support Topic Maps in any way in the US.
The participants of the SOA CoP forum understand our core assumption about (US federal) government interest in a common, simple example/demonstration of SOA principles. There is a recommendation (copied below) that an agreement be made that the simplest SOA be specified so that that specification can be demonstrated/simulated. We second that recommendation.
A demo should be developed in accordance with the suggestions as having the greatest understanding of these active in this discussion, with valued input from others.
Additional
recommendation:
I suggest, on behalf of others, and myself that a SOA CoP subcommittee also be formed with the mission to map out the possible future developments in the SOA and ontology mediated SOA domains.
This subcommittee should have a separate forum and should be futures oriented. Topic Map standard should be re-examined as part of the ontology description, and the TM standard should be used rather than the OWL (W3C) standard for ontology modeling.
The "futures of SOA" forum may wish to focus on "beyond the horizon" efforts. We propose that this effort would be supported by a public/private wiki.
We propose that the Federal (US) CIO Council provide (1) a wiki, (2) visibility to the forum, (3) exposure to results (presentations) developed by forum groups; so that we might be able to develop a forward looking exposure of what is "next".
Valid topics would be
1) SOA implementation methodology
2) compatibility between standards supporting SOA
3) XML acceleration techniques, marshaling and un-marshaling techniques
4) ontology interface to orchestration and service discovery
5) community and individual visualization of conceptualization of service webs
In a standard e-forum, such as this one (SOA CoP), it is difficult to preserve the knowledge exchanged by individuals, but the wiki resource could do this.
Several members, of the SOA CoP forum, have been working on a wiki architecture that starts out by seeding a "shell" wiki with a set of phrases/terms and then allowing members to make modifications to the information on each page. Alternative viewpoints could be exposed along with the mainstream viewpoints. A core team would have editing capability until the wiki is stable, and then the wiki is made available for open editing.
A spec on this wiki based conceptualization of a domain of discourse is being prepared.
f we can agree to the OASIS BCM as the foundational methodology standard, this would be helpful. But comments and viewpoints regarding BCM should be asked for and discussed.
www.businesscentricmethodology.com
The OASIS Technical Committee working on a SOA methodology (Business Centric Methodology) produced a model having four layers, the bottom one being "conceptualization".
A CIO Council sponsored "futures of SOA" wiki could have as its mission the development of the community conceptualizations about SOA, now and into the future.
This may be done with less effort than in producing a demo, and a knowledge management system based on a wiki may be found useful to the design community.
When done in parallel to producing a demo would allow the community (the federal space in particular) to see the demo and to also see the first part of a methodology guiding
1) Conceptualization
2) The formation of a common substrate for description of
"services" within a community
3) The issues related to "extension" from some existing
"service web" to new or other "service webs"
4) The (finally) informed implementation efforts needed for extending an
evolving and dynamic "service web"
5) Establishing for the first time membership
So in summary:
We do not feel a need to question the specifics of any demo that the active participants of the SOA CoP forum wish to define. The issues have been fully vetted. However, we note the absence of Topic Maps and a community process based on deep design.
We are proposing that the CIO Council sponsor and give exposure to a "futures of SOA: eforum AND wiki.