A few thoughtlets

[Don Flinn <flinn@alum.mit.edu>]

Some thoughts which may or may not be useful.

As we all know, the scientific method proceeds somewhat as follows:
1. Observe the real world in some domain
2. Form axioms, which we assume to be true and valid, about that domain
3. Use these axioms to form a mental structure or hypothesis
4. Addition axioms, which may be extracted from former axioms, may be 
needed to complete the mental structure
5. When done, we have a theory
6. Test the theory to see if it fits the real world and meets the goals 
of the theory
7. If it doesn't modify the theory
8. No amount of experimentation can prove a theory correct, but one 
experiment can prove it wrong (Einstein)

Our reference architecture has proceeded to step 5, i.e. we have 
developed a theory about how an SOA system can or should be developed 
and used.  Testing the theory will involve its being used to develop 
concrete architectures and then those being used to develop SOA systems; 
this theory's ultimate goal.

The more abstract the theory or model the more real world situations it 
will cover and the more useful it should be. 

Our Reference architecture is a darn good abstract theory.  It will be 
up to the experimentalist to test the theory by developing concrete 
architectures and system to see if it holds true and to find areas where 
it may need modification. So the picture is not complete (although it 
will never be absolutely complete re: item 8) until SOA systems are 
built and used based on this reference architecture.  Further, unless 
experiments are based on a solid theory they are just shots in the 
dark.  Note that some of these shots may work following the hypothesis 
of an infinite number of monkeys at keyboards.

There seems to be some misunderstanding by some in the outside world as 
to what this reference architecture is and should be used for.  Would it 
be helpful for users to understand the purpose of this reference 
architecture to include a sentence or two, such as the following, to 
section 1.1.1?  "From a scientific process point of view, this Reference 
Architecture is an abstract theory that is to be used to develop 
concrete architectures and from those SOA systems.  As with any good 
theory it must be tested to complete the process,  which in this case is 
the aforementioned development of concrete architectures and SOA systems 
based on this theory."

Don

-- 
Don Flinn
President
Mansurus LLC
e-mail: flinn@alum.mit.edu
Tel: 781-856-7230
http://mansurus.com