[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Will The Real SOA Please Sit Down?

• To: XML DEV <xml-dev@l...>
• Subject: Re: Will The Real SOA Please Sit Down?
• From: "W. E. Perry" <wperry@f...>
• Date: Tue, 10 Jan 2006 11:52:14 -0500
• Cc: "Bullard, Claude L (Len)" <len.bullard@i...>
• Organization: Fiduciary Automation
• References: <15725CF6AFE2F34DB8A5B4770B7334EE0BB1FD54@h...>

If SOA means anything particular (and I believe that it does) then it is that
architectural practice which separates the design of individual processes, or
services, from any specific aggregation or framework in which those processes
might be concatenated or otherwise cooperate. SOA is most emphatically not
about the design of the processes themselves:  it is 'service-oriented'
precisely because the services are pre-existing, or at least they are designed
with the primary goal of expressing a particular expertise in processing, and
significantly without reference to the form in which they might be aggregated.
Standards for SOA are therefore standards for the aggregation of processes not
built to those standards. The creation of any particular SOA framework should
not even contemplate the redesign or rewriting of the processes which that
framework might invoke:  those processes should be chosen for their specific,
idiosyncratic expertise, which should never be vitiated by reducing that
expertise to some common denominator of  a framework which is entirely alien to
the expertise for which that process was designed and built.

Respectfully,

Walter Perry

• References:
  • Will The Real SOA Please Sit Down?
    • From: "Bullard, Claude L (Len)" <len.bullard@i...>

• Prev by Date: RE: Will The Real SOA Please Sit Down?
• Next by Date: Re: Will The Real SOA Please Sit Down?
• Previous by thread: Re: Will The Real SOA Please Sit Down?
• Next by thread: Offtopic: IBM Steps Up To Plate To Help USPTO
• Index(es):
  • Date
  • Thread