Brian Leathem wrote:
> On 10/12/09 8:10 PM, Reinier Zwitserloot wrote:
>   
>> I always thought Service-Oriented-Architecture was a a devteam take-a-
>> break bullshit word.
>>
>> You know, something you tell the brass so they get off your back for a
>> month or two, giving the team time to spend some much needed time in
>> feature-freeze, cleaning up stuff, refactoring code to be more
>> maintainable for the future, and turning hacked together workarounds
>> into proper fixes. Once you're done, you tell management you have the
>> exact same thing you had before, except it's now all smoother and more
>> maintainable (true), and that it is because "it now has more
>> SOA" (which is false, as SOA is meaningless blather).
>>    
>>     
>
> I think a distinction needs to be made between well 
> modularized/refactored code that is architected to provide services; vs. 
> "SOA".  I interpret "SOA" these days as an architecture that uses things 
> like BPEL and workflow engines to wire various service endpoints 
> together.  I'd say the former is what you are describing, and the latter 
> is what the tool support was providing.
>   
I've heard many definitions for SOA, the only one I like is that you 
look at your business in terms of business processes and then break it 
down into small "services" which all have some kind of agreements around 
them. Where agreements are along the lines of SLAs, i.e. include things 
like availability and response times and not just the input and output 
formats.

The execution can happen with web services and BPEL/workflow enignes, 
but doesn't have to. In fact I have seen someone giving a talk on their 
SOA architecture, which turned to be implemented to some extent in CORBA 
(sic!). But with the above notion of SOA that is perfectly valid.

The problem is that most of the time people try to sell something using 
the "SOA" term. You sell technology, so you bind some technology to that 
term. That process somehow took over and suddenly everyone thinks "web 
service", "WS-*" or even "BPEL" when the term is mentioned. I somehow 
suspect is wasn't originally meant that way.

Which leads to the question: does someone have a reference for the 
origin of the acronym? I once heard that it came out of Gartner, but I 
can't find any proof of that.

  Peter

--

You received this message because you are subscribed to the Google Groups "The 
Java Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.


Reply via email to