Responses to SOA and the Reality of Reuse
Friday, September 29, 2006
I've been surprised by the responses I've seen to my latest Opinari
. I expected more people to disagree with me, defending the promise of service reuse. What I've seen instead is general (and sometimes enthusiastic
) agreement. Even the comments on Joe McKendrick's ZDNet entry
were largely in agreement, although I'm not sure Joe himself shares my perspective.
Microsoft's Harry Pierson wrote an interesting response
on the importance of context in reuse, something I fully agree with. His observation that reuse with services might well be more difficult to achieve than it was with objects also rings true, since the requirement for remote access to services does raise the bar. David Ing also wrote a useful response
, raising some excellent points about why reuse is valuable primarily for services that expose data.
For a view from the other side, I had a chance to talk recently with the author Thomas Erl
. We both presented at Gartner's Application Development Summit this week, and Thomas is significantly more bullish on the potential for reuse. As I understand it, his view (based on experience gained from his firm's consulting work) is that if an organization does an appropriate top-down analysis of a well-defined business domain, it's quite possible to discover and implement reusable services. This is a good argument, and I'm willing to believe that there are cases where this kind of analysis-led reuse can work.
What's harder for me to believe is that a majority of organizations will be able to do this. Instead, I've come to expect most efforts to take a more technically oriented bottom-up approach. Given that the coming wave of vendor platforms will all but force us to build service-oriented applications, it will be hard to avoid at least this much SOA.
Will this simpler, less ambitious approach offer all of the benefits claimed for SOA? Certainly not. But is it an improvement over the world we live in today? Just as certainly, the answer is yes. Like objects, services won't provide everything promised by their most fervent exponents. Nonetheless, this shift counts as a step forward.