VOSibilities

Yeah, baby! Ain’t nuthin’ like a good blog war-o-words. And a juicy one has just broken out between two influential voices: Nick Malik and Bruce Silver. And I suspect we haven’t seen the last of it. (At least I hope we haven’t. July is a slow month; we could use some American Gladiators-style trash talkin’ right about now.)

Apparently, Nick found the top dead center of the button you shouldn’t push in Bruce’s mind: he says BPM is never going to live up to expectations that non-developers will create applications.

In reply, Bruce — slappin’ Nick right upside the head – replies that Nick has to “prove” his assertion by showing that someone — anyone — in the “BPM community” has made a claim that modeling leads directly to completed applications.

While I hope the histrionics continue, this is really nothing more than two purists trying to keep their rivers from converging.  (I gotta admit that I find these near-screaming matches to be more educational than so-called “polite debate” for the very simple reason that they strip out the fluff in favor of direct frontal attacks everyone can understand.)

We all know from long, bitter experience that the “third rail” in the Microsoft world (touch it and die) is developers. MSFT will do what it takes to keep developers tied to the Windows API. Anything that could loosen that death-grip is a danger, and that includes end users working in standards-based tools that could care less about the underlying OS.

And from what I’ve read about the “BPM community” there’s a fair bit of wishful thinking there, too. Bruce is probably correct that no responsible entity has claimed what he believes Nick is claiming. Yet, you don’t have to say the “E” (execution) word outright to lead people to the conclusion that your BPMS does it directly from pretty pictures. Go ahead, spend five minutes on Lombardi’s site and tell me you don’t see it there.

What do we care? Well, let me be the first to pre-announce our upcoming ActiveVOS release, scheduled for mid-August, in which we actually converge the rivers. We will have the most complete BPMN modeling capabilities and, of course, we have the world’s best and most complete BPEL deployment, execution and management system.

ActiveVOS will make it possible for business users to come very, very close to execution via BPMN. And we believe that developers will take that non-executable model and “finish” it in a 100%-standards-based environment that frees them and their businesses from .NetJail.

Forgive me the nested platitude, but the issue boils down to that old saw that says, “Get the right tool for the job.” Developers need modern, standards-based languages that execute on the metal; business analysts need modern, standards-based ways to describe what systems have to accomplish. Being doctrinaire about which is the “correct” way to serve business and IT is beside the point.

So, while it’s fun to see the purists bloody each other, we intend to deliver an implementable, cost-effective and complete way to achieve what neither side really seems to want. And that, dear readers, is what a visual orchestration system is all about.

A fascinating post on the Inside Architecture blog lambastes SOA tools for creating JaBoWS (just another bunch of web services).

Nick Malik writes:

Don’t get me wrong. I don’t hate tools. For one thing, there are some tools that support Enterprise SOA. Not many, but a few. Those tools understand that Enterprise SOA is not about building one service after another, but building the right services, and building them in a manageable and non-overlapping way.

Nick goes on to say that companies that companies that do not implement a “comprehensive Enterprise SOA transformational program” end up with “tripe.”

Have you ever violently agreed with some one’s conclusions but disagreed as violently with the premise? Well, that’s where we find ourselves after reading Nick’s very passionate (and well-written) post.

In short, when you build a SOA up of piece parts, you would tend to believe that service orchestrations — the actual applications — can be built from piece parts. And they just can’t. What’s good for the architecture isn’t good for the developer.

Developers — especially Java types who are creating web services willy-nilly and then running into a wall trying to use them — need something both familiar and holistic to actually get some value from those web services. They need a visual orchestration system which is complete, standards-based and familiar. Something that masks the complexity of long-running transactions, includes human tasks, eliminates hand-coding of XML, offers discovery of available services and, above all, makes testing and deploying services-based apps easy.

ActiveVOS can do all that, and more. Yet some people call us a “tool.” (I’d prefer SOA development system, but at the end of the day, if you are the tripe-buster, what do you care what people call you?)

Nick is exactly correct that the tools people have been using before VOSs yield little but JaBoWS. As long as developers have to put all the pieces together, you can’t get anything else. But there’s magic in making SOA development integrated and familiar.

Don’t blame the tools. Instead blame those who become captives of their own thinking, by extending the assumptions they made in building their SOA to the development of applications to run in that SOA.