It’s been an interesting 24 hours, and if you’d like to catch our version of Twitter, go to James Taylor’s (the IT analyst, and brother of the opera singer, not the singer/songwriter) excellent play-by-play coverage of Ilog’s Dialog 2008. Taylor, myself, and several others sat down for a lunchtime roundtable with IBM’s Sandy Carter, plus an encore performance from Ilog CEO Pierre Haren.
At the start of the discussion, the question of whether SOA is old wine in new bottles came up. (We’ll be raising that issue again tomorrow as we moderate back-to-back sessions on IT Industry trends and SOA.) Carter claimed that, compared to earlier approaches like CASE, rapid application development, object-oriented or component-based design, that SOA is more evolutionary than revolutionary – and besides, there’s a real body of industry standards this time. Haren interjected, “I’ve never seen an object,” and that “service orientation” was the key. As if anticipating the line, Carter added that she’s encountering many customers who want their companies to be service-oriented; in case you’re wondering, IBM’s been holding a number of SOA Executive Summits to press that very point.
“We don’t talk the “A” word [of SOA] anymore, we talk process orientation,” Carter said. That was kind of an interesting statement, as consultants like David Linthicum have long railed that SOA is an enterprise architecture pattern – a point that yours truly, Dana Gardner, and Todd Biske had a lot of fun with at last summer’s Open Group EA conference in Austin. In other words, lets start painting SOA as less enterprise architecture, but more business process management.
Maybe that’s because, as our colleague Joe McKendrick just pointed out, that there’s a perception that SOA is a long journey. The not so-hidden question: has the business run out of patience with SOA as a journey? Yet another question we hope to pose tomorrow.
As this is a conference about business rules management, the topic waded back into focus when discussion shifted over the looming generation gap, and with impending retirement of the baby boomers, the fact that a collective amnesia is about to settle over the corporate world if we don’t take action now to imbed those rules and processes before they literally walk out the door. It’s a thought that almost makes one pine for the simpler days of Y2K, when at least there were a few aging geezers still alive to patch up those 30-year general ledger systems.
That’s where Haren pointed to a concept he raised during his keynote this morning, that having good data is not good enough. Not that there was any offense meant against the data quality vendors, but you can have conflicting pieces of data where both are technically right. He pointed to a customer, which had an SAP system that had a transportation management system that numbered trucking lanes starting with the number zero, while another package was designed to reject integer values less than one. Taylor added that the problem gets compounded when you implement master data management, at which point you’d better have a data validation rule before you literally broadcast out what is supposed to be the golden copy.
That’s the point where Haren referred to what we’d call the wisdom life cycle. That is, you have apps that convert data to information – for instance, you total up sales and you have information on how well sales performed during a given period. It becomes knowledge when the information is absorbed by sales and marketing groups, and becomes wisdom when those same groups gain the insight on what’s likely to sell when the weather report predicts 60 degrees F and sunny on the weekend before Christmas. (Hint: you probably won’t be selling many overcoats.)