This is out of chronological order, but I didn’t have my laptop at lunch so had to reconstruct this from memory and a few scratched notes.
I squeezed my way into the Pegasystems lunch, which was completely full, in order to hear Alan Trefler speak. He was very engaging, with lots of amusing graphics, but one phrase at the end of his talk really grabbed me: “you have to get away from the poisonous import/export environment so that BPM doesn’t become the next CASE”.
What he meant by this is that by modelling in one tool, then exporting it to an execution environment where there is imperfect round-tripping, there’s a danger of having the processes caught in the execution environment where you’ll be stuck maintaining it in a more code-like environment: presumably, the execution environment has imperfect modelling, or you wouldn’t be using another modelling tool in the first place. This makes the modelling tool useless except for the initial design process, and therefore hinders the future agility of the process. He makes the analogy to CASE (think back to the 80’s and 90’s), where nice-looking tools generated code could then be “tweaked”, but you then ended up doing any further code maintenance in the code environment rather than the CASE environment because there wasn’t proper round-tripping between the environments.
This is part of the problem that I have with external modelling tools: unless you can round-trip seamlessly, you don’t have process agility.
Food for thought.
Last year around this time Gartner’s Bill Rosser did a write up which raved about a company called Business Genetics and their modeling methodology – xBML. (They are a sponsor at the Gartner Conference this year.) It has the shortcomings of a modeling only tool that you describe above, creating proprietary models that can’t be imported into a BPMS. I didn’t understand Rosser’s excitement then, and I don’t understand it now. It may be a great methodology, but creating paper requirements and proprietary models seems like a dead end approach. And Lombardi’s Blueprint seems to encroach on their methodology but offers an execution path. Check them out if you get a chance – I’d be interested in your opinion. Nice work from Gartner, by the way. Keep it up! Paul Fisher
Paul, who are you suggesting that I check out — Business Genetics, or Blueprint? I’ve already blogged about Blueprint a few weeks back after an extensive demo by the product manager, and I’m waiting for my beta account.
I had an interesting discussion about models-to-execution with someone from Proforma yesterday before I left, and I still think that not enough is being done to make it dead simple to model in one tool and execute in another, with round-tripping.
Sorry about that – Business Genetics. But seeming them will probably have to wait for the next Gartner Conference:>)
I probably shouldn’t mention I got a Blueprint beta account from Lomardi – just starting to play around with it. Maybe I can invite you in for collaborative modeling!