Process 2006 Day 1 – Networking

I don’t usually blog about my casual lunchtime conversations at conferences, but there were a couple of interesting comments and conversations that came up.

Funniest quote of the day, so far, was when Derek Miers referred to one of the speakers as being a “bullshitter” (you have to know how deep that Derek can shovel it to truly appreciate how funny that is :) ).

Another completely inspired comment was when Peter Fingar (I think) said “SOA is just Simula over IP”, and Keith Harrison-Broninski the other person who I was standing with (help me out, I know that you read this) said “no, SOA is just BASIC over IP”. Geek humour, I know, but I was ROTFL.

I had a couple of conversations about my recent rant about over-customizing BPMS’, and might even have a few converts to that point of view. I had this chat with Keith Harrison-Broninski and he felt (as he said in his presentation earlier) that BPMS’ are not really necessary but could be replaced by a modelling/analysis tool that could translate directly to J2EE/.Net components. My view on this is that if you’re over-customizing your BPMS, then you might as well just build your own, but that there’s a huge potential benefit to using a BPMS if you use it the way that it was intended.

Keith and I went off on a tangent about how much more that people are going to have to know about technology in order to do their job as things progress, and I made the analogy to driving a car now versus 80 years ago. At first, he disagreed, stating that you need to know less about a car now in order to drive it (for example, when’s the last time that you had to find your spark plugs and dry them off to get your car started?), but I eventually won him over to the idea that cars are much more complex now, mostly due to the auxiliary systems such as GPS, stereos, lights and windshield wipers. To make it worse, I don’t own a car but rent when I need one, which means that every time I drive, I have to use my metamodel of automobile functionality to be able to get into a completely unknown car and drive it away within 30 seconds. “Metamodel of automobile functionality”, he he.

8 thoughts on “Process 2006 Day 1 – Networking

  1. Sandy, I absolutely agree with your rant about over-customizing BPM – I must have missed you blogging about it though. My experience with this has been around the UI, where rules and process start (inadvertantly?) getting embedded, breaking the ability to alter the BPMS representation of the process.

    I’m not sure if this lines up with your view of over-customization. Are you thinking that parts of the process end up written in code, or something else?. That would then seem to indicate a progression to Keith Harrison-Broninski’s approach of converting directly to deployed app server components.

  2. Phil, my rant was only a few days ago, <a href=”http://www.column2.com/2006/09/bpm-and-web-20/” rel=”nofollow”>here</a>. I completely agree with your post from June, I see way too many cases of custom code built on top of a perfectly good BPMS, turning into yet another piece of legacy code. Your insurance company example is almost identical to one that I was involved in recently, where the systems integrators completely hid the BPMS behind custom code, made it impossible for the business analysts to use the BPMS design, simulation, administration or monitoring tools, thereby rendering it completely un-agile. Get something simple in early, I say.

  3. OK, I own up (knowing Sandy, the truth will out sooner or later anyway). The “SOA is just BASIC over IP” crack was mine.

    This is a serious point, believe it or not. I have come to realise more and more over the last few years that most people working at the intersection of business and technology (i.e., with techniques such as SOA and BPM) don’t “get” object-orientation. I’ll blog about this soon – about time I got back to the blog anyway!

  4. Pingback: Tyner Blain
  5. Pingback: IT Directions

Leave a Reply