Friday, September 18, 2009

Day Two - FICO Decision Management Tools User Group


Wow. Yesterday was good but today was even better. One interesting this is that Oracle was presenting on an idea and their presenter did not know that Oracle had bought a rulebase (two of them, in fact: Rule Burst & Haley Expert Systems) of their own that is a direct competitor to FICO's Blaze Advisor. Weird, really weird.

OK, performance issues - I ran through a brief (55 minute) presentation on how to improve performance. More of how NOT to get into trouble although I did point out some time-proven tips on how to make even Advisor run faster.

Then came the Illinois Death March: We took a bus to the lake for a river cruise. (whater...) The bus dropped us off and we walked past ship after ship after ship. Now, understand, walking a mile is not such a big deal UNLESS you have a bad back, two knees that need to be replaced and pinched nerves in your neck. Then (after a rather nice meal onboard the ship) we had to walk BACK! So, thank the Lord for Hydrocodone and Soma. Between the two I made it through the night.

So, wrapping up today. Tonight begins Rosh Hoshanna so a lot of the guys have to get home in time for services. People are here from all over the world; Russia, Lithuania, Israel, France and even one or two from Texas. :-) So, this will be the last report on this conference. Nothing technical. You'll have to get that from FICO since everything is copyrighted into their vault of knowledge.


Thursday, September 17, 2009

Day One - FICO Decision Management Tools User Group


First day started at 13:00 but I couldn't get through traffic and arrived at 15:00. I went to the Decision Support clinic headed by Don Grice and Maarten Van Lier that started at 13:00 or so and I was so far behind that I'll have to catch up later.

The hotel is nice but the 17:30 cheese-and-wine was awesome. Lots of brisket, pasta bar, chips-and-dips, beer/wine/etc. bars made it nice. But with my bad back I could only stand up for so long. Lots of old friends there: Tom Travis, Vlad Silverman, Carole Ann Berlioz-Matignon, Carlos Seranno-Morales (Europeans are really big on double last names). Willie Hall and Mark Eastwood are supposed to arrive tomorrow.

The conference itself seems focused on business users rather than geeks. (So what am I doing here?) Several good talks lined up for today including some on Blaze Advisor and business optimization tools. My talk will be on optimizing Blaze Advisor at 16:00 today and then FICO has a river-boat tour and supper planned. I'm not sure if the PPTs or talks will be available to the general public but if they are I'll let you know. Maarten's talk last night was pretty good so I hope that they do release them to the great, unwashed masses. :-)


Monday, September 14, 2009

MIND Over Matter: If You Don't Mind Then It Doesn't Matter


If you don't already subscribe to the Mind magazine, may I suggest that you consider it. Drop by your local magazine store or book store and take a look at what is on the shelf and then save a few coins by subscribing to the magazine on an annual basis. Why? OK, here's a couple of thoughts that I got from one of the more recent issues:

Do you remember the last (or next to last) session of M.A.S.H. on TV? It seems that there were some villagers trying to hide from some enemy soldiers and, if caught, all would be killed. One of the young mothers had a baby with her and the baby, as they will do from time to time, was scared and began to cry. The rest of the villagers kept telling her to be quiet or they all would die. The mother ended up suffocating the baby in her attempt to keep it quiet. Questions: Was she right to kill her child to keep the other villagers from being found out and not only killed but probably tortured as well? What did this have to do with MIND magazine?

It seems that whenever the mind is presented with a conundrum such as this that certain portions of the brain are activated. In younger minds, they are, basically, stored in one location. In older minds, the job is distributed out to more than one part of the brain. What I think is happening is that older minds have to consider more than one answer and that the outcomes are different depending on which solution is implemented and when it is done and how it is done. This comes all the way back around to rulebased systems: It's the Senior KE, the Senior Knowledge Engineer, the Rule Architect that will make a huge difference in the overall outcome.

Why? Because that Senior KE has seen the pitfalls of what happens when certain logical paths are followed and has experienced the outcomes of failures in the past. What was it that George Satayana said? "Those who will not learn from history are forced to repeat it." (Often attributed to Ben Franklin but supposedly he got it from George...) We learn from our mistakes. The more intelligent ones learn from the mistakes of others. It seems that some never learn but have keep making the same mistake over and over and over. My favorite is the company that hears about off-shoring work at really, really cheap rates. What they don't hear about are all of the problems connected with different cultures, different languages, different work habits, the advantage of face-to-fact meetings, etc.

The other one is when a company refuses to pay the higher rates of a senior consultant and hires a programmer who is really, really good at Java or C/C++ and assigns them to do the work that truly needs a Senior KE to direct the effort. the project goes down the porcelain receptacle reserved for such sludge and management brazenly decides that a rulebased solution was not needed for a project such as this. No one really loses their jobs because the smart ones bailed out after the first few months and went on the bigger and better things. Those stuck with the smelly thing can then blame those who started it. Those who started it blame those who followed it with not carrying out their particular vision and, of course, it went bad.

Now, wasn't that fun? And I'm sure that most of the senior guys have been on the Project From Hell, Death March, etc. By whatever name, it wasn't much fun if you were on the final march and, as consultants called in to save such a fiasco, it's usually impossible with what is left of the budget. But, once in a while, it is possible and what remains of management is so desperate that they actually let you do the right thing. And you get to be the hero. One or two of those go a long, long way toward making up for all of the fiascos foisted upon you in the past.

So, what is the lesson to be learned here? Take care and be wary of what you accept when management is willing to pay a higher rate plus expenses. There HAS to be a catch somewhere and I'll bet it's because some ninny has already messed up the playground.

Sunday, September 13, 2009

More and more good news on ORF 2009


First, some GREAT news! We have re-arranged the Agenda so that we could accommodate a "Thursday Afternoon Think Tank on Needful Things." This would be a round table discussion (with some attendee participation) on various subjects that concern the more technical aspects of rulebased systems rather than their direct application to problems per se. Those who have agreed to be on the panel are Dr. Charles Forgy, Mark Proctor, Gary Riley, Dr. Jacob Feldman, Dr. Richard Hicks, Carole Ann Berlioz-Matignon, Carlos Seranno-Morales, Jason Morris and Paul Vincent. For example, a few of the topics will (might) be Complex Event Processing, Benchmarks in This Century, Patterns in Rulebased Systems, Finite State Machines as a Special Type of Rules, etc. This is focused more on the "hard core" developers and "thought leaders" in the rulebase field. However, ALL attendees and/or speakers are invited to attend and be part of the direction(s) that rulebased systems might take for the next decade.

Another change is that there will be another Drools Boot Camp (like last year) headed up by Mark Proctor. This will be co-located with October Rules Fest in The Adolphus hotel and will be all day SUNDAY and MONDY morning. Since Mark and Edson will be there anyway this was a great opportunity for another in a great series of DBCs that Mark has done in the past. The DBC will be absolutely free for ORF attendees but there will be a charge if you are attending DBC only so that we can pay for the room, coffee, etc.

Finally, the conference proper will start on MONDAY AFTERNOON rather than Tuesday morning. The lead-off speaker will be Lawrence Terrill on a really good explanation of the Rete Algorithm - one of the best that I have ever seen.

I've been reading through some of the early submissions of the White Papers for October Rules Fest and they are really outstanding! Dr. Charles Forgy's paper on parallel rulebased systems (need and solutions) is fantastic. As is that of Gary Riley on optimizing performance in CLIPS. Andrew Waterman submitted one on games and is looking for folks to submit more add-ons. Dr. Richard Hicks has submitted one on Validation and Verification that everyone should read. (He is giving each attendee a free copy of the game.) Dr. Daniel Levine's paper explores the inner-dimension of the brain itself, its own self-imposed rules and problems with "rational thought process." Remember, all of these will be available at no charge only to attendees at ORF 2009.

Anyway, you need to sign up (if you haven't already done so) and be sure to reserve a hotel room. We still have a few places left.