Think

Mostly on Rulebase Benchmarks, Politics, Religion and other "stuff" that you can't discuss in the workplace

! Java Rules

Greetings:

 

OK, I took the original name from another blog by James Owen who copied it from Daniel Selman.  But since I’m doing more on scripture and ontology than rulebased systems, I thought I needed to change the name.  The link is the same but just what shows up doesn’t lead anyone to think that I’m working full time in the rulebase space any longer.  You can’t do everything so I guess I’ll focus on scripture, some on politics (boooor-ring!) and maybe some on technology when a fever hits.  

 

Enjoy the blog anyway and always feel free to comment.  :-)

SDG

Yaakov

December 21, 2008 Posted by | Rule Stuff | Leave a comment

Bare Knuckle Push-UP

Just had a chat with a friend of mine over in East (almost) Texas.  Fort Worth, actually.  He said that he went to lunch with a couple of friends from the Dallas Rules Group who wanted to start up a course based on the Progression Chart listed on the DRG home page somewhere.  It’s a “kind-of” Sigma-Six thing where you study your whatever off and you get awarded varying degrees of completion that is based on Karate / Tae Kwon Do / Kung Foo kind of progression.  Basically it is a five-year or more plan where the student moves from apprentice to journeyman to master, or in their case white belt (three grades) to yellow, green, blue, purple, three levels of brown belt and finally to black belt.

I’m not too sure that anyone would want to do bare knuckle-pushups to strengthen their hands and wrists and focus in order to better perfect their punch.  Especially when the workout is on a gravel road leading to the sand dunes and then to the sea.  It all seems so Okinawan to me.  (Karate started with Kung Foo teacher from China who taught on Okinawa and the art migrated to Korea, Japan and, eventually, to the USA and Europe.)  Anyway, the Okinawa “masters” would repeatedly punch a 1″x4″ board wrapped in hemp rope, do spear-hand-thrusts into buckets of sand, do knuckle pushups on gravel, wood or something else damaging to human flesh, all in order to enable them to punch through bamboo armor worn by the Japanese soldiers occupying their land.

So, what does all this have to do with his luncheon date?  My friend carried away two things from the meeting.  One of the persons there asked, “Is there any way to compress this time frame from five years down to something shorter?”  He should have just gotten up and walked away at that point but, being a Western Gentleman he gave the lady the benefit of the doubt and simply answered with a comparison of shortening a Ph.D. career – some can, most can’t.  Obviously, she would not have lasted two weeks but he was too PC to tell her that – maybe at some later date.

The other problem was that it had to fit into a corporate environment.  Oh, yeah.  Now there’s a great idea.  Move a Knowledge Engineering school that probably belongs in a Ph.D. program in some high-end university to a corporate environment.  Never happens – no “short term” benefits; only long-term benefits if the students can last that long.  Think about it – most high-tech employees don’t last five years on the job.  What’s to keep them there to complete a five year course of study that will separate them from their family and friends?  

Anyway, like most high-school rock-musicians, programmers eventually have to get a day-job to support their habits of family, friends, nice homes and fancy cars.  The Karate stuff, and the studying, go out the window as soon as they leave college.  No, Jimbo, you’re going to have to keep looking for those few, those silly enough to commit their very lives to the study of knowledge for the sake of just knowing what they are doing.  I hope you have a long and fruitful life but I don’t think it’s long enough to find those Fearless Five in this century.  :-)

 

SDG

Yaakov

July 21, 2008 Posted by | Rule Stuff | Leave a comment

Drools, BRMS and Revolutions

Greetings, Programs:

If your are a member of the Dallas Rules Group (DRG) and you don’t subscribe already to the Drools RSS feed at http://feeds.athico.com/DroolsRSS, maybe you might should since most of “All Things DRG” are Drools-centric. Of particular interest is the “latest and greatest” early editions of version 5 from the Drools Team for two reasons:

(1) The Drools Team is beginning the late-September release of Drools 5 that will include a BRMS implementation. Admittedly, their initial implementation might be just a bit behind that of ILOG or Fair Isaac BUT, and this is the important part, it includes basic BRMS functionality AND (wait for it…. wait… here it comes…..) DROOLS IS FREE.

(2) The Drools team lives on the globe. Literally. Mark Proctor (their fearless leader) is in London. Edson Terelli is in Canada. Mic Neale is in Australia. And the others are scattered around the globe. BUT they work together to produce a product that is making giant leaps and bounds from that “freeware” tag into what used to be the exclusive domain of $20K or $50K starter BRMS tools.

Here’s another thing: More and more companies are using shareware or freeware products such as Linux, Open Office, Java, Eclipse, Net Beans, JBoss to run those $500K (final installed price) BRMS packages. Now they’re going to have another really good package to add in there; Drools. The line between commercial BRMS vendors and freeware vendors is becoming more and more blurred and, with the next year or two, should disappear altogether.

Remember when JBoss was free and big companies would not use it because it did not have “support” from a “real” vendor? So they paid through the nose for BEA (formerly Tuxedo) Web Logic (now owned by Oracle) or IBM Web Sphere. And all of the software ran on either Sun Solaris or IBM AIX. But then “A Funny Thing Happened On the Way to the Forum” – small companies, then bigger companies and bigger companies started using Linux, that Unix work-a-like for Linus Torvalds. Totally unexpected. Fast forward three years and now major companies are not only using Linux, they are beginning to use JBoss. But they continue to use ILOG, Fair Isaac, Pega, Corticon, Visual Rules, etc. BECAUSE the OPSJ, Drools, CLIPS and Jess folks just could not produce a “real” BRMS product.

So, fast forward three more years – companies at that time were beginning to use Jess and Drools in places where they never did before BECAUSE Jess and Drools were using the Eclipse interface and debugging tools and Ant and all of that other stuff that their geeky programmers loved so much and used anyway – until they got caught and reprimanded. (They used it even after that because they could not get that vaunted “tech support” from the commercial vendors.)

Today we, the BRMS and rulebase programmers, rule engineers and knowledge engineers, stand on the cusp of an historical incident much like the Rete Algorithm gave us almost 30 years ago in 1979: A FREE BRMS !!! It’s coming! The first editions are here! That “pre-alpha” part where there are more problems than features. But judging on the history of the Drools Team, those problems will go away. If you are a BRMS consultant, get ready and be prepared. Meaning you had better download Drools and begin getting used to their syntax rather than that of the big commercial vendors.

My prediction is that within the next three years those dinosaurs of the BRMS world will be rapidly slipping into the tar pits of obscurity and we will have copy-cat BRMS freebies. They won’t be as well done as Drools, but we will begin to see more and more BRMS products (note that I did NOT say more and more rulebased products) that are, for the most part, free. Maybe the Drools logo now should be a giant purple dinosaur surrounded by a large red circle with a bar through it. Meaning, STOP THE DINOSAURS. :-)

SDG
Yaakov

July 4, 2008 Posted by | Rule Stuff | , , , , , , , , , , , , , , , | 12 Comments

October [Technical] Rules Fest

The Dallas Rules Group (http://www.dallasrulesgroup.org) is sponsoring a Rulebase Conference in October called, of all things, the “October <Technical> Rules Fest” – maybe it has something to do with the German community in South Texas that has the October Fest.  But October Fest in Munich is actually from September 15th to September 30th.  All very confusing. 

BUT – and here’s the really cool part – this could prove to be the “Woodstock of the Rulebase World”, as one of the “leaders” of the Jess group called it.  Why?  Well, check out http://rulesfest.org and you’ll see that the list of speakers includes almost all of the major rulebase leaders today.  Just to name a couple:  Dr. Charles Forgy [invented Rete and Rete2/Rete III], Mark Proctor [inventor of Drools], Gary Riley [inventor of CLIPS], and many more.  There is even a day of tutorials on Wednesday.  All included for only $150 per attendee.  That’s dang-near free in today’s $5K conference fee level.  Even the El-Cheapo BR Forum is $1,750.  

 Only a couple of world-class God Fathers (sorry ladies, I just don’t know about any God Mothers in the rulebase world) are missing and DRG is trying to get them.  The really, really cool part is that it was the speakers themselves who kept asking, “Well, did you invite so-and-so?”  We invited so-and-so and he would ask, “Did you invite who’s-it?”  And it just grew all by itself.  

Check out the link.  Check out the speakers.  This probably will NEVER happen again – meaning there are not plans right now for an “October <Technical> Rules Fest – Der Zwi”.   OK, so my German is really bad movie-subtitle-level German, but you get the idea.  There may not BE another conference like this ever again.  

SDG

Yaakov

June 10, 2008 Posted by | Rule Stuff | , , , , , , , , , , , , , , , , , , , , , | Leave a comment

What’s Wrong With Rulebased Systems ??

Answer:  Absolutely nothing!  The problem is that if you are a poor schmuck of customer and you don’t have anyone on staff who has been trained in “real world” AI, then you have to believe something of what a salesman / client partner says, right?  Of course, right.  So, during all of the meetings, the client partner tries to find out what you really want, where are your hot buttons, and proceeds to promise you the sun, moon and stars while pressing all of the hot buttons.

OH, did I mention the other poor schumck in this scenario?  The poor guy from professional services who has to implement what the client partner promised.  Can’t be done.  But the PS guys can’t tell you that it can’t be done because he’ll lose his job.  Or get a bad review, which is tantamount to being one of the untouchables class in India.  Or having leprosy.  Something like that.

So, what happens in real life?  The PS guy fools around for about three months, has you send all of your staff off to training (at $5K a pop) at someplace nice (like CA) and then, with nothing more than the rudimentary basic parts up and running, hands if off to your crew to “save you money.”  Guess what comes up in the second verse?  A year later it still isn’t doing what was promised so “another” PS guy arrives after a few visits by the client partner.  

AH!  You (the poor schmuck of a customer) didn’t implement this properly after the original PS crew left.  So, now you get the higher paid guys (the senior KE guys) at a higher price to “show you the shining path.”  After about three or four weeks of this, you own people have to admit that they were wrong and now they have seen the light.  (If they don’t, then apparently they are incredibly obtuse and should be fired at once.)  So, the senior PS guy leaves and your guys have to try (with the help of some Technical Support Staff who were hired last month) to implement what the senior PS guys said. 

By this time, you will either bail out and write it off to a bad experiment or keep going back to the well.  Time after time after time.  Yes, I will have to confess; there are a few, well-done installations.  But most of the ones that I have seen in the past have been abortions that are just covered up and no one EVER admits to failure.  Why?  Because the standards for success or failure were never established in the beginning.  Why?  Because it would take more time and money to do it right than to fail a few times and then just sweep everything under the rug.

OH – what about the guys who started all of this in the first place?  Well, they declared it a screaming success, were promoted within the company to another division and their position now is that it worked fine when they left – you (the poor schmuck who inherited this fiasco) must have screwed something up.  Or they went to work for another company where they will now do the same thing for them that they did for you – and all with your highest recommendations.  Schmuck!!

SDG

Yaakov

May 7, 2008 Posted by | Rule Stuff | | 1 Comment

White Paper Link

KBSC has posted a page of White Papers at http://www.kbsc.com/whitePapers.html that contains, among other really nice stuff, all of Dr. Charles Forgy’s 1979 Thesis – by chapter and one huge zip file of all of it.  Of particular interest might be the Parallel stuff by Forgy / Newell / Gupta as well as Doorenbos’ 1995 thesis that I use to teach the Rete Algorithm.  

 

SDG

Yaakov

April 10, 2008 Posted by | Rule Stuff | Leave a comment

New Rules Group in Dallas

There is a new Dallas Rules Group, http://www.dallasrulesgroup.com and I plan on attending.  It’s a long drive from Sweetwater to Dallas but it will give me an excuse to go over and shop before the meeting.  From what I’ve seen about it, they are focused and organized.  Right now they are teaching Drools but I understand that later they will teach CLIPS/Jess as well as MIKE and maybe some others.  Regardless, if you can make the monthly meeting, it should be worth your while.

 

SDG

Yaakov

April 7, 2008 Posted by | Rule Stuff | , , , , , , , , | Leave a comment

JAB (Just Another Blog)

Just started this blog – more than likely it won’t be devoted exclusively to rulebased systems but about anything technical that strikes my fancy.  Maybe it will be helpful, maybe not.  Later I plan to add a lot of “stuff” to it to help those just getting into rules to discern the difference between normal procedural programming like Java and C++ to declarative programming concept that should be used in a rulebase.  My pet peeve is that concept that a BRMS (Business Rule Management System) is a rulebase on steroids.  Not true!  It’s a dumbed-down rulebase so that regular programmers and business managers can understand it.  Simply put, BRMS is to a rulebase as COBOL is to programming.   Well, that should be enough for the first time out.  I’ll clean all of this up later, add some cool links, start some conversations, etc.  Drop by, leave a comment, let me know what you think. 

SDG

Yaakov 

December 22, 2007 Posted by | Rule Stuff | Leave a comment

   

Follow

Get every new post delivered to your Inbox.