Thursday, November 9, 2017

Graham Glass

Greetings, Programs:

[dateline 171109:1500CST]
Just thinking about Graham today.  Don't know why...  I was just sitting in my LaZyBoy this morning drinking my morning coffee, having my usual morning Einstein Blueberry Bagel that my Spousal Unit picks up for me every Tuesday on her way to the Irving Art Association and he crossed my mind.  I remembered that either he drove me home or I drove him home after a JavaMUG meeting at the Sun offices one Wednesday evening.  [I "think" that the meting might have been about Voyager that evening - maybe not - I don't really remember.  Maybe GreG can correct me one this one.] All that I can remember of that evening was that we had a remarkable chat on the way home about rulebased systems and how they might fit into what he was thinking about doing with a future project.  (I don't think he ever got around to doing anything with rulebased systems nor AI nor the future project for that matter.)

All of that is neither here nor there but I did go look up where he is now and what happened to him after that evening.  At that time he was still with Object Space.  It seems, after reading his bio, that what he should have done back then was sell out his shares of Object Space after his dispute with the board (if he could have done so) and gone his on wonderful way.  He has a remarkable mind and could have done anything with the proper capitalization and good business team behind him, much like many other high-IQ techies that I have been fortunate to meet in my long and varied career to date.

Bottom line: Whatever you are doing, if you are not enjoying it then you are either doing it wrong or you are doing the wrong thing.  [What an old cliche'!] Find something that you truly enjoy doing and do it to the the very best of your ability, even it that something is just fly fishing, brewing beer, making wine, crafts, painting, wood working, programming (whether rulebase stuff or security or special applications or GUI stuff - whatever) or biking or whatever.  Find the best parts, even the "commercial" parts, and build a business out of it that exploits the absolute best - the real craftsman part.  If it is teaching then find out what part of teaching really reaches others.  If it is the political spectrum, then be the best, non-corrupt politician in the country, NOT what exists in today's culture!

I cannot teach children nor can I teach high school students - and rarely college students.  But I can teach those who want to learn rulebase systems (RBS) so that is something that I love.  Unfortunately, a lot of folks say that they want to learn RBS but it is only so that they can charge the Fortune 500 companies lots of money for that knowledge.  Those folks I cannot teach. They want the QuAD (Quick And Dirty) way to learn and I have not found that way yet.

Most BRMS companies try to teach a one-week school followed by another one-week school a year later but I have found that those who have followed that path in those schools have learned a monkey-see, monkey-do kind of programming for RBS, not anything of any kind of depth.  When I am working on-site and I ask them to actually read a book outside of class or outside of the working environment, they look at me as though I have lost my mind. Or they take the book home and never read it and I get it back when the gig is up, untouched and unread.

Like I said back on the first of the month - these days I am writing on almost anything.  Graham memories got me started on this so I can blame him for it.  :)  So, from GG to RBS to QuAD teaching methods.  Have a Happy Whatever you are having this week. 

Shalom!
jco

Wednesday, November 1, 2017

Rete Goes GNU

Greetings Programs:

Yepper, you heard (OK, read) that correctly: The Forgenator himself, Dr. Charles L. Forgy is, as of today, releasing the Rete-NT Algorithm as GNU source code to the world.  It is totally free  BUT you cannot make changes to the code and then try to keep your changes proprietary(See the link to the licenses below.)


An article (soon to appear in the on-line version of InfoWorld per my over-worked and under-paid editor) should confirm that Dr. Charles L. Forgy, the original inventor of Rete (pronounced Ree'-tee in our industry regardless of how much Latin you had in high school or college) has released his invention into the wild.  Not just the original Rete but Rete-NT, the latest and greatest!  Yes, you read that right!  Rete-NT is now available for download from PST on the GNU license.  (Go to https://en.wikipedia.org/wiki/Comparison_of_free_and_open-source_software_licenses for a comparison of all of these license.) 

This is basically a R(E)evolution in the BRMS / Rulebase World.  Are you old enough to remember where you were when John Kennedy was assassinated in 1963?  Or where you were when Neil Armstrong first walked on the moon in 1969?  Or Gene Cernan last walked on the moon in December of 1972?  (Most folk don’t remember Gene Cernan!)  Actually, there are a few of us still around who remember the news that Pearl Harbor had just been bombed in 1941 or that we had dropped the first (and last) war-time atomic bomb in 1944.

OK - back to the technical side; have you been around long enough to remember the advent of the Rete Algorithm way back in the 1960’s?  Or the coining of the term “AI” at Dartmouth College in 1954?  What about the introduction of Java back in 1996 or 1997?  (It was kind of vague back then.)  No?  Well, try to remember where you were today when you heard that the Rete-NT Algorithm was released as GNU code to the world by Dr. Charles L. Forgy via KBSC.

Why?  Well, I asked him the same question on a Tuesday afternoon at lunch in Dallas at a local Bar-B-Que.  His answer: “It’s about time.”  For now, the Rete-NT algorithm will be released via the GNU license, meaning that you can use it at almost any academic or non-commercial venture AS WELL as for any commercial purposes.  Sounds fair to me. You can also use it commercially but this should be negotiated with PST, Dr. Forgy's company based in Pittsburgh, PA. For downloading and the OPSJ 8 manual and/or code, please contact/email Dr. Forgy at cforgy@pst.com for more information with your name, company/university name and phone number.

Now, for those who cannot remember the history of AI and want everything compressed, I shall try and compress the history into as brief a passage as possible that includes all of the contributions of Dr. Forgy et al made along the way.  Back in August 31, 1955, there was a short (2 month, 10-man) study of AI at Dartmouth College by such notables as John McCarthy (Dartmouth College), Marvin L. Minsky (Harvard University), Nathaniel Rochester (I.B.M. Corporation) and C. E. Shannon (Bell Telephone Labs) to study how computers could be programmed to use languages to manipulate words as human thought processes, neuron networks, abstraction, randomness and creativity and other “original” ideas.  They “coined” the term “Artificial Intelligence” – or “AI”.  Drs. Newell and Simon also introduced their Logic Theorist Program which later became the GPS, General Problem Solver – which were really advanced programs for that period.

Then, in 1958, a stunned USA realized that the USSR was ahead of them in the “space race” and established ARPA, or Advanced Research Projects Agency which became D(Defense) ARPA.  And, yes, ARPA/DARPA was the origin of the internet, NOT VP Al Gore, when they contracted with BBN Technologies to build the first routers in 1969.  DARPA also was highly involved with Stanford, MIT, Boston University and CMU in Rulebased Systems where a young Charles L. Forgy was working on his Ph. D. with Dr. Alan Newell, one of the founders of AI.  During that process, Charles (later, Dr. Charles Forgy) was tasked with optimizing the process of running the rules because it was taking DAYS to run simple rulebase tasks, even using the Symbolics LISP machines. He came up with a method of swapping memory space for optimizing processing time using a network (ergo, "rete" meaning "network") of objects.  You can read about it in his dissertation or many other simpler but not as detailed explanations on the web. (eg, Robert Doorenbos Thesis)

More later as Dr. Forgy makes it available.  Stayed tuned for more earth-shattering news from KBSC as it becomes available.  :)

jco
(c) KBSC 2017