Friday, October 26, 2012

Corticon / Progress Rebuttal


There seems to be a link at at Progess/Corticon  ( ) where I am quoted (page 14, figure 8...) as having said in an InfoWorld article in 2006, "Corticon Plays By Different Rules", (see for the link) and that I conducted certain benchmark tests on Corticon and that they were much faster than other BRMS vendors at that time.  If you go to the link at  you will see that InfoWorld did not publish ANY benchmarks with that particular article.  As a matter of fact, against my objections, the article was slightly re-written to remove reference to performance so that Corticon would not seem to appear so poorly in the eyes of the readers.  At the end of the revised article, our editor posted the following note:  "This review has been changed to qualify statements in the original regarding Corticon 4’s performance and scalability and its suitability for complex applications. The scoring remains unchanged."

However, pertaining to performance, when I first informally reviewed Corticon I did note that Corticon used DETI (DEsign Time Inference) rather than Rete (Ree'-tee), meaning that most of the inferencing process takes time before the timing process begins so that any time to compare would be totally unfair to another benchmark process.  Also, if the number of rows were to be the number of rules, then Corticon used an inordinate number of rules.  Ordinarily Miss Manners uses only 8 rules and Waltz50 uses only 32 rules.  Corticon used WAY more than that (something over 1,000 rows for Waltz50) even though they did solve the problem.

I compare this to running a marathon.  It is much like a runner running the first mile, catching a subway to the 25th mile, then running the last part of the marathon and claiming victory.  I just isn't fair to the other competitors.  Sure, he completed the race but NOT according to the rules.  Sorry, Charlie; no cigar.

And NONE of this showed up on the review in InfoWorld - they are far too nice a magazine to ever publish this kind of thing.  So, as you might notice, we totally cut the part of performance out of the evaluation during both versions of the article.  After all, they were a small start-up company at the time, struggling to compete with the Big Boys of the BRMS world and we just wanted to give them a leg up as it were.  Little did we know that we had grabbed the wrong end of a snake.

So, for the record:

First, I DID write such an article for InfoWorld in 2006 and InfoWorld DID, in fact, publish that article.

Second, I did conduct certain benchmark tests but they were NOT published in any form.

Third, Corticon was not the fastest on Miss Manners benchmark - far from it.

Fourth, InfoWorld did NOT publish any of the benchmarks in the published article.

Fifth, I did not make such a statement nor did I authorize such a publication by either Progress nor by Dr. Mark Allen.

Sixth, I hereby request that Progress and/or Corticon and/or Dr. Mark Allen please remove this article from their website forthwith and without hesitation.

James C. Owen
Senior BRMS Consultant / Architect

No comments: