Kicking around the vendor booths, I made it to every single purveyor of
business rules engines present at the conference. I tried to dig a
little bit on how each one was unique and how they differentiated
themselves from the other guys. In this post I’d like to give some of
what stuck out and was noteworthy.

• Yes, the big two are still ILog and FairIsaac, but there were a number of other speciallty players worth a look.
• If you wanted a solution that handled both business rules and business processes, there were two players at the conference: RulesPower and Pega.
  For my money, this is where the action is; business rules only really
  make sense in the context of business processes. Rules are the sail, if
  you will, and processes are the wind. So here were two companies that
  provided middleware that allowed you to integrate all your systems into
  flows governed by rules. More on this later.
• If you stepped up to the InRule
  booth, they would tell you how they were the only true .NET solution.
  That’s not terribly interesting, since all the major players either
  already have or are about to release .NET platform versions of their
  products. No, InRule’s real selling point, in my opinion, is their
  strength in quantitative and statistical matters. If you are looking at
  doing a very basic business rules apps based on simple predicates,
  etc., you may be better off playing it safe with one of the big boys.
  If, however, you have a lot of quant and stat heavy processes and
  decisions in your business and you make extensive use of SAS or SPSS,
  e.g. insurance underwriting or chemical engineering, this one is worth
  more than a look. Of course, you won’t get any of that information from
  their website or their product literature.
• The folks at Haley
  have a small memory footprint rules engine that might fit in a PDA. I’m
  not sure that this is where the low hanging fruit is for business, but
  if that’s your thing, have a look.
• After sitting through several "anatomy of a business rules engine" type presentations and watching the folks from Sapiens
  sigh and roll their eyes every time the subject of the Rete algorithm
  came up ("It must suck. After all, Forgy has released Rete-2,
  Rete-3,…Rete-78, yada, yada…"), I got the idea that they don’t use
  Rete. So I resolved to stop by their booth and get the dope on what
  exactly they do. Well, they don’t do Rete, and what they say they in
  fact do do is "data driven." In the whirl of their terribly high level
  diagrams and brochures, I got the impression that they do some sort of
  lazy forward chaining. If you have a truly huge transaction based
  systems with millions of transactions flowing through the system per
  day, then these guys may be worth a look. However, on the eternal
  business rules question of "where the heck do I plug this into my IT
  infrastructure and what do I do with it once I have it in place?",
  Sapiens didn’t give me the warm and fuzzies. Maybe it’s just bad
  marketing materials that make their product look overly complex and
  opaque.

That’s it as far as highlights. I’ll touch on some
of the other players, such as the methodology folks and the rules
repository cats in future posts.

Share/Bookmark