OK - I have calmed down <TeeHee> Either you missed the point of my original post entirely or I made it so vague as to invite misinterpretation.
w/r/t to a rules-based system, the rules themselves are a specification of behaviour. However, since rules may be written to produce a system that is definitely non-deterministic, IT management will probably endure a significant amount of pain trying to explain to a Federal auditor why the system behaves differently depending on how (the ordering of) facts are asserted. Is this not a property of rule-based systems (using forward chaining) ?? Would a Federal auditor not view this as suspicious if the IT shop can not explain the behaviour ?? Hmmm, are you cookin' da books Bubba ????????????? I'm still calm, since I am NOT a CEO nor a CFO - I'm not going to jail, they are. OBTW - I did enjoy reading your "Micro Workflow" papers some time back. ----- Original Message ----- From: "Mariusz Nowostawski" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, March 30, 2004 4:31 PM Subject: Re: JESS: A New Dawn Breaking in Software (Rules) Engineering ?? > You seem to be implying that given a formal specification of a system > expressed in one formal language and an implementation of a computing > system (in another formal language) it is possible to test if the > implementation meets the specification. This has been proven almost 80 > years ago, and over and over again to be not true, i.e. there is no > effective procedure which could tell you if given implementation meets > given specification. Assuming both to be formal and deterministic. > Therefore, your message and described "vision of software era" has no > sense whatsoever, and your "fears" are unjustified. You can calm down now. > There are some things which cannot be done, even if US Congress may be > stating otherwise.... > > However, it is easy to meet the requirements of Federal auditors: when > asked for the proof that the software meets the specification, one needs > to show to the auditors the source code of the system, and tell them > that this is a specification of the system (written in such and such > language). If the software itself is the specification of the system, > the implementation, by definition, "meets the requirements" of the > specification. > > > > Rich Halsey wrote: > > I am being advised that the dawn of a whole new era of software > > engineering is about to happen, i.e., legislation by the US Congress is > > now forcing US corporations to comply with new financial reporting > > standards and by extension it means that IT shops will now have to prove > > to Federal auditors that their applications are not "cooking the data" > > which could be used to "cook the books". I would take this to mean that > > Federal auditors can demand to be shown that the existing IT system code > > corresponds EXACTLY to the engineering documentation (if it even exists > > !!) and corporations that cannot produce proof of this will fall of a > > cloud of suspicion. This seems to be leading to the P.O.P. Syndrome > > (Probability of Prosecution) as witnessed by ENRON, WorldCom and others > > because the legislation directly targets the CEO and CFO of any publicly > > held US corporation. > > > > So, how many IT shops can document how their (non-deterministic) rules > > systems actually work ?? Or, how many IT shops can do a "show and tell" > > on any of their systems to the satisfaction of Federal IT auditors ?? > > OBTW, all systems that have been off-shored now have the same > > requirements if the data ultimately feeds to the financial reporting > > systems - just how much does NOT ???? So, it seems that all off-shored > > work will now be under a glaring microscope just like the US work. > > > > This all seems to imply that the days of "web speed" software > > development may be nearing an end, and the re-emergence of "acid tested" > > software is about to begin. Is the industry up to this - from what I > > have seen over the years, IT shops may need an "attitude adjustment" > > from the CEO/CFO levels. > > -------------------------------------------------------------------- > To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]' > in the BODY of a message to [EMAIL PROTECTED], NOT to the list > (use your own address!) List problems? Notify [EMAIL PROTECTED] > -------------------------------------------------------------------- > -------------------------------------------------------------------- To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]' in the BODY of a message to [EMAIL PROTECTED], NOT to the list (use your own address!) List problems? Notify [EMAIL PROTECTED] --------------------------------------------------------------------
