RE: Formal methods



Date view Thread view Subject view Author view

Mike Ellims (mike.ellims(at)pitechnology.com)
Thu, 7 Feb 2002 14:28:02 -0000


In reply to response to my "slightly frivolous" last posting, I shall dig a little deeper :-) Martyn Thomas wrote... > > Is this meant to be an argument against the use of formal models? It > could be taken to be one. > No, I'm not using this as an argument against but rather as a caveat. It was noting that a formal specification is just a model of a system and that models have a habit of not always being right. > > Would you buy a significant bridge that hadn't been analysed in this way? > Anything by Telford :-) I'm not sure I'd feel completely happy buying off someone who had never built a bridge at all... experience counts for something but unfortunately it's not quantifiable. Peter B. Ladkin wrote... > > Do you have a reference? Chapter 15 of To Engineer Is Human (entitled > From Slide Rule To Computer: Forgetting How It Used To Be Done) lists > the pitfalls of using computers uncritically to perform engineering > estimations and evaluations, but there is no such statement as the above > there. Neither is there one in his Design Paradigms. > I don't know for sure, a side effect of having two offices and a mobile set of books. I'll try and find exactly what he says, the wording is not his (no quote marks) but I believe that is essentially the message. Michael Schmidt wrote... > > I _must_ be missing something here. A large, expensive error is a desirable, > nay, a "standard" route to correcting a mathematical model? > Not desirable, but Petroski would argue that there is always a non-zero probability of it occurring when you try something very new. The Millennium Bridge was very new in the arrangement of suspension members. The problem is that you can't incorporate in a model something you don't know about. The full title of the Petroski books are as follows. To Engineer Is Human : The Role of Failure in Successful Design Design Paradigms : Case Histories of Error and Judgment in Engineering ------------------------------------------------------------------- If a little knowledge is dangerous, where is the man who has so much as to be out of danger. Thomas Henry Huxley 1877 ------------------------------------------------------------------- Mike Ellims Pi Technology mike.ellims(at)pitechnology.com www.pitechnology.com phone +44 (0)1223 203 913 (direct) phone +44 (0)1223 441 434 (reception)


Date view Thread view Subject view Author view