One can argue that what we get with computation, if done right, is a good model for the world, our systems, their processes, and a whole bunch of other things. Then, this model can be used for analysis, for decisions, and learning. That last applies to the theme of this blog.
It is very easy for unsound modes to come into play with computation added as a resource, to wit product modeling as we've seen with the project motivated this blog and the accompanying mania about life cycle management's new expertise, called systems engineering.
You see, folks, if we're going to overlay computation on the world, then, we need to be aware that 'undecidability' is a stronger, and more prevalent, phenomenon than allowed by the 'can do' thrusts of engineering (and the associated business mind that wants something from nothing).
'must' and 'may' come into play. We'll be looking at a technical paper recently published by the ACM (Intro, Article) that shows what is necessary to statically analyze a program, that is a computer program. But, do we not use 'program' for labeling real world things, such as an 'airplane' program?
Let's be real. There are strong analogues between the cyber and physical (to wit, cyber-physical systems). So, we need to get a handle on the basic notions related to 'undecidability' for the sake of project success (that is, increasing the may) and for continued peaceful existence of that much beleaguer'd thing, called capitalism.
If we were truthful, we would see that 'must' is not frequently the case. As in, 'that' must follow from 'this' (except, tautologies are strong - yes, though trivial, this is a big set). Somehow, power (that is, top-down enforcement -- as opposed to the middle-out of engineering) seems to think that 'musts' are a larger set than what is real.
We'll be addressing this, too, from the truth engineering viewpoint, as the need for such endeavors arises from the main quandary that we face.
Systems engineering got to a state of hubris by thinking that it handles the 'may' in a strong fashion, it seems. Let's back up, folks, as they have no better way to handle undecidability than does any other discipline (it's somewhat amusing, as a few years ago, Scott Carson was going on about what we don't know - atta boy, who has 20-20 foresight (actually, who has this in hindsight?)?).
We can use the current state of, and the known (as in public) history of, one project to discussion the important issues, all in the name of progress.
11/21/2010 -- Three years ago, it was said: Computational foci raise miraculous need. Still applies.
09/27/2010 -- Capitalism is for the good of us, let's bring that forward.