Yes, this applies, as well, to the open issue of the rolling labs (autos, ...) upon which, and into which, we put ourselves daily; the idea would be to lessen the experimental aspect (we want existential).
Ah, where is the old DOD where any computation required e-m shielding plus special handling of OS, and resource, issues? Gone with the Internet which is open to mis-'app'lication.
Many talk algorithms when they discuss apps. What? Algorithms are strong; but, they have severe pre-conditions. Even if you had an infinite set of powerful algorithms, there is still the hard problem of matching the algorithm with the problem (assuming that the problem can be defined 'app'ropriately) and then even interpreting results.
Ontology and epistemology are a couple of important issues here.
Now, heuristics (rule of thumb, even if support by Bayesian, and other, methods) can be strong, too, if they are founded on learned frameworks. In fact, we probably have more use for these.
Unfortunately, a lot of the software behind 'apps' is ad hoc. We'll go more into that.
By the way, none of this stuff is as easy as the youngsters try to make it (ah, MS, how many failings are under your paradigmatic responsibility?).
CEO apps? Well, we could have computational voting (at some point - consider the DOD issues that relate to security and stability and truthfulness). People, some, will still need their hands held.
As said before, we have three problems that are not yet resolved: qualification, frame, and ramification. That is, we have to deal with pre-conditions, closure, and then use of any (and in almost all) computational events.
Some pushed data base (static view - behind the numeracy craze (consider: numeracy leads to hubris, hence we need quasi-empiricism; innumeracy is not idiocy (..., as the commercial says -- ..., priceless ... (as in, numberless))) which even if it has operational behavior does not get outside of the 'apps' problems. Then, behavior (which includes those things related to dynamical systems - of which we, of course, are the epitome) analysis comes along with a whole set of age-old problems.
Some modern techniques try to cut to the quick (thanks, Occam - or, are you sorry for all of this modern mis-use?). Such reductions, as these are wont to apply, to the sufficient can throw out the necessary.
That, folks, is probably the most basic cause for a lot of mis-apps and their dire consequences.
Two important concepts, to look at more closely, will that we need to have man-in-the-loop (woman, too), including borgs (yes) and that humans have the uniquely held talent for truth evaluation (computation has an in-laid bit of vertigo that is impossible (yep) to overcome). The latter is trainable; the former is much more than augmented reality.
01/23/2015 -- Software? Well, we are talking more than apps (latest craze). We are dealing with fundamental questions which, then, gives rise to normative issues in mathematics (and, by extension, to the computational).
11/21/2010 -- Three years ago, it was said: Computational foci raise miraculous need. Still applies.
07/25/2010 -- That some have been allowed to misuse the situational uncertainties associated with modern technology, and its use, needs to be discussed. The ca-pital-sino result (the basis that we see for daily gaming a near-zero reality) was, almost, inevitable. How to extricate ourselves reasonably to a more sound foundational framework is of prime importance.
07/02/2010 -- Stunned? Hubris or stupidity (or, are they the same?). Meaning what? Well, this is a simple little thing, of no real consequence. How many problems lurk amongst all of those computational elements that have been spread around the economic world? Who cares?