Software profession suggests GMOs are untestable (Score 1)

So a stray bit of DNA got through without the rigorous testing regime noticing it? That's hardly surprising. It's high time genetic "engineers" reviewed the lessons of software validation. We know that only the most trivial programs can be tested empirically. Serious software validation involves a spectrum of methods that go way beyond testing, including formal proofs (a rare luxury) and, most particularly, code inspection. Nothing beats a human brain (or better still, a team of brains) poring over source code, exploring it mentally, looking for logic errors, or stray elements that just don't make sense. But you have to understand the code you're reading. So I've long been concerned that they just cannot know enough about how DNA works to be able to meaningfully code-inspect the products of recombinant genetic engineering (although in the reported case, perhaps a simple inspection of the sequence would have spotted the viral gene fragment; does anyone even read the sequence before a GMO goes into production?). The great surprise reported in Nature last year than junk DNA isn't junk goes to show how immature the basic science of genetics still is. It's presumptuous indeed to call it genetic "engineering" when they swap sequences (sub-routines) between organisms without a full understanding of how genes are expressed. Empirical black box testing is the only tool available for validating GMOs, but any software engineer will tell you, that's not enough. There can be very little confidence in any amount of testing of such a complex system as a modified genome. See a longer presentation of the argument here: http://lockstep.com.au/blog/2011/01/15/not-ready-for-gm.