A Craftsman hammer vs. a Stanley hammer
...either will still pound nails; it's just a question of which one feels better in your hand.
The point is...we're still in a <a href="http://en.wikipedia.org/wiki/Craft">craft-work</a>, solidly pre-engineering phase of development. The software for the Jules Verne ATV or the Boeing 787 Dreamliner may be the most complex, needful-of-near-perfection software we have yet created; amazing achievements - but still, a monumental achievement in the literal sense, on the order of <a href="http://en.wikipedia.org/wiki/Reims_Cathedral">the cathedral at Reims</a>, where major features were over-engineered by now-seemingly "ridiculous" levels because structural and materials knowledge at the time was not in any way up to today's levels.
All the "patterns" and "best practices" and so on that have yet become reasonably widespread in our craft do not yet combine to form a body of knowledge on the same level as, say, a civil engineer's ability to predict how a dam built of certain materials to certain dimensions and located on a specific place on a river will hold back that river's water. In other words, we haven't yet met the definition of "engineering" as "using a known process to take sufficiently known inputs and produce a sufficiently knowable product, by individuals whose knowledge and/or education meets an industry-wide set of standards."
"Software is different," you say. "Anybody can learn to do it, and we don't need no stinkin' gubbamint regulation." The first part is (technically) true, as long as quality of resulting work is ignored. As for the second.... imagine a failure scenario such as this: a "market-leading" spreadsheet, used to build the software that presents information to millions of small/midsize businesses around the world, ships an update that includes a subtle, seemingly intermittent defect. That bug happens to get exercised in complex calculations across multiple pages of a spreadsheet - say, the kind of thing that these companies use to plan and track some budgetary areas. As a result of the problem, errors are introduced into these calculations that cause American businesses collectively to "lose" billions of dollars, often going bankrupt in the process. Does anybody seriously believe that the government wouldn't - or shouldn't - intervene <em>once it were shown that the defect was a root cause of the fault</em>?
"Poppycock", you say (if not something less printable). "THEY wouldn't ever screw up that badly." To which I have two, engineering-related, responses. First, remember the FDIV bug. Then think about pumps and dikes in New Orleans - a very engineering-aware failure.
Finally...consider this: At what point does a recently-widespread technology or practice (such as software construction) acquire sufficient impact on and importance to public policy, safety, health and welfare that any government that fails to at least actively monitor its development and use becomes negligent in its duties? Would you want to drive a car built by home hobbyists that lacked sealed headlamps, seat belts or safety glass? Would you, as a member of the motoring public who saw that others were driving such a car, feel safe and unaffected?
Is a world largely dependent upon a single vendor's software which exercises no effective oversight or regulation of that software, or the dependencies being built on it, truly sane? Or has it degenerated into a miasma of Randian "individualistic" groupthink that may well cost it dearly later on?