When I say "if everyone follows this philosophy," I include both VCs and college acceptance boards in that antecedent. Then, presuming you don't have a trust fund, the only way to get your start building stuff is to do it on the side while being underemployed; this is a net efficiency loss for the economy.
Firstly, colleges are not in the business of making admittance decisions based on whether you are a good programmer or not; secondly, 40 hours a week is the standard for being reasonably well-employed, which leaves a whole lot of hours for doing your own thing in life (not to mention the years you're in primary and secondary school, during which many programmers build a pretty passable CV of personal work.)
I feel guilty even typing this much, because I don't think you're being intellectually honest at all. The vast array of open-source and free projects in the world speaks plainly about the ability of people to build useful things outside of work.
Okay, let me now be completely clear on how I interpreted your original statement:
Let "an institution" be "a place where one may work on projects safe in the knowledge that one's base needs for food and shelter will be met."
I(x) := x is an institution
S(x) := x seeks to join an institution
A(x,y) := x will accept y
P(x) := x has completed a project
1. ∀xy( (S(x) ∧ I(y) ∧ A(y,x)) → P(x) )
2. ∀xy∃z( (S(x) ∧ I(y) ∧ P(x)) → (I(z) ∧ y≠z ∧ A(z,x))) )
∴ ∀xy∃z( (S(x) ∧ I(y) ∧ A(y,x)) → (I(z) ∧ y≠z ∧ A(z,x))) )
Or more prosaically, "every institution that will accept a person (and thereby let them create a project) requires that there exists a project they have already completed (and thereby another institution willing to accept that person.)" Obviously there's a problem with that statement :)
That being said, I agree completely with your first paragraph. I was being half-satirical—not intending to find a flaw in your argument, so much as desiring to provoke you into strengthening your point by clarifying it (and trying to go one step beyond the common-sense assumption of "our current society" to see where this philosophy would take you.)
Do you think the world would be better if programmers could not transition into the workforce until they had created something that no one had asked them to create? The idea behind a company hiring you is that they need what you have, and so they pay you for it. If you're creating something and no one's paying you for it, then, presumably, either you aren't marketing it well enough, or it's something nobody needs (or you've made it as an act of goodwill, but the likelihood of Joe the Programmer writing FOSS as anything other than a prerequisite to a job—which, therefore, means that he really would rather have been paid, but just couldn't find a taker, again implying that he was underemployed—is slim to nil.)
I admit that I smiled when I saw a big response to me with chunks of predicate logic in it.
Do you think the world would be better if programmers could not transition into the workforce until they had created something that no one had asked them to create?
I'm not sure if the world would be better. There's an overwhelming demand for competent programmers, and not anywhere close to enough supply to fill it. Is it better to lower our standards, resulting in shoddy and sometimes dangerously bad work, or to leave things undone?
However, for any given single institution, I am very confident that that institution would be better off hiring only people who had created things which nobody asked them to create. I agree that these are usually things which either nobody needs, or which are marketed poorly; but I don't see why that would indicate anything negative about the person in question (as far as the company is concerned.)
I think this is trickier than just saying "do this, you obviously have time", but not for the reasons that derefr gave. I think you are assuming things that may not be true of all people, and then using that to draw very broad strokes which may have a steep rate of false predictions.
Forty hours a week is an theoretical standard; it says little about the actual hours put in by programmers. Urban lore suggests the accurate number is higher, sometimes much higher.
The free time a programmer isn't guaranteed to be available in large chunks for hacking. There are a number of things that can take up that time; some of them are bogus, but some are quite reasonable. Enough of that time gets chewed up, the programmer really doesn't have the kind of time they need to focus on deep, substantial projects. Some do find the time, but there are more factors involved than simply motivation.
As for primary and secondary school years, this too depends on a number of factors, not all of them matters of personal motivation and ability.
Assessment is hard enough when you don't have preconceived notions of the molds other people fit into; if you start assessing people already cherry-picking assumptions about them based on what societal fashion suggests is "normal," then your ability to fairly assess them is already screwed. You're not trying to fairly assess them for them; you're doing it for you, because doing it fairly is your best shot at not screwing up.
This is where the author's string of side projects comes in to play. I mean face it, it's kind of difficult to make something really cool in your spare time without getting paid to do so.
