The first, that programmer hacking on their own projects, I agree is a very positive indicator. If they don't, it may or may not be a negative indicator, depending on the circumstances. However, I've found the combination of someone who really has the itch to hack in their own time, and that same someone who honestly can't find the time is rare. More often, after enough prodding, they'll admit they just aren't that interested, and that it is "just a job like any other".
The second, that selecting candidates based upon social connections is an optimal strategy, I approach with some caution. This is not to say that it isn't the best way for raganwald to do things, but I have seen people develop insurmountable confirmation biases simply because someone they know gave a candidate the thumbs up. A lackluster interview performance wasn't enough to overcome this; poor performance once hired wasn't enough either. The trust the hirer puts in their network to suggest the right people must be for the right reasons. This, in a person's profession network, is rather rare. More often, two people knowing one another is the basis for a referral, rather than what the two people know about one another. This works ok in situations where the primary goal is to find a good social fit, but it is like throwing dice when determining if the candidate is capable of performing the required duties. In a lot of cases, both parties just don't care; the referral is a social, not economic or professional, transaction. Luckily for both parties, many companies don't care that much either. Mediocrity is not really affected by the introduction of more mediocrity, so neither party ends up with a tarnished image. This sort of politicing sucks.
My perspective on this is much the same as Joel Spolsky's -- or at least the perspective he used to have. If a particular candidate comes pre-recommended by someone, I'm really hesitant to know that. If I'm the person doing the recommending, I always wonder whether I'm doing it because this person is a genuinely good fit, or if I am just feeling the pressure to be a good friend. If I was both introduced to him and one of the one's vetting him, I'd be really concerned that my bias is way too strong for me to make an honest, objective decision. Hiring people based on a blind resume submission sucks, too, and I don't know of any other alternatives. But I worry we have a blind spot to the biases of this method of finding candidates, simply because we want to believe the best in everyone we know.
Elsewhere, Giles Bowkett described hiring through networking, HN, blogging, twitter, and so forth as being "inbreeding" and would produce "retarded" results.
p.s. Giles was positively kind in comparison to discussions elsewhere. It really shows that individual communities do have a very strong cultural bias, which I think supports your point and his:
The first, that programmer hacking on their own projects, I agree is a very positive indicator. If they don't, it may or may not be a negative indicator, depending on the circumstances. However, I've found the combination of someone who really has the itch to hack in their own time, and that same someone who honestly can't find the time is rare. More often, after enough prodding, they'll admit they just aren't that interested, and that it is "just a job like any other".
The second, that selecting candidates based upon social connections is an optimal strategy, I approach with some caution. This is not to say that it isn't the best way for raganwald to do things, but I have seen people develop insurmountable confirmation biases simply because someone they know gave a candidate the thumbs up. A lackluster interview performance wasn't enough to overcome this; poor performance once hired wasn't enough either. The trust the hirer puts in their network to suggest the right people must be for the right reasons. This, in a person's profession network, is rather rare. More often, two people knowing one another is the basis for a referral, rather than what the two people know about one another. This works ok in situations where the primary goal is to find a good social fit, but it is like throwing dice when determining if the candidate is capable of performing the required duties. In a lot of cases, both parties just don't care; the referral is a social, not economic or professional, transaction. Luckily for both parties, many companies don't care that much either. Mediocrity is not really affected by the introduction of more mediocrity, so neither party ends up with a tarnished image. This sort of politicing sucks.
My perspective on this is much the same as Joel Spolsky's -- or at least the perspective he used to have. If a particular candidate comes pre-recommended by someone, I'm really hesitant to know that. If I'm the person doing the recommending, I always wonder whether I'm doing it because this person is a genuinely good fit, or if I am just feeling the pressure to be a good friend. If I was both introduced to him and one of the one's vetting him, I'd be really concerned that my bias is way too strong for me to make an honest, objective decision. Hiring people based on a blind resume submission sucks, too, and I don't know of any other alternatives. But I worry we have a blind spot to the biases of this method of finding candidates, simply because we want to believe the best in everyone we know.