Yeah, I totally missed your point. An interesting idea.

I think it would only work when trying to build a team initially, to get the first few good people in the door.

If you bring him in after you already have a team, then it would be a huge slap in the face to your existing developers.

hire a programmer on a contract basis to help them weed out candidates.

That is logical in a lot of ways, and it solves the problem of getting rid of candidates you don't like before they hurt your organization.

However, it may eliminate a lot of very good candidates. At the younger end, it eliminates the hotshots who want stock options; and at the older end it eliminates those with families who want good benefits. And it eliminates anyone who needs to relocate or otherwise make a significant commitment (and turning down other job offers is a form of commitment, as well).

It's a competitive marketplace, and a contracting job now with the possibility of an unknown salary later is a weaker offer than a high salary now. If the goal is to get good programmers, a weaker offer is not a great start.

go back to suspecting women are lying when they call rape

Never said that.

To be fair, at one point it was pretty standard to put the accuser in a case like that more on trial than the accused.

When it's one person's word against another person's word, then the presumption of innocence should prevail. Once a trial has begun, a person has a right to question their accuser (that's in the Constitution).

If there is no hard evidence, then the prosecutor shouldn't drag the victim through the whole process, even if they are really a victim. It doesn't help anyone.

Yes, it sucks to let criminals go free.

But it sucks more to allow one group of people (in this case, women) to put anyone from another group (in this case, men) into prison by simply lying.

society's trying to find an appropriate equilibrium.

Oh, so it's society's fault?

No. It's the fault of the witnesses who lie and the prosecutors who press charges when there is no corroborating evidence. And both should go to jail.

Considering that Social Security is a multi-trillion dollar program that exists mostly so that children can avoid living with their aging parents, these devices seem like a small price to pay by comparison.

those rates would almost certainly be a fraction of injuries and deaths attributed to modern automobile accidents.

Once you divide by the number of people-miles traveled in both cases, it may be quite close.

Cars are quite safe and quite clean (for the environment). The only reason that they are a leading cause of death and environmental concern is because we drive so much.

As an interesting side node, Afilias uses PostgreSQL to run .org and .info.

(I can't find a recent link that directly says that, but clearly they are still involved with PostgreSQL: http://www.afilias.info/news/events/2008/03/29/postgresql-conference-east-2008 )

PostgreSQL needs some form of MERGE/UPSERT/REPLACE. MySQL has REPLACE and INSERT ... ON DUPLICATE KEY ....

Also, postgresql should support index only scans (or similar), which MySQL supports.

I still prefer postgres by far, but those are two legitimate issues.

I made a lot of mistakes that I would not have made with MySQL, mostly with privileges.

Can you elaborate? The complaint that I saw most often was that there was no good way to manage the permissions for a lot of objects. In 9.0, that was addressed with GRANT/REVOKE IN SCHEMA.

Also, keep in mind that the standard has a lot to say about the privilege system, so postgresql won't want to stray too far from that.

but if MySQL works for you...cheaper...and easier, why not use it?

Of course, go for it.

Just don't let inertia keep you from learning something that might really be an improvement. "Works" is not binary -- some things work better than others. SVN works. Git works better. Notepad and GEdit work. Emacs and vim work better.

not all autism is crippling. It can have a fairly broad spectrum

I'm skeptical when I hear things like that. Broad-spectrum conditions that are diagnosed through subjective tests (especially mental/social tests) leave a lot of room for misdiagnosis and over-diagnosis.

In the US, this is a particular problem because there is a financial incentive to both the individual and perhaps to the school if the diagnosis is autism. For instance, the thought process might be something like "well, we're not really sure, but if we don't say it's autism, then we don't get any money to treat it".

Also, it's far from clear that "mild cases" are really a problem at all.

Money is a good incentive when there is a direct and immediate relation between your paycheck and your output

I think this misses the point. Money (and other incentives) work well to encourage inputs. For those tasks where there is a direct connection between inputs and outputs, money can improve outputs (like with your widget example). However, if the connection between inputs and outputs is looser, such as a task that requires creativity rather than brute force, incentives generally don't work as well.