When incentives don't work, when CEOs ignore the
long-term health of their companies in pusuit of
short-term gains that will lead to massive bonuses,
the resonse is always the same:
get smarter incentives.
The truth is that there are no incentives you can
devise that are ever going to be smart enough. Any
incentive system can be subverted by bad will.
I think part of this talk is directed at the messed-up American legal system which has trickled down to affect a lot of life in America. You can't make an exception for fear of a lawsuit. The rest of the world doesn't have a culture of lawsuits, and as a result, wise exceptions flourish.
If we remove lawsuits with large monetary penalties, leaving small penalties and, of course, criminal penalties, frivolous lawsuits will die on the vine.
Does anyone know of a movement to make this change? And if it doesn't exist yet, who's with me to start one?
That's a good idea but it could have the consequences that a poor person would have to front money to go to trial, just in case he or she lost.
I think getting rid of the large settlements (replaced with criminal charges) is the way other countries have it. Again, I'd be interested to see if work can be done towards this.
> That's a good idea but it could have the consequences that a poor person would have to front money to go to trial, just in case he or she lost.
Not necessarily. Some "loser pays" systems allow the plaintiff's lawyers (or an insurance company) to take the risk of paying the defendant's costs if the lawsuit is unsuccessful. (The key is that those systems allow such folks to get paid for taking that risk.)
In the US, when a contingency lawyer loses a case, s\he is typically out his time and occasionally expenses as well. Loser-pays "just" doubles the hit from a loss.
I'm not saying that loser-pays is a good idea, I'm just saying that it doesn't necessarily exclude poor folks.
I think much of what this guy talks about applies directly to programming, too, and especially speaks to the friction between management and programmers. Managers want hard rules, and programming is not amenable to this.
I've been writing an essay on this topic for a while which really reminds me of this video, and one of the main thrusts is that you can't just learn rules, and it targets the question of how do you develop the wisdom to program. It's not all that great and it's not done (mostly unposted). It's not stuff that would necessarily be news to anybody here, but it's been interesting writing it all out.
There's no substitute for trusting programmer judgment, and that bothers people. Programming isn't unique that way, either, it's just the field I know.
I think a really great example of this is the "One Minute Manager" and similar management books on the market. They try to give you rules for managing the people you work with, without considering the human / personal element of the job. I have known people with managers like these, and they always fail. People catch on pretty quickly.
One of the good trends in software (IMO) is that in rejecting rule-driven ways of running projects (like waterfall, RUP, etc.), people across the industry have basically agreed that the right way to build software is to make it a human endeavor again, thus Agile, XP, etc. The bad trend I see is that some people are trying to create rule-driven systems out of Agile and XP (e.g. 99% unit test coverage required, or you must pair program every line of code in the system). This makes my head hurt, since it is so against the spirit of the Agile manifesto.
If we are going to count on hope and virtue to get us out of this mess, then we might as well give up because it will only be an exercise in wasting time.
