[einpoklum]

> to encourage success

Scammers, arbitrage stock-traders, and opioid drug marketers can also achieve success... On the other hand, some endeavors "succeed" by not-failing and being out of sight and out of mind, e.g. a plumbing installation. You don't think of your "daily plumbing success".

So, I'd say not success which should be encouraged, it is well-directed and well-applied effort.

PS - The plumbing example can well be adapted to nearly any pursuit, definitely including software.

[dgb23]

> You don't think of your "daily plumbing success".

Initially I had to laugh at this one, but why not?

Robustness is something that seems to be very hard to measure, communicate and sell until something bad happens.

I think one of the IT fields that is struggling with this is security, because the problem is most apparent there.

[yourapostasy]

> Robustness is something that seems to be very hard to measure, communicate and sell until something bad happens.

This is endemic to business and is especially widespread in contemporary American-style business, and the general case seems to be a failure in pricing all kinds of tail risk. I suspect a key challenge in this space is the unsolved principal agent problem.

I think the principal agent problem is rampant because pricing information is extremely lossy information compression; only the grossest outcomes are recorded in the market. Those who exploit the principal agent problem rely upon externalities to succeed, and those externalities are not timely or fine-grained reflected in the compressed/price information.

Reputation systems lossily add the missing information back into the market as a side-channel. But what externalities to record is a whole other can of worms.