Alexander J. Glass on the benefits of taking chances in technology development programs.

Glass:

That's really another subject entirely. I don't think peer review — I think peer review has to be used with a certain amount of caution, where peer review tends to be a way of reducing risk, and in technology development programs you need risk, so I think too much — you get a situation where NSF is, where they just won't try anything that's new. If you have to have some professor say that it's as good idea, then it's an old one.

Bromberg:

Is that something that was well understood for example in IDA at that point?

Glass:

I think it was well understood at ARPA, in the early days of ARPA. The atmosphere we're talking about was a collective atmosphere of IDA and ARPA when ARPA was first set up, and it was, that was the great advantage of it, was that a program manager like a Glenn Sherwood and his successor John McCallum could take chances.

Bromberg:

That's very interesting, it's a new aspect —

Glass:

In the job I’m in now, one of the biggest problems we have in the company is encouraging an atmosphere in which people are not risk averse. And at the same time, we can't be profligate in our use of resources, but the more I see peer review in action, the more I fear it's a great leveler that keeps new ideas — in many cases, they erect too great a barrier for new ideas, when you really would like to go into the laboratory and try things. Some of them will work, some of them won't. So I think peer review has a negative side. I think program managers should be able to fail.