Subject: Re: ok, FSBs created schwag; now bet the farm on something better
From: "Jonathan S. Shapiro" <shap@eros-os.org>
Date: Mon, 17 Sep 2001 10:35:14 -0400

> > This isn't historically true. Historically, all of
> > the major research labs
> > have amply paid for themselves (and yielded high
> > returns, even). The problem
>
> i would have to take issue with this, or at least ask
> for some proof. innovation often benefits someone
> other than the company (or lab) doing the innovating.

Others definitely benefit, but this doesn't contradict what I said.

Actually, I made an overbroad claim. What I know is that a study was done
within Bell Labs research to determine which groups had failed to pay for
themselves. I believe the results were published in the economics research
journal that Bell Labs then still published. I do not now recall the article
title, and I haven't the time to track it down. Feel free if your interest
is sincere -- I'ld be interested to reread it.

One of the hidden questions in the study was "what is the event horizon for
research payoff?"

Over a 5 year period, the study found many groups did not pay for
themselves, but the focus of this study was long term payoff. Much to
everyone's surprise, when evaluated over a 25 year period, not a single
group could be found that failed to pay for itself over the long haul. The
big surprise was that even the number theorists had paid for themselves --
they came up with the cable weave used on local trunks to minimize cross
talk in the 50-pair cables. This is probably the *only* useful thing to come
out of the number theory group over those 25 years, but there it is.

I believe that similar results have emerged from IBM studies, but I'ld have
to check with some people at IBM on that.