Subject: quick replies to various messages (Re: How to run an FSB R&D lab (maybe))
From: Tom Lord <>
Date: Thu, 21 Mar 2002 15:51:32 -0800 (PST)

	Aren't you describing something like Bell Labs?

I wouldn't know.  Public accounts suggest it is not much like that
now, though parts of it may have been more like that in the 70s and
early 80s.

	What you described is Xerox Parc of the 80's and early 90's.

	It produced great technology and ideas, and didn't run in the

Ultimately, I wonder if that isn't more of an accounting problem than
anything else?  How much office equipment did Xerox sell as a result
of Mac and Windows and all that they led to?  What has the escape of
unix into the world done for the sale of telecommunications services
in the last 30 years?

	And note also that those environments (Athena, Andrew) 

Your message said interesting things, but just to clear up our
respective meanings, I'm not talking about development that was part
of either of those (or any other) sponsored project.  Universities
facilitated in the sense that computers were available, meeting
spaces, equipment graveyards, etc -- but I'm talking about work that
wasn't part of any grant-receiving projects.  Work that took place on
or near the same campus as, for example, Andrew -- but that wasn't
part of the Andrew project.

       Corporations tend to only justify this kind of overhead when
       times are good; in lean times [the last 2 years show it well]
       many commercial research institutions fade away

My strategy doesn't need a castle-on-the-hill research facility, and
doesn't need many of the expensive components of the big corporate
labs (like expensive apparatus for bleeding-edge fab processes or
basic physics and chemistry).  It needs competitive salaries and a
decent petty-cash fund.


	Seriously though. I think allot of that "magic" you seek is
	only very prevalent in a student/mentor environment.

Agreed.  A mix of experience levels is a good idea and a reasonable
expectation is that the more senior researchers are self-motivating
for things like paying attention to the business context, good
engineering practices, good experimental practices, good
dissemination, and balancing those things against a wild-ride for
transient members.  Speaking of transient members, cycling a subset of
the engineering staff as a whole through stints at the lab seems like
a good way to encourage cross-fertilization.


	If universities are such a great model for creating saleable
	value, then why have they been unable to hold tuition
	increases to the rate of inflation?

As an aside, some big universities with an open source track record
are rethinking that position for exactly revenue reasons.  That's
something FSBs who figure universities into their strategic thinking
should keep in mind.

More to the point: the environment I'm pointing to is outside of the
projects "owned" by the universities.  In some famous cases, students
do in fact take these projects and turn them into businesses or
commercially significant projects.

	My point is that if you want to make the model work, you must
	change it so that you can PAY the workers, instead of charging

Right.  Here's the essence of my sales pitch: "Look, it's a good
idea.  Something valuable will probably come out of it." :-)

	3. The behavioral protocols that are established and practiced
	will, in large part, determine the success of the venture.

Absolutely.  The people employed in this situation have to be
internally motivated to work on that issue -- that's their
responsibility trade-off for the institutional degrees of freedom they

In that light:

	CS grad students do have the additional motivation of getting an
	advanced degree.

And employees of my lab have the very strong motivation of having it
not fail to do anything of recognized value.