[8254] in Commercialization & Privatization of the Internet
A Note for Randy Windell on the difference between coop agreement & contracts
daemon@ATHENA.MIT.EDU (Gordon Cook)
Fri Nov 12 11:42:37 1993
From: cook@path.net (Gordon Cook)
Date: Fri, 12 Nov 1993 16:41:57 GMT
To: com-priv@psi.com
Just a semantic tweak for one of Randy Windell's comments on contracts with
NSFnet. As those of us who have been down these paths before know, the MERIT
pork barrel is *NOT* a contract but rather a **cooperative agreement**. The
difference is hugely important. The new solicitation will also be for
cooperative agreements.
NSF evidently maintains that its 1950 authorizing legislation requires it to use
this form where a kind of loose cooperation is agreed upon. IE the parties
state goals, but exactly what will be done and under what conditions is not
specified because the subject is research into NEW technologies where, by
definition, outcomes are not known in advance.
Now why could NASA and DOE sign a **contract** with Sprint for delivery of an
ATM backbone? Apparently because their founding legislation does not require
them to pursue everything in a research cast. Can Congress do anything to
change the NSF picture? In theory sure. In reality unlkely. The Congressional
staffers like Mike Nelson and Jim Wilson, who would have to be won over to tell
their bosses that new legislation should be passed, are generally phd science
graduates of the universities that the current situation serves very well and
whose interest is in seeing that the status quo is maintained.
Occasionally the NSF does put its foot down.... I wish it would happen more
often. The Princeton John von Neumann Supercomputer Center is a case in point.
The consortium for scientific computing was one of the last groups to get a
proposal in. They wanted a Cray center but realized that the chances of getting
funded were very small if they did anything else but bid for a Control Data
Cyber 205 Center to be followed by the ETA 10. They were awarded a control data
center.
When they got the award, apparently there was a contingent who thought they
could go under the flexibility of the cooperative agreement and say, it really
looks like we could serve our community better with a CRAY. I have been told
that they actually started to spend big dollars on installing the kind of
cooling system necessary for a Cray. They told CISE that they wanted to dump
control data and buy a CRAY and **to CISE's eternal credit**, they were told on
**no uncertain terms**, you cease and desist immediately or we will void your
cooperative agreement, close you down and put it back up for rebid. The center
got of on such a bad start with NSF that when Control Data did exit the super
computer business in may 1989, NSF was delighted to shut us down which they did
without further ado or repercussion. Personally, it was too bad. I enjoyed
working there. But from a point of view of a responsible use of taxpayer money,
the NSF in that decision to close von Neumann was absolutely correct.
Unfortunately that situation didn't have the high profile that the NSFnet
backbone SERVICE now has.
By the way, NSF notified MERIT in August 1991 to get ready for an extension of
the cooperative agreement that was not scheduled to expire until Nov 1 1992. [I
know this from documents obtained from NSF under FOIA.] This was some months
before NSF announced in on Nov 26, 1991 that the coop agreement would be rebid.
Now I would imagine that NSF has already informed MERIT to get ready for its
next extension. Why? because now one who has talked with me belives that the
four new contracting entities will be able to achieve a cut over to the new
order of the universe in less than 4 months. Looks like its time for a new FOIA
request.
