[262] in DCNS Development
[cec@MIT.EDU (Cecilia d'Oliveira): ["Hal Abelson"
daemon@ATHENA.MIT.EDU (Jon A. Rochlis)
Fri Oct 16 18:51:49 1992
Date: Fri, 16 Oct 92 18:42:52
From: jon@MIT.EDU (Jon A. Rochlis)
To: developers@MIT.EDU, jon-dist@MIT.EDU
------- Forwarded Message
Date: Fri, 16 Oct 92 18:08:12 EST
From: cec@MIT.EDU (Cecilia d'Oliveira)
To: dcns@MIT.EDU
Subject: ["Hal Abelson" <hal@martigny.ai.mit.edu>: In case you haven't seen
this. Please forward.]
------- Forwarded Message
Date: Fri, 16 Oct 92 15:02:44 -0400
From: "Hal Abelson" <hal@martigny.ai.mit.edu>
To: acs@MIT.EDU, jis@MIT.EDU, cec@MIT.EDU, jdb@MIT.EDU
Subject: In case you haven't seen this. Please forward.
Reply-To: hal@martigny.ai.mit.edu
Address: MIT AI Lab, NE43-409, 545 Technology Square, Cambridge, MA 02139
Phone: (617) 253-5856 Fax: (617) 258-8682
DATE: Thursday, October 22, 1992
TIME: Refreshments at 2:00 pm
Talk at 2:15 pm
PLACE: Room NE43-518
War Stories from Andrew
James H. Morris
Professor of Computer Science
Carnegie Mellon University
Abstract
The Andrew Project was Carnegie Mellon's version of Project Athena.
I was its first director.The project started in 1982 and grew to about
30 people. Eight were IBMers, ten or so were newly-minted Ph.D.s of
Computer Science. We built Andrew, partially in the model of the Xerox
Alto system. It achieved some real successes but didn't accomplish all
that we'd hoped. Much of the problem was intrinsic to the way the world
is, compared to how we thought it was. I often meet people who still
have some of my 1982 misperceptions.
I will reconsider some of the early, crucial decisions of the Andrew
project related to wiring, protocols, workstations, file systems, and
software. I'll assess the impact of Andrew on education and discuss
what IBM and Carnegie Mellon ultimately got out of it. I will argue that
UNIX, the Macintosh, and Electronic Mail will not survive to the next
century.
Finally, I'll offer eight opinions that computer scientists who venture
into
the real world should consider:
1. Slowness Kills.
2. Big, integrated systems lose.
3. Virtual << Real.
4. Programmer productivity is not so important.
5. Technology transfer rarely works.
6. Engineers don't buy; they log roll.
7. Industrial funding is difficult.
8. 2nd System < 1st System < 3rd System.
HOST: Prof. Jeannette Wing
------- End of Forwarded Message
------- End of Forwarded Message
