Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email firstname.lastname@example.org
Subject: NSA/IBM vacuum cleaner: "Stretch-HARVEST" lecture: 8Nov PARC 6pm
From: John Gilmore (gnutoad.com)
Date: Wed Nov 01 2000 - 12:08:30 CST
- Next message: R. A. Hettinga: "CALL for Papers: The Millennium Mac-Crypto Conference"
- Previous message: Vin McLellan: "Re: RC4 - To license or not? (OT)"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
The "very large data processing application" this machine was designed
for is being the center of the NSA vacuum-cleaner: Figuring out which
few of the millions of NSA intercepts is worth reading. The machine
amazingly continued in that job until the late 1970's or '80s --
possibly because it was so custom that they couldn't port the job to
any newer hardware.
STRETCH was the biggest IBM computer built at the time.
HARVEST was a fascinating co-processor designed for text search, matching,
I haven't heard the speaker, but the topic is rich in cryptanalysis history
(the branch of cryptography that we know *least* about).
From: Eleanor Weber Dickman, Office of Public Relations, Computer Museum
Subject: Computer Museum History Center Lecture at Xerox PARC
The Computer Museum History Center and The Institute for Women in
Frances E. "Fran" Allen
Senior Technical Consultant, IBM Research's Solutions and Services
speaking on The Stretch-HARVEST Compiler
Wednesday, November 8, 2000 - 6:00 p.m.
at the Pake Auditorium, Xerox PARC, Palo Alto
Reception to follow
Advance reservations are required -- Please RSVP by Monday, November
6, 2000 to Wendy Ann Francis, 650-604-5205 or franciscomputerhistory.org
In response to government requests, IBM Research designed a system for
a very large data processing application, known as the HARVEST system,
including Stretch, which was delivered to the National Security Agency
in the early 1960s. The combined Stretch-HARVEST Project created a
milieu for developing new technologies, new hardware architectures,
and new software to meet the challenges of both systems. One of the
guiding principles of the project was to make programming easier by
the use of a compiler to generate code automatically from statements
in the user's language. Allen was a member of the ALPHA language
design team which created a very high level language featuring, among
other things, the ability to create new alphabets beyond the system
defined alphabets (e.g. English, decimal, integer, binary) and treat
complex, heterogeneous data in high-level statements. In addition to
an overview of Stretch-HARVEST, the talk will describe some of the
lesser known aspects of the project the people and institutions
involved, the political climate, and the shared knowledge, views, and
value systems which were part of this interesting project at an
interesting time in the history of computing.
Fran Allen is regarded as a pioneer in the field of optimizing
compilers. Her achievements include seminal work in compilers, code
optimization and parallelization. She helped create one of the first
automatic debugging systems and, as a member of the Stretch-HARVEST
project, developed an advanced code-breaking language known as
Alpha. In the early 80's, she founded the Parallel Translation Group
(PTRAN) to study compiling for parallel machines. This group was
recognized as one of the top research groups in the world dealing with
parallelization issues. Allen currently serves as senior technical
consultant to IBM Research's Solutions and Services vice-president. In
1989, she became the first woman to be named an IBM Fellow. She was
also elected president of the IBM Academy of Technology in 1995. Allen
earned her BS in 1954 from the State University of New York in Albany
and her Masters three years later from the University of Michigan.
She received an honorary Doctorate in Science in 1991 from the
University of Alberta.
The Computer Museum History Center: www.computerhistory.org
If you would like to respond to this memo electronically, send
mail to dickmancomputerhistory.org.