| From: HANNAH::HASTINGS "DECprint Arch. (508)467-8299 MR01-2/K20 06-Dec-1994 1735 -0500" 6-DEC-1994 17:40:26.46
To: @GOODBYE.DIS
CC: @SYS$PUBLIC:HCE.DIS,HASTINGS
Subj: I'm retiring 16-Dec; I'll miss you all
This is a note to the many people I've worked with over the past 27
years that I'm "retiring" effective this 16-Dec. My wife has wanted
to move back to southern California, where she grew up. I'll be
working for Xerox in El Segundo, as Principal Scientist in their
Architecture & Document Services Technology Center.
I'm sad to leave Digital after such a satisfying career and as Digital
starts to turn the corner. I see many positive signs and wish
everyone continued success. I've enjoyed my various assignments and
groups over the years. I've actually worked for Digital three times.
So let me indulge in a few reminiscences:
I started working for Digital as the first programmer (what we called
ourselves then) as a summer job in 1961. Ken did the interviewing. I
wrote a floating point package for the 18-bit ones-complement integer
PDP-1. Looking back, I sometimes wondered if maybe he thought that a
graduate student could produce all the software that the company would
need in a summer. I've never joked with him that I'm sorry I let him
down...
The company had contracted with Ed Fredkin for an assembly program.
Ed named it FRAP for Fredkin Assembly program. There was another
contract for a one-pass Algol compiler called Decal (pronounced dee
cal). The linker performed the second pass by having the user put the
intermediate paper tape in the reader backwards so that the forward
references could be resolved. Linked lists hadn't been thought of
then.
In 1964 I started full-time on the PDP-6 operating system. The PDP-6
was a joy to program (in assembly language) with its many
well-organized instructions. We produced the first commercial
multi-user multi-programmmed operating system. Eight DECtapes with
read-write addressable blocks were used for user and system program
storage. Disks came later. Bill Segal was in charge of compilers and
application software.
The PDP-6 kept the engineering department (Gordon Bell, Alan Kotok,
Bob Clements, Dave Gross, Tom Eggers ...) hopping in order to keep
the hardware running in 20 or so installations. Wisely, the company
decided to re-engineer the hardware more conservatively to produce the
PDP-10, and keep the software that we had developed for the PDP-6,
which now included program swapping and data files on a fixed head
disk.
Larry Portner and I hired Dave Stone to run the operating system group
for the DECsystem10. (Helping to hire your own boss was an example of
Digital's flexibility and acknowledgment of the alternate technical
career path). Peter Conklin also joined our group. For TOPS10 on the
DECsystem10, we re-wrote the file system (calling it by the prosaic
name of Level D). Many of the ideas were re-used in the FILES-11 file
system used by RSX-11M and VMS.
We brought TENEX in from BBN for the DECsystem20, along with Dan
Murphy, to get a jump on a demand paging operating system. Peter
Hurley was project leader.
Later Gordon Bell asked me to join the architecture team (Bill
Strecker, Dave Rodgers, Steve Rothman, Richy Lary, myself) to design the
VAX, which we knew would ultimately put the DECsystem10 and 20 out of
business. Designing the VAX and participating in the early
development of VMS were immensely satisfying to everyone involved.
For VMS V1.0, I led the development of the VAX Common Run-time
Environment. We worked closely with the compiler people: Rich Grove,
Ron Schaefer, Kevin Harris, Ron Brender. With the VAX procedure
calling standard and the VAX CALL instructions, we had the means for
inter-language function calling; and industry first. First came
FORTRAN, then BASIC and COBOL. PL/I and PASCAL followed.
Meanwhile, I also helped develop the ANSI/ISO escape and control
sequence standard working with Dave Hughes who made the VT100 and
follow-on terminals become industry standards. The LN03 and follow on
Digital laser printer used the same standard.
I worked with Bob Travis, Lois Frampton, and Tim Lasko, on developing
the 8-bit DEC Multinational coded character set. Then, as chairman of
the ANSI committee on character sets and codes, I helped get agreement
in ANSI and ISO on ISO Latin-1, now used in Windows. We also managed
to change the ISO two octet draft standard to agree with the industry
Unicode standard and thus avoided a major schism (like the EBCDIC vs.
ASCII schism for the one byte set of two decades previous).
Most recently, I've been working on the ISO DPA and POSIX standards
with Brian Handspicker and Pete Smith for a distributed printing
system as implemented in MIT Palladium, IBM Palladium, and now a new
printing system for OSF/1.
I'll miss the company and the people. We've done some pioneering
products in the past and will continue to do so as we become more
competitive in the new world of desk-top and client-server computing.
The loyalty and dedication that the engineers continue to demonstrate
makes it clear that Digital will be a winner once again.
Goodbye and good luck,
Tom Hastings
P.S. My e-mail address after January 1, will be:
hastings.escp10@xerox.com
|
| Tom worked for me _three_ times in our careers. I have always found
him to be an incredibly collaborative person--one of the nicest with
whom I have worked.
Our first collaboration was during the heyday of the Decsystem -10s
and -20s. To put this in perspective, Tom not only introduced
Timesharing software to Digital. Tom is the co-author of the seminal
papers at MIT that created the concept of timesharing. During our
tenure together in the -10 group, I was proud to be able to hand Tom
his promotion to be Digital's very first consulting engineer. At the
time, he was the first engineer at Digital to take a "sabbatical" to
return to school for a semester to maintain his currency with the
state of computer science.
After his success with the industry's first commercially viable time
sharing system, Tom then joined with Dave Cutler on a research project
to explore a state of the art operating system advanced development.
Out of this came some of the leadership concepts that Cutler used in
the creation of VMS. Tom joined the VAX architecture team, again
working for me. Tom held to the concept of "everything can call
everything." Out of this, he created the concept of the VAX common
runtime environment, as instantiated in the common runtime system.
This capability led the industry by at least 15 years. Today, we would
call this object oriented programming.
Later, I was again privileged to "direct" Tom's work when he worked
for me in the VIPS (video, image and printing systems) arena. Tom not
only led the development of the DECprint model, but he worked to
establish this as the only basis for a distributed printing system in
the industry. He obtained the complete consensus of such intense
competitors as Xerox, Siemens, Kodak, IBM, Digital, HP and Adobe to
the vision of one distributed printing system for all networks.
|
| Tom hired me in 1978 to work on the VMS Run-Time Library. At that time he
was, I was told, one of only five Consulting Software Engineers in the company.
Much of what I am today is a result of Tom's leadership, guidance and
insistence on "doing the right thing". I owe Tom an enormous debt of
gratitude.
Good luck, Tom, wherever life takes you.
Steve
|
| Farewell gathering for Tom Hastings
Wednesday, 14-Dec, 5:00 to 8:00 PM
The Atrium
Best Western Royal Plaza Hotel
181 Boston Post Road West, Marlborough, MA 508-460-0700
Drop in and chat, with nibbles and a cash bar.
RSVP appreciated: David Larrick, HANNAH::DCL, DTN 297-8340
Cost: $3.00 to defray the cost of the event, plus whatever you like for
a gift for Tom
Pay in advance if possible to David Larrick, MRO1-3 pole KL23
or Mary Morton, MRO2-2 pole B13 (look for the balloon)
Please suggest gift ideas, whether classy or corny!
Directions to Best Western Royal Plaza Hotel:
From I-495:
Take Rt. 20 westbound exit. After 3/4 mile, turn right at Best Western
Royal Plaza sign. (The hotel has a long driveway, and is not visible from
Rt. 20.)
Your destination is the main (second) building, not the "Royal Plaza Trade
Center". Park in "Function Parking", but use the main (lobby) entrance, not
the "Function Entrance". We're in the Atrium, which is part of the lobby.
From Digital MRO1 facility:
Turn left out of the parking lot onto Forest St. At the traffic light, turn
right onto Ames St. At the traffic light, turn right onto Rt. 20, then
immediately left at the Best Western Royal Plaza sign. (The hotel has a
long driveway, and is not visible from Rt. 20.)
Note: the parking lot for ViewLogic, straight across Rt. 20 from Ames St.,
does NOT connect usefully with the Royal Plaza.
Your destination is the main (second) building, not the "Royal Plaza Trade
Center". Park in "Function Parking", but use the main (lobby) entrance, not
the "Function Entrance". We're in the Atrium, which is part of the lobby.
|