Stretch/Harvest Stories

Here are stories that have been sent in since the reunion. Some of these stories can also be found on the Stretch/Harvest Newsgroup.

060114  Bill Collier               Obituary for Ken Plambeck 
060105  Ernie Stephens             Look Ahead Memory 
051215  Ernie Stephens             From Ernie Stephens 
051009  Turget Burakreis           From Turgut Burakreis 
030819  Bill Palmer                My dad, Ralph Palmer, is alive and well 
030706  Turgut Burakreis           Pleased to find you 
030630  Fred Brooks                A Brief Memoir
030413  Bill Collier               John Cock's New Car
021207  Bob Bellizzi               To continue Don Peterson's story
021204  Bill Collier               Stretch Conversion Routine
021204  Don Peterson               Another Stretch Story
021129  Dr. Barlane "Ike" Eichbaum IBM STRETCH SUPERCOMPUTER (THE REST OF THE STORY!)
021020  Don Peterson               Three Stories from STRETCH
021010  T.C. Chen                  The Duke Connection
021010  Joe Kusmiss                A tinge of remorse
021010  John Taggart               Contact for Sam Snyder; impact of TRACTOR?
021009  William Stevens            My questionnaire response was missing from the book 
020830  Drew Major                 The Death of Stretch
Last updated January 14, 2006.

[Site map].

=========================================================================================

From: Bill Collier [collier@acm.org]
Date: Sat January 14, 2006, 11:00 AM
Subject: Obituary for Ken Plambeck

I do not often look at the obituaries in the Poughkeepsie Journal, but this morning I glanced at the page and was very sad to see the following obituary for Ken.

Bill Collier

Kenneth Ernest Plambeck

Poughkeepsie. Kenneth Ernest Plambeck died peacefully at Vassar Brothers Medical Center after a long illness.

Born in Chicago, IL, on December 21, 1933, son of Ernest Plambeck and Helen Armbrust Plambeck Wagner. He was the stepson of Gladys Plambeck and Gale Ray Wagner.

He was a 1955 graduate of the Georgia Institue of Technology with a Bachelor of Electrical Engineering degree. In 1958 he earned a Master of Electrical Engineering degree from the University of Illinois, ChampaignUrbana, where he was President of Tau Kappa Epsilon.

On December 29, 1962 he married Isabel Miller Foshee at the Second Presbyterian Church, Louisville, KY.

He is survived by his devoted wife of 43 years and three loving daughters: Virginia Plambeck and companion, David Brown, Arcata, CA; Susan Barnes and her husband, Abednego, Milwaukie, OR; and Margaret Zaepfel and husband, Richard, Leesburg, VA; four grandchildren, Julia, Jessica, and Jonathan Barnes and Jillian Hua Zaepfel. He is also survived by the loving Foshee family, Louisville, KY and a cousin, Ernest Ott, Schaumberg, IL.

He was beloved by his family and many friends and colleagues. Kenneth served the military on active duty from 1951-1953 in the Army Special Forces and Airbourne and spent four years with A.R.O.T.C. at Ga. Tech. He joined IBM Poughkeepsie as an Electrical Engineer in 1958, retiring in 2002 as a Senior Software Engineer. During his distinguished 44 year career, he wrote the compiler for the "Stretch" and contributed to the diagnostics for the "Harvest Machines". He co-authored "Principles of Operation", the architecture manual for many IBM mainframes. He received three "Outstanding Contribution Technical Awards" in 1989, 2002, and 2003 and holds 17 patents with IBM.

He was an accomplished contract bridge player, becoming an American Contract Bridge League Life Master in 2003. He was a 40 year member of American Radio League and served on the TKE Fraternity Advisory Board, Marist College, for seven years.

There will be a memorial service at First Presbyterian Church, 100 Cannon St., on Saturday, January 14, at 1 p.m. followed by a fellowship reception.

There will be no calling hours.

Burial at Cave Hill Cemetery, Louisville, KY will be at the family's convenience.

Memorial donations may be made to the First Presbyterian Church, 100 Cannon St., Poughkeepsie, NY 12601 or Habitat for Humanity of Dutchess County, PO Box 70, Fishkill, NY 12524.

Isabel can be reached at 7 Daisy Lane, Poughkeepsie, NY 12603.

=========================================================================================

From: Ernie Stephens [ern1es@aol.com]
Date: Thur January 5, 2006, 1:30 PM
Subject: Look Ahead Memory

I'm trying to find a way to get a copy of a Patent tilted "Look ahead memory". I headed a small group that wrote the application for this Patent. It was awarded the out-standing IBM patent of the year 1963. I've searched Pat Gov but they cut off the library at 1975.

Would appreciate any ideas you may have on the subject.

=========================================================================================

From: Ernie Stephens [ern1es@aol.com]
Date: Thur December 15, 2005, 7:45 PM
Subject: The Lost Ernie Stephens

Hi, Bill,

Thru Bob Jones I heard that I was missing so did not get to go to the reunion of Stretch. So sorry. I did find the roster on Google. w/ your invite to communicate. I was there from the beginning to the installation. I'm alive and well building Timber Frame Homes on the West Coast.

I seem to miss some names that have stuck in this cloudy brain of mine. Max Femmer, Clarence Frizzel--maybe my timing is off but I feel certain that they were in the same era as Palmer, Stetler, Slobodzinski, Sweeney, Stringfellow, Stephens. I did follow w/ 701,702,3,4,77xx.etc in Field Eng and Machine Technology. In the Stretch era I was responsible for Maintenance Design-Error correction ala Hammer (Bell Labs) and the field engineers.

I am very happy to find this trace of history. I some how disconnected as I migrated to STC in CO to the West Coast. I don't know where Google got all this but it's great technology to which we all contributed. We had our share of great patents that IBM enjoyed for a couple of decades.

Regards,

Ernie Stephens
Residential Design Associates
Authorized Representative of Timbercraft Homes, Inc.
15650 Camino Del Cerro, Los Gatos, CA 95032
Phone: (408) 358-4143
Fax: (408) 885-1010
Cell: (408) 255-0122
ern1es@aol.com
=========================================================================================

From: Turgut Burakreis [turgut@alum.mit.edu]
Date: Tue October 4, 2005, 2:03 PM
Subject: Contact Info for Stretch/Harvest Alums: Turgut Burakreis, Istanbul

October 4, 2005 - Istanbul

Dear Bill,

Thank you very much again for inviting me to the reunion meeting in 2002.

I would like to present my new address for your future mails as follows:

Permanent email address: turgut@alum.mit.edu

Home address:
Turgut & Fatima Burakreis
Aygaz Koop No. 11
34450 Kilyos, Sariyer
Istanbul, Turkey

With my very best wishes and regards,

Turgut Burakreis

=========================================================================================

From: Bill and Sharon Palmer [wp@berk.com]
Date: Tue August 19, 2003, 7:18 PM
Subject: IBM's Ralph L. Palmer-Project Stretch Reunion

Dear Mr. Monroe:

I just stumbled upon upon the website for the Stretch Reunion and I was amused to see that my father, Ralph Palmer, was listed as "deceased" and that this information was provided by you. Well, I'm pleased to report that he is still very much alive and healthy at the age of 94 and is enjoying his retirement in Delray Beach, Fla. I know that he wanted to maintain a very low profile after his retirement, but is not ready to be known as being "deceased" quite yet.

Hopefully, you can reinstate him to "still living" status.......

Best Regards,

Bill Palmer

Kinderhook, NY

-----------------------------------------

From: George R. Monroe [ak2lar1@attglobal.net]
Date: Wed August 20, 2003, 4:15 PM
Subject: Re: IBM's Ralph L. Palmer-Project Stretch Reunion

Dear Bill & Sharon,

Please accept my most abject apology for this terrible blunder. For the life of me I can't recall having reported this but it must be so if Bill Collier says so. I am delighted to hear that my friend and mentor is alive and well and enjoying his retirement. He taught me so much about project management and more, about intellectual integrity. Among other valuable lessons, he taught me perseverance not to mention how to divert one's anger to fruitful endeavor. I cherish my days learning at the master's knee, so to speak.

Please pass to him my warmest regards and best wishes. I am a mere 76 years old now but still active with my three PC's, networked and active on the Internet.

I would dearly love to send him a note of good cheer and thanks for mentoring a crass youngster during the late fifties and sixties when IBM was a Company like no other.

Regards,

George

-----------------------------------------

From: William W. Collier [collier@acm.org]
Date: Wed August 20, 2003, 8:05 PM
Subject: Re: IBM's Ralph L. Palmer-Project Stretch Reunion

Dear Bill and George, Being a husband and a father, I am unaccustomed to being referred to as infallible, but I will say it has a nice ring to it.

I am not certain of the source of this error. I regret any distress it has caused others and any missed opportunities for communication with Ralph. Ralph's updated contact information is:

Ralph L. Palmer

Delray Beach, FL
Bill, does he have an email address that I could list? Bill Collier

=========================================================================================

From: Turgut Burakreis [tib03@superonline.com]
Date: Sun July 6, 2003, 3:24 PM
Subject: Pleased to find you

Dear friends,

I am so pleased to find you through my daughter's google search.

After living in the U.S.A. for forty years and working in the software industry, I have returned to Istanbul. I have two daughters and two granddaughters by my first marriage and two daughters through marriage to my second wife.

I can be reached at this email address: tib03@superonline.com.

I would be delighted to hear any news from you.

Best wishes,

Sincerely,

Turgut Burakreis

=========================================================================================

From: Fred Brooks [brooks@cs.unc.edu]
Date: Mon June 30, 2003, 10:25 AM
Subject: A brief memoir

For Sharon Codd, Libby Codd, and all the Codd children:

I'm very sorry I can't come. Mrs. Brooks and I will be returning from England that date.

A brief memoir:

It was my pleasure to work with Ted on Stretch, where I developed a high respect for his wisdom and insights.

Then, after I had a year's sojourn in the Research Division in Westchester, I moved back to Pok to become Data Systems Division Architecture Manager. In this role, I became manager over Ted's project of building a multiprogramming operating system for Stretch, the first such ever built, I believe. Ted needed no managing, only an occasional advocacy with the higher management.

Of course, I had no techical role in the project, which was well along, but I was proud of it anyway. I still cherish the Conference Proceedings where Christopher Strachey of Oxford, one of the U.K.'s best computer scientists, stood up during a paper discussion and postulated that it should be possible to build a multiprogramming operating system, and one could even imagine time-sharing a machine among users. Ted then stood up and explained that his team had in fact done that very thing!

As the Stretch special OS project was winding down, Ted and I talked, and it was my privilege to nominate Ted for the IBM Ph.D. program, which he received. I have often thought that this, rather than the System/360 project, may have been the most useful thing I ever did at and for IBM.

My sympathy to you all and my prayers for your comfort, and best wishes for a deeply meaningful service honoring a great man.

Fred Brooks

=========================================================================================

From: Bill Collier [collier@acm.org]
Date: Sat Apr 13, 2003, 11:15 PM
Subject: John Cock's New Car

I just came from the 50th anniversary celebration of the Unitarian Fellowship in Poughkeepsie where I talked with Don Ehleben. He wasn't on Stretch, but we started the game of "Did you know ...?" and he had some good stories about several of of the names I brought up. One of the stories was about John Cock. John was known for being independently wealthy and for being too absent-minded to cash his paychecks. One day on his way to work John bought a new car. When it came time to go home, he couldn't remember where in the 705 parking lot he had parked it. Worse, John, being John, could not remember the make of the car or even the color of the car. He did remember the location of the dealer he had spontaneously stopped at, so some folks called the dealer and got enough identifying information to enable them to find John's car and to let him get home.

=========================================================================================

From: Bob Bellizzi [bklynboy@bellizzi.net]
Date: Sat Dec 7, 2002 2:27 PM
Subject: To continue Don Peterson's story

The initial main CB panel was one of IBM's standard colors. Safety decided that it had to be 'international yellow' but it looked to me like one of our simian relatives' nether quarters in heat. Hence the Mfg group started calling the color 'B.A. Red'. We have all been creative and imaginative people so I need to go no further with this.

The PDU was an OK piece of equipment but it made things r_e_a_l_l_y s_l_o_w to 'come up' because each box was cycled individually up with a motor-driven rheostat from 0 to standard voltage from the 400 hertz MG unit. Seems slamming 400 hertz 3 phase into power supplies caused some spectacular results.

bob bellizzi (adv mfg)

=========================================================================================

From: Bill Collier [collier@acm.org]
Date: Wed Dec 4, 2002 4:23 PM
Subject: Stretch Conversion Routine

In the reunion book I described the need on Stretch to convert characters 0-9 on punched cards to the characters A,1-9 internally. Specifically, it was necessary to convert:

  
    0 -> A
    1 -> 1
    2 -> 2
     ...
    9 -> 9
Jim Havender came up with a solution using five instructions and three constants which converted eight characters in a word all at once. After a great deal of effort, I improved on this result with a routine that used only four instructions and one constant. So many people have asked me how I did this that it really was a great stroke of good fortune when I found the original five cards from 1961 as I was getting ready for the reunion. Here, at last, to satisfy everyone's curiosity is the routine:

         LF(BU),($2),5($1)
         &, CONST,5
         C!0001(BU,27,3),187245.45,1
         C0111(BU),($2), 8($1)

CONST    DD(BU,64),(16)7F7F7F7F7F7F7F7F
=========================================================================================

From: C. Don Peterson [Ceedeep@aol.com]
Date: Wed Dec 4, 2002 1:59 AM
Subject: Another Stretch Story

The time came at the South Road Road Laboratory to integrate the parts of Stretch under one power supply. A 400-hertz generator was to supply power to the entire machine via a large cabinet, the main feature of which was three large circuit breakers. A crowd gathered to celebrate this milestone. Someone volunteered to step forward and activate the three circuit breakers. He heaved and the circuit breaker made a loud click. To his surprise, it fell back to the off position when he released it. He tried another with the same result. The third acted the same way. It took about an hour to find and fix the problem, but by that time the crowd had dispersed. It seems in their anxiety to comply with IBM's constant emphasis on safety, the power guys had ignored an important requirement, namely that although the circuit breakers would quickly trip on an overload condition, it took a tiny amount of power to hold the circuit breakers in their "on" position. In summary, power wouldn't come up because power wasn't up! A small push-button fixed the problem. One had to keep the button pushed while activating the breakers.

=========================================================================================

From: Dr. Barlane "Ike" Eichbaum [BEichbaum@aol.com]
Date: Fri Nov 29, 2002 8:19 PM
Subject: IBM STRETCH SUPERCOMPUTER (THE REST OF THE STORY!)

This book ("Wasn't That a Time--STRETCH/HARVEST RETROSPECTS") tells much of the unwritten story about the First Supercomputer developed by IBM Scientists and Engineers in the 1950/60 Era. I (B. Eichbaum) made written inputs, as have several others, telling of important contributions and accomplishment and many of the problems encountered with such a major project as STRETCH. Many of the contributors have, truthfully, expressed their views of the engineering accomplishments as well as their disappointment in management including the President, Thomas Watson, Jr. In 1961 President Watson publicly announced that he was disappointed in the Stretch computer because it did not meet all of the proposed specifications hoped for in 1956 (pages 14-17, 82). Unfortunately, people, like Mr. Watson, who are business and marketing oriented do not always understand the importance of technical and scientific breakthrough, and also the limitations of what is available and possible at that time period. As a consequence, Steve Dunwell, the Project Manager, was essentially "put out to pasture," as many others who helped to develop this advanced "state of the art " computer which was a major transition toward newer and more advanced computers that we have today. At a later date, Mr. Watson, seeing an error in his ways, apologized to Mr. Dunwell and made him an IBM Fellow which he deserved for his major contributions in making The Stretch a reality (pages 28, 82,83.) STRETCH did make its mark in the advances and breakthroughs of computer development, and IBM financial gains. This was followed by CDC's development of Control Data's Model 6600 Supercomputer, having three times the power of STRETCH, after IBM dropped the ball, by not improving and upgrading the STRETCH COMPUTER. This was a poor decision by top management! (pages 14-17,82)

Joining IBM, I had the opportunity of working with a *small group of engineers under William Lawrence (who didn't get credit), who did what many said "it can't be done." We developed a key building block, SILO, the Core memory, with much higher speeds, driven with transistors, and reducing the power input while increasing the speed over earlier computer memories. (pages 84-87, 36; also C. Allen, page 58) With this breakthrough it made possible, the next phases of Stretch engineering development to complete the system.*Some of the engineers on our initial Stretch Memory project in 1956/7, developing this initial building block were: Charlie Allen, Dick Booth, Ed. Councill, Greg Constantine, Bob Flaherty, Dick Lamy, Bill Lawrence, and Lang Stallart. See the index for their contributions and comments. Unfortunately, this group of engineers was dispersed onto other projects and were not available to continue to pursue improvements, as needed, to meet the original specifications. In my case, although I developed the needed memory, logic and matrix switch cores(page 88 Fox, B7), I was not allowed to publish these breakthroughs, and other people were promoted and took the credit for my contributions. Ed Councill (pages 35, 76) who was in charge of the memory packaging R &D , in a recent email (attached) acknowledges these contributions. In fact, I went to IBM World Headquarters, to complain, to the President of IBM, on these and related matters, and forming a Component Div., resulting in my being fired from IBM. I, too, was a victim of Mr. Watson's wrath & misunderstanding of innovation and technology, and he closed his Father's (Thomas Watson, Sr.) "Open Door Policy"! A Component Div. was started, shortly after!

Fortunately, Gene Amdahl, former project Manager of Stretch, who had earlier been sidetracked in favor of Steve Dunwell, left the company, and was interviewing, in Poughkeepsie, NY, for Ford Aeronutronics, Div. of Ford Motor Co... I interviewed with Gene, which resulted in my being hired to setup and do memory research & development for their advanced BIAX and other advanced computer systems. One system went into the Apollo space capsule. In the 1960s Gene Amdahl came back to IBM as architect & designer of the 360 system, which was to be what Aeronutronics was to build, but decided not to. After making that program a success, using much of the Stretch technology, he was pushed out of IBM again, because his brother had a small company building peripheral equipment for computers. He was successful in that venture and 3 other companies, indicating that IBM lost another key engineer due to poor management decisions. Some INFORMATION IS REPORTED BY GENE AMDAHL.(pages 58-60)

Many advances were made in development and engineering and in the software to operate this computer system. It was a marvel of which many engineers made contributions for its success. My comments relate to the hardware development of STRETCH.

Ironically, Charlie Sciandra, who had worked at IBM as a technician, and came to work with me at Aeronutronics, as a BIAX development and processing engineer, had a major part to play for Control Data. On my recommendation to CDC, of Charlie's capabilities, he became the manager in charge of development and manufacturing the memory cores for Control Data which were used in their CDC 6600 (Pg.17, 85), the new Supercomputer to supersede STRETCH.

In 1962, I, too, had the opportunity of joining Bob Noyce, of Fairchild Semiconductor Co. and later (see attached correspondence) Intel to help start the IC Memory Division, because of my directing the advanced memory Group, at Ford Aeronutronics,. using advanced technology, similar to present IC manufacturing, We had developed a very high speed thin film memory at Aeronutronics, operating at 10 times the speed and one tenth of the power of the best memories at that time in 1960s. My decision, in not joining Noyce, because my wife, Bea, had been injured in a car crash, and was very depressed. I decided that her health was more important than my career -- and we were blessed with her recovery. A result of my 1962 input and the vision of Bob Noyce, Intel developed and now produces the Pentium memories, which are used in most of the new and present day PCs (Personal Computers), as well as being used in: Military, Research, Industrial and other automated equipment components. Incidentally, Bob worked at Philco Corp R&D. where he saw the need to develop integrated Circuit Technology for future electronic technology. They would not support his program so he joined, Shockley and Fairchild Semiconductors Corp. to fulfill his dreams. About 5 or 6 years later, Philco-Ford contracted and paid royalties to Fairchild to use the IC technology which he helped them develop. A poor management decision by Philco.

This has been written to put together some of the puzzle, from my experiences and others, as told in this book, about the Rest of The Story. It indicates that major advancements can be made, and improved on if proper management decisions are made. It also indicates that competent technical personnel, can go out on their own and be successful, when a large bureaucratic organization, makes poor judgments. They can be successful in the United States, if they don't give up.

=========================================================================================

From: "Don Peterson" [Ceedeep@aol.com]
Date: Sun Oct 20, 2002 4:34 pm
Subject: Three Stories from STRETCH

Customer Engineers from the Texas/New Mexico and Washington, DC areas were called in to help debug the first STRETCH computers. As a CE who accompanied the first STRETCH to the Los Alamos, NM, AEC Laboratory and helped to maintain it for several years, I recall vividly three stories from the development of STRETCH. They are presented here.

One

The "roll-a-gon" panels that held the printed circuit cards making up STRETCH were supplied their various voltages from a multi-layered bar that ran through the middle of each panel. Taps were taken off that bar at each row of cards. A label was pasted on the tap group at each row to distinguish each voltage. Colors were used to highlight these. Such labels, when turned sideways, resembled campaign ribbons. Someone thought of the idea of snipping off the colors and pasting them to the bottom of his IBM admission badge. Soon many were doing this, and just for fun I did it too.

I was making my way along the hallway in the Poughkeepsie Development Lab one day when I met two men whom I'd never seen before. As they passed me, I heard one of them remark, "Gosh, now there is a guy with some real service time."

Two

My area of expertise in STRETCH was the Basic Exchange, the precursor to the "channel" idea in IBM's 360-370 Series computers. One day at the Poughkeepsie Lab I was approached by two men who said they were experimenting with digital sound. "Could I," they asked, "explain the protocol by which input/output devices interface with STRETCH?" I did so, and within a week or so they were back, asking whether I could help them attach their digital sound device to a channel of the Basic Exchange and to simulate an I/O command as if it had come from the CPU. When I did, I heard the computer talk. I can still remember what it said. "This is an example of digital audio output. BEEEEEEEEP." The beeping noise used up about 3 seconds of the remainder of the STRETCH memory, one megabyte. In retrospect, I do believe I was hearing the very first example of digital audio.

Three

The bugs we removed from the prototype STRETCH computer were also removed from a STRETCH computer in the next room. That computer was equipped with HARVEST and went to the National Security Agency. Except for knowing it consisted of a lot of fast, wide-width tape reels and "streamed" its data, we did not know how it worked, and we didn't ask.

But one day over a beer after work, as I recall, one of the Customer Engineers charged with debugging and maintaining that computer asked me a strange question: "Don," he said, "have you ever played Wizard?"

I had not, and I asked him what it was. Without replying, he went to a nearby phone. I saw him dial a number and heard him say, "Is this the Wizard?" He chatted amiably with "the Wizard" for a few minutes within earshot, then suddenly turned to me and said, "The Wizard wants to talk to you."

Almost expecting something like being squirted with water in the ear, I went to the phone and said a hesistant, "Hello."

The voice on the phone said, "Hi, Don. You're Don Peterson, aren't you?"

I said, "How did you know my name?"

He said, "I'm the Wizard. I know all."

My friend then took the phone back and thanked the Wizard.

I had to know the secret. He explained that when he greeted the person answering the phone as "The Wizard," that person was clued in. He began to recite the alphabet, and of course I could not hear this. When he said the letter "D", my friend said, "How are you, Wizard?" and that clued this man to the fact that the first letter of my name was a "D". The conversation continued in this way until he had my full name, and the rest is a no-brainer.

It was years later in Los Alamos that it dawned on me one day that, in playing Wizard, my friend may have been telling me how his HARVEST machine worked. What follows is conjecture, but I can't help but wonder if I'd figured it all out.

Suppose that, as a method of communicating with the intelligence community around the world, the NSA would broadcast an unending series of random letters. While broadcasting this stream of letters, NSA also records them on tape. Anyone wishing to transmit a message back to Washington need not have a "code book," merely a small electronic gadget that would accept the plaintext, "listen" for the stream of letters going by, and "time-mark" the intervals between the desired letters. It would then transmit the time marks. With no way to collate these time marks against a previously transmitted stream, no enemy would be able to decode the intended message. In other words, it's a kind of "Wizard" game.

If the NSA ever played "Wizard" they could not do it nowadays; the ability to store huge amounts of data by the "enemy" would surely preclude it.

=========================================================================================

From: "T.C. Chen" [tcchenuccu@hotmail.com]
Date: Thu Oct 10, 2002 3:01 pm
Subject: The Duke Connection

Dear Eric,

Please send the Stretch book and mug to my USA address.

Thank you and all the organizers very, very much for organizing the Stretch Reunion, and for or the information about old friends. Steve Dunwell talked about the Duke Connection even when he was heading the Stretch project, but I was the least influential of the Duke team. It was so nice of Julia to mention it. Ever a gracious hostess, she had ample inner strength to weather political storms before the worth of the Stretch project was recognized by Steve's IBM Fellow nomination. While talking about the Duke connection, let me point out that Fred Brooks also was a Duke graduate, but he joined IBM after graduate school at Harvard and a stint at Texas Instruments.

As I wrote earlier, I last met Steve last in St. Petersburg, Russia in an APL Conference. He took off his watch, put it on a bench, and took a shower; after the shower the watch was nowhere to be found. I let him have my card-sized hand calculator-cum watch so that he could get to the flight home on time.

While still at Duke, John Cocke, though a mathematician, was well known (notorious?)for looking behind the shoulders of physicists and suggesting improvements on their research. He insisted on getting the same starting salary as mine, and I thought it a terrible mistake. Among the first thing he did after joining IBM was to ask Jim Pomerene to teach him computer design.

Sully of course, inspired us into choosing computing as career. I was struggling with hand machines in calculating molecular integrals, and a friend suggested that I should seek advice from a visiting mathemtician from Oak Ridge called Sullivan Campbell. Sully gave me a programming manual of the Oak Ridge Oracle machine. I remember well how I was stuck on the unfamiliar concept of a branch (called a "jump") instruction, and had to ask Sully for help. I wrote dozens of pages of programs on the Oracle but they were never run, as the head of computing at Oakridge was not on friendly terms with Sully. It was hard to imagine that Sully could have enemies.

Thanks again. Cordially, Tien Chi

=========================================================================================

From: "Joe Kusmiss" [jkusmiss@mitre.org]
Date: Thu Oct 10, 2002 2:57 pm
Subject: A tinge of remorse

Eric and Bill,

I got my mug a few days ago and I liked it. It got an honored place among the awards and mementos in my office. I got the book yesterday and browsed through it. You know I've never gone to a re-union. I guess I prefer remembering people as they were without benefit of current reality. Also I never do interviews or questionnaires (which has gotten me into trouble with some of my management over the years). When I see there are others in the same category I don't feel so bad. Yet as I leafed through the book I had a tinge of remorse. I saw my name in the index and saw that Bob Manente had mentioned me and it helped. I was there, I was part of it, and I'm very proud of it. I looked for all the other people I knew in the index and was glad they were all there. The people from Church street and then NSA and back to Poughkeepsie again. The only name missing was Alex Kotwas. He was one of the first people I worked with along with John Carter and Dick May. I'm sure they all realize I remember them. The work was very interesting and stimulating. I wish the rest of my assignments were that good. But above all I cherish the memories of the camaraderie, the laughs, the happy hours, yes even some of the problems. You two (and any and all other contributors) are to be commended for putting this thing together. Best wishes, Joe Kusmiss

=========================================================================================

From: "John Taggart" [jltagga@aol.com]
Date: Thu Oct 10, 2002 3:10 pm
Subject: Contact for Sam Snyder; impact of TRACTOR? - from John Taggart, NSA

Eric/Bill,

I want to thank you for the HARVEST Reunion. I enjoyed it very much. I know that it took a lot of work.

I am interested in any people that you turn up that worked on the TRACTOR system, since I worked on that part of the system. I am interested in a summary of the impact of TRACTOR on future systems, along the line of Erich Bloch's presentation on STRETCH.

I have also talked to Sam Snyder's son, Joel. He reiterated his appreciation for the Harvest reunion people considering his father. He said that his father would appreciate calls and/or emails from HARVEST Reunion attendees. Sam does not have an email address, but Joel will pass email on to his father. Joel's email address is jsnyder@audiodescribe.com.

Thanks again for your efforts on the reunion.

John Taggart jltagga@aol.com

=========================================================================================

From: "William Stevens" [w.y.stevens@acm.org]
Date: Wed Oct 9, 2002 4:19 pm
Subject: My questionnaire response was missing from the book

Somehow, my questionnaire response never reached Eric, so it was not included in the book. Below is what should have appeared!

A1 William Y. Stevens, Gwendolyn H. Stevens

A2 South Portland, Maine 5 Nov 31

A3 Bates College BS, Physics 1953
Cornell University MS, Engineering Physics 1955
Cornell University PhD, Engineering Physics 1958
A4 Two

A5 None yet, one on the way

B1 Before Stretch, I was a graduate student at Cornell. Previous professional experience was a summer at General Electric in radar system analysis and four years as a research assistant at Cornell in electron tube reliability.

B2 I believe I was hired by Joe Logue, but it might have been Bob Henle.

B3 All of my work on Stretch was in Poughkeepsie. I was there for two summers, 1956 and 1957, before coming permanently in August 1958. I remained at IBM, mostly in Poughkeepsie, until my retirement in 1993.

B4 The summer of 1956 was in Joe Logue’s circuit design group, analyzing drift transistor transient response at a basic physics level.

The summer of 1957 was in Werner Buchholz’s architecture group, with much of the time spent preparing a patent application for Fred Brooks’ invention of the Stretch interrupt system.

When I came permanently in 1958, it was again to the architecture group, where I worked on CPU architecture.

B5 Bob Domenico, Werner Buchholz, Gerry Blaauw

B6 First year: 1956 Last year: 1960

B7 Since the architecture was truly a team effort, it’s hard to single out my own accomplishments. I contributed in various ways to the CPU architecture, particularly in the areas of branching and interrupts. I helped edit and maintain the architectural definition documents.

B8 The architecture group consisted mostly of young PhD’s without a lot of previous experience. We tended to socialize as a group more than most IBM groups I was in later. I made many lasting friendships, several of which survive to today.

B9 Several of us in the architecture group, along with some other IBM friends, formed an investment club. We called it Sigma Investment Club, after a code name that was being used at the time for the Stretch CPU. It existed for 10 or 15 years. I think we all learned a lot, but it wasn’t the primary investment vehicle for any of us, and we didn’t make much money.

Not exactly a humorous anecdote, but one experience sticks in my mind. As engineering models of Stretch were finally being built, I was sent one day to Steve Dunwell’s office. It seemed that release of some printed circuit designs was stalled because the design automation programs couldn’t figure out how to route the wires. Steve was apparently getting a lot of finger pointing and wanted someone to make an unbiased evaluation of how the logjam might be broken. For reasons I never knew, I was chosen to look at this area I knew nothing about. After asking a few questions I determined that a manual "override" was possible. I reported this back to Steve, who was much relieved! I always treasured this expression of Steve’s willingness to give an opportunity to a young, unproven engineer.

C1 As the architecture for Stretch was completed, the same basic group of people turned their attention first to the 9000 Series and then S/360. I left the architecture group in about 1961 to manage the microcode design for the Model 50 CPU. My later assignments for IBM were mostly at a system design level, including reliability/availability/serviceability (RAS), Future Systems (FS) Source/Sink design, teleprocessing system design and various staff assignments.

My changes in assignment were usually triggered by project cancellations, other reorganizations, or the desire to remain resident in Poughkeepsie.

C2 Having my first "permanent" job as a young engineer in the small group that included Werner Buchholz, Fred Brooks and Gerry Blaauw had a profound impact on my career. Their insistence on clear thinking, precise specifications, clean designs and intellectual integrity set a standard that I tried to maintain for the rest of my professional life. Fred Brooks, in particular, had a strong influence. The work I did with him in the summer of 1957 gave me the topic for my PhD dissertation, and he acted as my principal (unofficial) thesis advisor.

C4 My "extracurricular" activities seem to have gone through a series of phases. First there was community theater, where I did lighting design and was president of the group. Then I got involved with volunteer ambulance and fire department work, where I became Fire Chief and then Chairman of the Board of Fire Commissioners. Now I am concert manager of a group that produces chamber music concerts, which my wife and I founded and I was the first President.

Bill Stevens

=========================================================================================

From: Drew Major, Last Stretch System Engineer [Drew_Major@volera.com]
Date: Fri 8/30/2002 2:11 PM
Subject: The Death of Stretch

I owe much of what I know about computers and programming to the four years I worked on the Stretch while I was going to school. My duties included writing and maintaining the system software and helping Bill Ivie and others track down and fix the almost daily failures.

I had the privilege of running the last program on the last Stretch at Brigham Young University. I did a program that wrote "goodbye world" or something like that (it was always a little tricky writing with the lights, having to factor in the parity lights in the middle of the accumulators) which I planned to have running when it was shut off. We didn't get to run it because the video crew brushed up against some of the wires in pushing in one of the cabinets and broke something which made it so that the Stretch wouldn't boot.

In the end, I set the switches to do a single instruction loop doing the square root of the time register and that's what it was doing when the power was turned off for the last time...