return to main 1401 Restoration Page - Major Topics
- Photos from IBM 1401 era
- IBM 1401 Archive Pics
- 1960s Org Charts with Q/A
- late 1950's Supporting Technologies
- Fran Underwood - 1401 - Design Philosophy
and IBM 1401 1950s Team Bios
- WWAM - (World Wide Accounting Machine) - the hardware design effort by engineers from France, Germany and the U.S. that resulted in the IBM 1401 card (and tape?) system.
Table of Contents
- Pre-1401 History: World Wide Accounting Machine (WWAM)
- Early - 1401 History: Stored Program Accounting and Calculating Equipment (SPACE)
- 1401 Product Announcement
- After announcement:
- Pre-1401 History: World Wide Accounting Machine (WWAM)
- A Motivation for WWAM
- IBM France -
- extracts from "IBM's Early Computers"
- WWAM Directory from Maurice Papo June 2014, 470K bytes
- WWAM Logic, IBM Eng. Symposium , Paris Jan 9, 1957 from Maurice Papo
- "Giro", Giro, Giro_Specs, Giro_Cost_Compare_fr, Giro_Hist_fr
- Oral History and collection of e-mails from Karl Ganzhorn
- Oral History and e-mail from MauricePapo 2008 March 27
- Oral History of Jacques Maisonrouge (.pdf 2.4 MB added Oct 20, 2008)
- Oral History of Max Paley collected 1960s, On pages 10-14, he talks about his involvement in the MAC (which he headed up), 750, WWAM, and 1401 projects.
- A collection of e-mails working to find answers
- e-mails about Joe Logue, Max Paley, Vince Learson, Bob Evans, ..., & gold plated contacts, wirewrap, ...
- WWAM-Photos - April 23, 2008 & Dec 19, 2014
- Estrems 1956 WWAM memo "Presented at the Syracuse University Joint Engineering Meeting June 17-22, 1956", .PDF version, 7.7 megabytes, added October 11,2008
- WWAM Architecture by Robert Weaver "Here's my try at it."
- WWAM Report, by M. Flynn, May1957 - Components, circuits, logic, packaging
- W W A M Project # 7002, Final Report, Jan 1960
- WWAM Patents
- Early - 1401 History: Stored Program Accounting and Calculating Equipment (SPACE)1 megabytes
- Why a (new, incompatible) 1401, not an IBM 650 compatible upgrade?
- late 1950's Supporting Technologies
- "Solid State in '58" - the "quote"
- Oral History - Ralph Mork an IBM 1401 project manager, got 1st three people going
- Main 1401 Story by Charles Branscomb
- Charles Branscomb - early 1401 program manager
"After assessment of WWAM, turned attention to what became the 1401. Directed development of 1401 from the time there [were] three people on program until it was in production." - IBM Career
- Oral History w Saphire, 3.6 MB pdf
Dialogs - Charles Branscomb (1401 program manager) and Robert Garner
- Two IBM 1401 Principle Founders found!! April 16, 2008
- Fran Underwood was designer of IBM 1401
- Design Philosophy, by Fran Underwood
- Fran Underwood, Oral History, 1968 .pdf 3.7 megabyte, HTML 50 K Bytes
- Fran Underwood, Interview June 2008 .mp3 39 min, 4.5 megabyte, HTML 30 K Bytes
- SPACE-1stMemo-Aug1957--.pdf 6 megabytes
- SPACE-2ndMemo-Oct1957l--.pdf 10 megabytes
- SPACE-OpCodes-Oct1957.pdf 2 megabytes
- SPACE-PrelimDescription-June1958 6.5 megabytes
- SPACE-OpsManual-June1958lq--.pdf 9 megabytes
1401 Price Target $2500/Month SPACE/1401 engineering story in Endicott, 1958 ? 59 by George Ahearn Sheldon Jacobs was Product Planning, "Evangelist" ;-))
- Shel Jacobs, presentation June 2008 .mp3 33 min, 3.8 megabyte, ,
- A selection of 1401 development letters and org charts from John Pokoski
- Those thick cables? ;-))
- B.O. Evans' Special Systems Development Org Chart from John Pokosli
- "Monday Morning Quarterbacking the IBM 1401 Design, 50 year later ;-))"
1401 Product Announcement
- Computer History: IBM 1401 Product Announcement 1959, Data Processing Mainframe (& 7070) Mark Greenia & Computer History Archives Project - YouTube
- Presentation to the Product and Market Planning Seminar, Deauville, France - by Sheldon Jacobs, Friday. May 6, 1960
- 1401 - A Product Announcement - .jpg - 943 KBytes
- IBM Products 1959 - .pdf - 1.2 MB
- IBM RAMAC 1401 1959 - .pdf - 575 KB
- Status Symbol
- How much tab room equipment could an IBM 1401 replace? Sept 2013
- Product Test - Chuck Branscomb
- Continuing Engineering 1401 product life after introduction.
- IBM 1401 Archive Pics
A Motivation for WWAM from Chuck Branscom (SPACE/1401 project manager) - Apr 24, 2013
Correspondence between Robert Garner (1401 restoration lead) and Johann Johannsson (composer of the 1401 Symphony), April 6, 2013
Trivia: When I brought this [recording] up with the 1401 restoration team members several years ago, none thought it was for a 1401 and instead suggested it might be for an older punched card machine. After I listened to the entire recording, as provided by your father (cc'd), at the very end of the tape, the instructor states that it's for an IBM 421 Tabulator: http://en.wikipedia.org/wiki/IBM_421
-------- end of background -----------
The reference to the IBM 421 is interesting to me. It is my understanding that one reason World Trade was encountering more difficulty with competitive accounting machines in the mid-50's was that the IBM 407 was not made available in Europe. They had decided that the potential sales volume did not justify tooling, etc. for 407 manufacturing. Bull was gaining some success in Europe and WT alerted Corporate IBM to this problem and asked for help. This led to the huge US/Europe market study which in turn led to the WWAM effort.
"Solid State in '58" - the "quote"
On 7/3/2015 4:48 PM, Robert Garner wrote: > Guy, > > Although TJ Watson spearheaded the policy, the infamous memo was penned by Wallace McDowell in Dec 1957, before the 608 shipped (which was years late relative to its announcement): "It shall be the policy of IBM to use solid-state circuitry in all machine developments. Furthermore, no new commercial machines or devices shall be announced which make primary use of tube circuitry... Any variations of this policy must be approved in writing through this office." > > But IBM designers and CEs were experiencing it the most in 58, thus the "Solid State in '58" jingle. > > The WWAM designers had already decided, after thoughtful deliberation, to build a transistorized machine in ~1955. (The 1401's data paths came from the WWAM. fyi, we're talking with its co-designer, Maurice Papo, cc'd, 93, on Tuesday mornings. We think they may have designed one of the first transistorized oscilloscopes too.) > > And, Joe Lougue, who introduced transistors to IBM in the early 50's, and his students had designed and built the world's first fully transistorized calculator (demo'd around the US) in the mid 1950s (which eventually became the 608). > > As George Ahearn, Jud McCarthy, John Pokoski, Earl Bloom, and Jim Ingram, 1401 designers/managers, and Rick Dill, cc'd, where all there, they may recall the general response/reaction... > > - Robert
Extracts from "IBM's Early Computers" by Bashe, Johnson, Palmer, and Pugh
notes about the book and this web page
- extracts are from pages 459 - 474 which may whet your appetite to see more.
- the book is still available second hand for a modest ($8-10) price)
- the book references interviews with Dr. Karl Ganzhorn (see Table of Contents) several times.
- Emerson Pugh is still active (2008) with a relatively new book "Building IBM: Shaping an Industry and Its Technology"
- references have been removed
- Specification, architecture and design were considerably different at IBM than at Dilbert's organization ;-))
- Choosing what to delete was painful - every sentence is interesting !!
1st paragraph of page 459
In the early 1950s, even the smallest stored-program computer was more expensive than a typical installation of punched-card machines. The $3250 minimal monthly rental of IBM's 650-the least expensive computer the company had installed by 1956-would support an impressive array of punched-card machines. That figure would cover, for example: two accounting machines (Types 402 and 407); two reproducers, usable also as summary punches for the accounting machines; two sorters; a collator; an electromechanical multiplying punch; half a dozen keypunches and verifiers; and several hundred dollars for miscellaneous equipment and supplies.last paragraph of page 465
12.2 The Type 1401 Data Processing System By early 1955, when the MAC program was launched, the domestic argument over small-to-medium electronic calculators had become largely self-sustaining and independent of the European crisis that had triggered it. This did not escape the notice of the French and German engineers, who proceeded to take the initiative in solving the competitive problem posed by the highly effective Bull Gamma 3 electronic calculator. The German IBM design group at Sindelfingen played host to a group of ten or so engineers and planners from France, Germany, and the United States during June and July 1955. The purpose of the seven-week conference was to propose several accounting machinetop of page 466
designs. The idea was that all candidate designs would later be refined and the best one chosen.^25Broadening the Base 467
From each of the three countries came a design proposal. The most distinctive came from representatives of the French laboratory in Paris, who endowed their machine with variable field length, a characteristic of the 702 and 705 with which they were familiar. They felt, however, that the 702-705 method, with end characters to delimit fields and with special instructions for adjusting field lengths, was wasteful of memory space. They believed the familiar plugboard could be used to advantage and at lower cost. With serial handling of data, plugwires would be required only to indicate field boundaries and not for the parallel transmission of all characters of a field as in the traditional card machine control panels. The German and American proposals, which did not differ greatly from each other, were based on fixed field lengths.
The report produced by the conference in Sindelfingen explained that the objective of that effort was to gain economy of manufacture by satisfying both European and domestic U.S. market areas with a single machine. ...
Although a particular machine design was yet to be selected from among the three proposals, the anticipated product of the Sindelfingen conference was known thereafter as the Worldwide Accounting Machine (WWAM).
To take advantage of the transistor experience of circuit design groups in the U.S. laboratories and to consider additional machine design alternatives, the members of the Sindelfingen conference reconvened in Poughkeepsie in September 1955. By year end, they had examined the engineering prototype of the transistor 608 and thus acquired a first-hand appreciation of transistor circuit design and packaging requirements. In December, the group produced a joint recommendation to select the French proposal-including transistor technology, ferrite-core memory, and the variable-word-length characteristic.
The French engineers stayed in the United States for about one year, working in Poughkeepsie and Endicott. ...
468 Chapter 12
the machine would be the responsibility of the French laboratory. The German engineers, meanwhile, had agreed to develop the card reader and printer. The latter was to employ the same stick printing mechanism-based on a development by H. S. Beattie in Poughkeepsie- that was being used in the RAMAC system. ...
A model was eventually operated at 300 lines per minute."'
As matters stood in June 1956, the WWAM was to have a serial-by-character arithmetic unit, being in that respect unlike the 604, the 607, or the 608 from which its circuits derived. It was an add-to-memory machine, having no accumulator and therefore requiring that each arithmetic instruction specify two memory operand locations (one of which was also the result location). As a plugged-program (control panel) machine, it relied on its serial electronic design to keep the size, complexity, and cost of the control panel small by comparison with the panels of traditional punched-card accounting machines. Magnetic-core working storage (used entirely for data, since no stored program was involved) could be as large as 1900 seven-bit characters.
The WWAM project had hardly become established in the European laboratories when it encountered serious problems with estimated product cost, including the cost of electronic circuits and packaging." ...
Furthermore, the WWAM project suffered from a lack of confidence at corporate headquarters in the ability of the relatively small and inexperienced European laboratories to manage a major product development program, one on which a large amount of revenue might depend over the years. As early as April 1956, sales vice-president T. V. Learson had told his superior L. H. LaMotte that, from the viewpoint of the
sales department, WWAM was the "number one" item in importance, outweighing the 750 "by a country mile." But he cautioned against putting the program in "unknown hands" (overseas) and expressed skepticism that the machine could range downward to meet its low-end cost and price objectives, when the 400 series accounting machines had never been able to do so: "The stick [printer], I assume, will be given as the new and different reason. I don't believe it."'Broadening the Base 469
Shortly after acquiring responsibility for the accounting machine area, Mork turned his attention to the WWAM product-cost problem. He learned that the plugboard-along with all the electronics to drive the signals and to ensure against damage arising from an operator's plugging errors-accounted for several thousand dollars of estimated product cost. Rather than upsetting the French laboratory ...
Thus, for at least the fourth time, a variable-field-length stored-program computer was designed-the TPM, the somewhat different
702-705, and the 305 RAMAC having been three earlier examples.^ Underwood dismissed the 702-705 method of field delimitation as too expensive. What was wanted, he later recalled, was a machine capable of replacing three 407s at the roughly equivalent rental of $2500 per month, including card reader, punch, printer, and processor.470 Chapter 12
in the method he chose for marking the boundaries of fields, Underwood retained a good deal of the logic of the plugboard. He proposed simply to add an eighth core plane, called the "word-mark plane," ...
A few weeks after Branscomb became responsible for accounting machines, a market research report made it clear that, since the lowest estimated monthly rental for the European laboratories' WWAM was some $900 more than that of roughly equivalent capability in the form of two 407s plus one 604, the market for the machine would be nonexistent. ...
he introduced the idea of variable-length instructions as well as variable-length data words. it was important to minimize the memory requirement because core memory could be expected to account for around $10 of monthly rental for each one hundred characters. Underwood's proposal was still preliminary in nature, but the memory-conserving principle of variable instruction size had been established."'Broadening the Base 171
Palmer and others in engineering continued to express concern over various aspects of the WWAM project in Europe-not only the rising estimates of product cost but such matters as the German laboratories' departure from the new standard designs, developed at Endicott, for card feeding and reading units. But a top management decision would be required to move so important a project from World Trade to a domestic laboratory; and LaMotte, for one, was being told that substantial delays might be the price to be paid for an unnecessary move.'" Then, late in 1957, Tom Watson, it., expressed dissatisfaction with the program. ...
472 Chapter 12
The newly assigned project, which for a time carried the acronym SPACE (for Stored Program Accounting and Calculating Equipment) and which was to culminate in the announcement of the 1401 data processing system,
The development of core memory, too, was timed conveniently for the new machine. Magnetic-core tape buffers of 800 characters had been in use as early as 1955 in the 702 system, and core main memories of much larger capacities had been part of all 704 and 705 systems, first shipped early in 1956. (See chapter 7.)
The standardized approach to design of card feeds and other card-handling devices had just yielded such machines as a high-speed collator
lower part of 474
with two independently clutched card feeds operating at 650 cards per minute (cpm), much faster than any predecessor. A reader-punch was designed for the SPACE system, employing the same feed for reading and replacing the second reading unit with a punching station; it read and punched cards at 800 and 250 cpm, respectively." By contrast, the fastest electric accounting machine (the 407) read 150 cpm, and the fastest punches available with unit-record systems operated at 100 cpm.
Perhaps the most fortunate aspect of the timing of the SPACE project was the completion, in time for announcement with that system, of the "chain" printer, a machine that was to introduce radical innovation into the design of high-speed printers for computer systems. At the time Branscomb became manager of the SPACE project, this fast, relatively economical, and versatile printer was far enough along in development to provide an admirable answer to the printing requirements of the system.
much lower rental for much higher performance than three 407 accounting machines plus a 604 calculator. It is no exaggeration to say that the 1401 opened the world of electronic data processing for the first time to a broad range of small and medium-sized users of IBM's punched-card systems. It was not the first IBM computer to use transistors instead of tubes-the 7070 and the 7090 computers, as well as the 608 calculator, were announced and delivered ahead of it- but it was undoubtedly the product that gave IBM its first realistic glimpse of the size and importance of the computer market that was unfolding.
Oral History and collection of e-mails from Karl Ganzhorn
- Oral History, Part 1 4.4 megabytes .pdf
- Oral History, Part 2 3.7 megabytes .pdf
- from Karl Ganzhorn 07/29/2007 - Subject: IBM History
- from Karl Ganzhorn 08/02/2007 - 1401 History_2
- to Karl Ganzhorn from R. Garner - 1401 History_2Link
- from Karl Ganzhorn 08/07/2007 - History_3
- from Karl Ganzhorn 08/14/2007 - History_4
- (Research and Development in IBM Germany) Prof. Dr. Karl Ganzhorn
Karl.Ganzhorn AT t-onlineDOT de (Karl Ganzhorn) 07/29/2007 05:16 AM
To Robert B Garner/Almaden/IBM@IBMUS
Subject IBM 1401 History
thank you for your e-mail (address was correct, obviously). It is interesting to learn about the planned 50th anniversary of the 1401 announcement, and I compliment you about the efforts of recollecting the history of this break-through product.
From left to right:
All I can contribute is a photo of the initial WWAM task force in August 1955 in Boeblingen, Germany, attached as with some comments on the participants. It shows all the members involved except Jim Ingram (he probably took the picture, his chair is empty). - Paul Manieri, Director Marketing IBM Europe, Paris (Italian origin) - Walter P. Scharr, Director of Engineering, IBM Germany (he died long ago) - Bill C. Christensen, Product Planning IBM WTC, New York - Maurice Papo, Engineering IBM France - Joe B. Fernbach, IBM Endicott, Head of the WWAM Task Force - Karl E. Ganzhorn, Engineering IBM Germany - Ted Einsele, Engineering IBM Germany (later professor at Technical University of Munich) - Gene E. Estrems, Engineering IBM France (he died several years ago) - Ted Buley, IBM Poughkeepsie.
The picture is also contained in the first volume of a book series which I undertook/edited: "Research and Development in IBM Germany", 7 volumes, see attached flyer
Paul Lasewicz has a complete set of the book series in the IBM Archive. From Volume 1 the article in the "Annals of History of Computing", July 2004 is condensed.
Karl.Ganzhorn AT t-online DOT de (Karl Ganzhorn) 08/02/2007 10:48 AM
To Robert B Garner/Almaden/IBM@IBMUS
Subject 1401 History_2
Thank you for your prompt response. Attached I am sending a new BMP scan of the WWAM task force 8/1955 which might help with a better resolution. The original picture comes from my private archive, I don't remember the photographer, but I assume you may use it as appropriate. Regarding the task force there are of course many episodes as our guests from USA and France were assigned to Boeblingen for many weeks with weekends of pleasure and exploration of Germany.
The room ?shown in the picture was a large hall emptied by IBM Germany for the purposes of the task force. We equipped it with drafting boards, laboratory benches for demonstrations, and a large conference table.
Here are some episodes:
Some technical reminiscencies:
- During the hot summer afternoons we served of course drinks. Yet, in postwar Germany the ?bottles were not yet cooled. So our American friends asked for some ice. But in those days ice cubes were not available from refrigerators in Germany. Nevertheless I called the kitchen for help. Ten minutes later the head cook apeared with a commercial ice block on his shoulder 40x10x10 inches and a 3 pound hammer, put it on the conference table and disappeared. We had some extra work with much fun.
- One weekend Joe Fernbach drove to the village in the Black Forest from where his grandparents had emigated in their younger years to USA. On Monday morning he came back completely flabbergasted, saying: "Karl, con you imagine - I understood their language! I never knew that before." What had happened: As a young child he had his grandparents heard speaking their German dialect and had picked it up unknowingly.
- Maurice Papo, an extremely brillant French engineer, was quick-witted and proud that he had some knowledge of the German language from school. The group often went out for dinner and explored German gastronomy (which, by the way, was not too exciting those days). Maurice always insisted that he ordered his dinner in German. So, one evening his American colleagues quietly instructed the waiter: "No matter what Mr. Papo orders - you bring him fried eggs!" Maurice saw that and immedeately responded: "See the problem wit these Germans: They don't understand their own language!"
- A story of typical IBM friendship: One morning I had to call and excuse me as my wife was in alarming condition having a blood poisoning in her throat and needing immediate treatment in a special clinic in Stuttgart. Ted Buley, knowing that I had no car yet in those days, immediately jumped in his rented car, rushed to my home and took my wife to Stuttgart breaking all speed limits. She was saved in the last minute. Thank you, Ted, again!
In 1955 we investigated three different approaches for an electronic accounting system: The American and the German concepts were both preoccupied by the large computer systems like the IBM 701: One address memory and consequently equipped with an accumulator. The French approach (for which Estrems and Papo later received high patent awards) was based on a two- address concept, using the second memory address as accumulator - similar to the relay accounting machines like the IBM 407. On top, a variable word length would be foreseen where the word length was determined by wiring on a plugboard. All three systems had the plugboard for memory definition and control. A year later, Underwood in Endicott, upon request of R.G. Mork, in a new systems study recognized that the necessary electronic drivers for the large plugboard were the prohibitive cost source. In 1956 he replaced the plugboard by stored program control. Together with the French two-address variable word length concept and a magnetically stored word mark in an extra plane of a ferrite core memory the costs could be brought down to what became the 1401. While the initial cost objectives could not fully be achieved the 1401 nevertheless found a huge market, admittedly not for the originally aimed smaller customers for which Jonny Dayger's 1403 printer was too powerful and expensive, but the only one available. This was solved with the follow-on systems IBM 1430, IBM /360 Model 20 and /360 Model 15 later.
The WWAM task force continued in Poughkeepsie starting in September 1955 under Peter DeGeorge as Manager exploring the three approaches further. In 1956 the Corporation decided that this program should get highest priority, benefit from all resources available and be "put in best experienced hands", as Vin Learson stated it. But this is all documented in Ch. Bashe`s and E.W. Pugh's "IBM Early Computers".
May I stop here threshing old memoirs.
Best wishes for your undertakings.
To Karl.Ganzhorn AT t-online DOT de (Karl Ganzhorn)
Subject Re: 1401 History_2Link
Thanks for your special remembrances of the early development of the 1401 architecture! This early history is wonderful, fun, and interesting and what I was seeking so we can better understand and interpret the machine.
May I ask: ?Can we please post your 1401/WWAM history and your anecdotes on our 1401 web site? ? http://ibm-1401.info/index.html (probably in a new section "Early 1401 Design History"). ?We have excerpts from Pugh's IBM history book there.
Some questions (I hope not too many!), if you may remember so far back in time...
What was the origin of the task force's name "World Wide Accounting Machine"? ?Were there high-level goals and marketplace motivators for an international machine that came from Watson/New York? ? Were there aspects of IBM's unit record accounting machines that were a problem for world markets? ?Or was the task force's genesis an effort of the new Boeblingen lab to get involved in IBM's mainstream machine development? ? How/why did the French members get invited?
> The French approach (for which Estrems and Papo later received high
> patent awards) was based on a two-address concept.
Were there any existing computers (possibly paper designs) in 1955 with two memory addresses per instruction that the French or WWAM team were aware of?
> All three systems had the plugboard for memory definition and control.
I'm still grappling with how a plugboard could would be programmed to perform all the functions necessary for a WWAM. ? Was it even larger than the 407's? ?Were these plugboard designs really workable solutions, or, in addition to the costly electrical problem interfacing to the plugboard, was it ever deemed to be too limiting a design constraint by members of the WWAM group?
> In 1956 Underwood replaced the plugboard by stored program control.
Was there anyone else on the WWAM team who had wanted stored program control? ?Was this a controversial topic? ?(It was interesting that the RAMAC did have a plug board, somewhat of a hindrance and a relic in my mind.)
What were the most controversial topics between WWAM members?
(besides German cuisine ?;-)
Any controversy on decimal addressing and arithmetic? ?Was there any debate of doing binary, or did decimal = business, no discussion? ? What drove variable-length word vs. fixed width? ?Was there any info on how much memory variable length words might save? ?Or just a gut sense it was cheaper (by who)?
> While the initial cost objectives could not fully be achieved the
> 1401 nevertheless found a huge market, admittedly not for the
> originally aimed smaller customers...
Why do think this happened? ?(Perhaps stored programmed architecture at a good price allowed for explosion in application program development?)
Do you know how we might be able to find any other surviving WWAM members?
My apologies for all the questions, and thanks for all your help and stories!!!
IBM Almaden Research Center, San Jose, CA
[attachment "WWAMTskF55.bmp" deleted by Robert B Garner/Almaden/IBM]
Karl.Ganzhorn AT t-online DOT de (Karl Ganzhorn) 08/07/2007 05:18 AM
To Robert B Garner/Almaden/IBM@IBMUS
Subject 1401 History_3
attached please find my comments to your questions
---------------Converted Carner809.doc to HTML, from Karl Ganzhorn -------------------
Prof. Dr.rer.nat. Dr.-Ing. E.h.
Gluckstr. 1, 71065 Sindelfingen
Tel.: 07031-812022 /Fax:-812024
e-mail:Karl.Ganzhorn AT t-online DOT de
Robert B. Garner
IBM Almaden Research Center
San Jose, CA
U S A
August 6, 2007.
Subject: 1401 History_3.
Thank you for your e-mail of August 6. Here are some answers to your questions:
You may use them for publications except the one on my wife?s illness.
- Origin of the name WWAM (needs a lengthy explanation):
In the early 50ies IBM?s Unit Record product line was different in USA and WTC due to the war separation. But the need for faster calculation and printing was recognized in both areas. During and after World War II the development of Unit Record punched card machinery had followed different routes: In France J. Ghertman and his engineers had designed calculator punches like the 626 and especially a new accounting machine Type 421, functionally similar to the 407, all in relay technology with plugboards for wireable data flow, storage allocation in electromechanical counters and control functions. In Germany a series of special features for the 421 accounting machine were designed.
IN 1951 the Corporation had encouraged several European IBM companies (France, Germany and United Kingdom) to establish engineering competencies in electronics. In France J. Ghertman, in Germany Walter P. Scharr and in UK J. Elliot (an outside hire) were implementing this directive.
By 1954 it was recognized in USA, in France and in Germany independently of each other that electronics had to be investigated also for use in Unit Record machines, replacing the clumsy relay technology. In USA several engineers around Jim Ingram became active, in France John Ghertman and his excellent engineers were searching ways and in Germany I had been given the task to seek ways for introducing electronics and physics into IBM?s product line technologies.
In early 1955 the Corporation set an objective for a common future product line for the Unit Record markets in the world. In the following early deliberations the name of "World Wide Accounting Machine" emanated.
For the first time a multinational engineering task force was conceived, the WWAM task force, with participants from USA, France and Germany. Walter Scharr, head of the German engineering department, was asked to host the first workshop of the WWAM task Force (about 6 weeks) in Boeblingen, Germany. Both Watson brothers, Tom and Arthur, took an active interest.
The strongest market pressure emanated in France from the national competitor Bull who had an electronic machine out in the European market.
- Two Address Concept:
In the early days of computers 1-, 2- and 3-Address schemes were investigated all over the world. The most frequent calculation in Unit Record applications consisted of a card stack with various information on each card. The stack was sorted in subgroups. Each type of information had to be added to an individual counter (accumulator). I.e. a varying number of accumulators was required to which the respective numeric information could be added/subtracted. This represents a multiple two-address operation. The French idea was to establish any storage cell to act as an accumulator to which the contents of an input word could be added. This way multiple accumulator registers could be avoided. Calculation went directly from the card input to memory. (Today every pocket calculator has this function, namely "memory +"). Estrems and Papo just combined this calculator function with a variable word length in memory allocation. (By the way, 8 years later in /360 one of the machine instruction formats represents exactly the 2-address format required for the "Add to memory" operation.)
- Plugboard Control:
The plugboard used to be the generally accepted means of control in all Unit Record machines (405, 407, 421, the earlier German D11, etc.). By plugging wires the information flow, the memory allocation and all functional control signals were determined. In the early design deliberations for WWAM this concept was never even questioned. It was Underwoods independent mind touching this holy cow in 1956/57.
- Stored Program control:
The WWAM concept also had in mind that for the world-wide customer community the established way of handling control by plugging wires on a plugboard should be maintained. The 1401 finally took away with this, based on Underwood?s design proposal, which meant a breakthrough in the Unit Record world.
- Controversial issues in WWAM conception?:
Decimal arithmetic was never questioned for the wide spread accounting business.
Fixed versus variable word length was a long debate. Primary reason for variable word length: Saving expensive magnetic core memory capacity. (Even three years later the 1401 came out with a lowest cost memory version of 1400 characters!).
- Cost objectives:
The costs of electronics technology with ferrite core memory and stored program control were still barely acceptable for the Unit Record market in the mid 1950ies.. Its low cost end actually could not be reached with the 1401, particularly in Europe. But thanks to the booming high end Unit Record market in USA this did not matter. Instead, a low cost version 1430 was planned to be developed but this submerged in the /360 design processes for several years. Eventually the /360 Model 15 and later the /3 took care of this.
- Other WWAM members?:
Here are two e-mail addresses:
m.papo AT ieee DOT org
rgmork AT ibm DOT net (this is an old address. Ralph Mork recently did no longer maintain our former connection via e-mail for unknown reasons. If he responds he may also know about Underwood).
Re address of Underwood: he might be a primary source of 1401 info. Maybe E.W. Pugh might know).
I hope this helps you with some of your questions.
Karl.Ganzhorn AT t-online DOT de (Karl Ganzhorn) 08/14/2007 05:17 AM
To Robert B Garner/Almaden/IBM@IBMUS
Subject 1401 History_4
Thank you for your extensive e-mail information on the fresh-up of 1401 history and design.
I am afraid that I am not really the right person to answer all your questions around the 1401 design. The reason: The WWAM task force comprised two consecutive phases:
The German Lab had been in the middle of this compatibility battle 1401- versus /360- architcture which was taking place in 1963 between GPD Endicott and DSD Poughkeepsie. It escaped by a tricky processor design: Its new processor had all functions implemented exclusively in microprogram control stored in a hardware read-only store (stored function concept). Thus, when the decision was taken to go /360-compatible early 1964 only this hardware control store had to be exchanged in order to switch to the other architecture. It took only about 6 weeks redesign of the hardware control store, yet long enough that the product could not make it for the /360 great announcement in April 1964. So, only in November 1964 the /36O compatible product could be announced as the /360 Model 20 (which in the following years reached over 15 000 systems shipped world wide). But then, the struggling for lower cost continued, and Boeblingen developed a /360 Model 15 which came out later.
- August-December 1955 in Boeblingen and shifting in September to Poughkeepsie with all initial participants.
- January - Summer 1956 in USA with only E.E. Estrems (Gene) and M. Papo from France still involved from Europe. The task force ended in summer 1956, when Underwood's stored program solution was accepted. Already after phase 1 (December 1955) the German Lab's electronic department was no longer involved in the development efforts which resulted in the 1401, while the printer developments went on. Instead it immediately engaged in the studies for a new low cost electronic machine for the low end Unit Record market. Focus was on a low cost printer and a low cost stored program machine for several years, aimed at a "1430 product" which finally was given up in the final battle 1401- versus /360-compatibility in 1963. So, the 1430 (the GPD/Endicott preference until 1963) was never anounced and given up when the /360 systems concept (DSD/Poughkeepsie) prevailed in January/February 1964. (I don't recall why the 1430 product number was in use while the product was not announced yet?)
Regarding 1401 alias: As mentioned above I am not the right person for this, as after WWAM I had turned to other areas in building and managing the German Lab in 1956. Your planned paper on the 1401 history would of course be of interest to me and I gladly would review it for the part in which I have personal memories.
Regarding 626 and 421: Both machines were IBM ?badged products developed in France for the WTC market after the war. The 626 was an electronic calculator card punch (electron tubes) and the 421 was the analogon to the 407 accounting machine with electromechanical technology. The printer and counters of the 421 were different from 407's.
Buchreihe: Forschung und Entwicklung in der IBM Deutschland
(Research and Development in IBM Germany)
Organisation/ Herausgeber: Prof. Dr. Karl Ganzhorn
Inhalte (Contents): → ./.
Research and Development in IBM Germany
Organization/Editor: Prof. Dr. Karl Ganzhorn
Volume 1: The IBM Laboratories Boeblingen - Foundation and Build-up (K.E. Ganzhorn)
- EarlyYears of Electronic R&D
- Solid State Design -Technologies and Systems
- The Last Punched Card Systems
- Exploration of Alternative Technologies
- Consolidation & Missions of WTC Laboratories
- The IBM Science Group Vienna
- Implementation of the Mission
- Product Development
- External Relations
- Transition and Management Change.
(z.Zt. vergriffen/sold out)
Volume 2: Die IBM Laboratorien Böblingen: System-Software-Entwicklung (A. Endres)
- Software für das System /360 Model 20
- Übersetzer für Formelsprachen -.
- Weltweite Software-Verantwortung mit LEOS
- Software-Entwicklung im DOS/VS Verbund -
- Konsolidierung der VSE Entwicklung
- Systemnutzbarkeit mit SSX-8. UNIX als offenes System
- Bestandspflege & neuere VSE-Systeme
- Komponenten für MVS- Systeme.
(z.Zt. vergriffen/sold out)
Volume 3: The Heidelberg Science Center: User Oriented Informatics and Computers in Science (A. Blaser)
- Introduction - HDSC Mission, Strategy and Research Areas -
- Chronological Overview of Some Milestones and Highlights
- Early Application Research - The Motivation for User Oriented Informatics
- Languages and Styles of Dialogue - The Obvious Features of System Usability
- . System Structure and System Components - Decisive Factors for Usability -
- Scientific Computing and Problem Solving
- 7. Summary and Acknowledgement
- HDSC Staff Members and Visiting Scientists.
(Preis: 25 ?)
Volume 4: Die IBM Laboratorien Böblingen - System-Entwicklung (H. Painke)
Aufbaujahre (1952 - 62)
- Die /360 Systeme aus Böblingen (1962-68)
- Die bipolaren /370 Systeme (1969 - 80) - Zeit des Umbruchs (1980-86)
- Durchbruch einer neuen Technologie (1984-88)
- Die Enterprise 9000 Systeme (1986-96)
- Neuere Zeit bis Gegenwart (1997 - 2003)
- Persönliche Erinnerungen (zu jedem Kapitel)
- Persönliches Nachwort.
(Preis: 29 ?)
Volume 5: The IBM Laboratories Boeblingen - Semiconductor and Chip Development
(H. E. Barsuhn/ K.E. Ganzhorn)
- 1. Prehistory
- Semiconductor Component Mission for Boeblingen
- Boeblingen?s Entry into Semiconductor Technology
- Bipolar Logic and Memory Chips
- MOSFET Memory Chips -
- CMOS Microprocessors for System /390
- Chip Development in a Global Environment -
- A Retrospective on various Facets of Chip Design
(Preis: 39 ?)
Volume 6: IBM Informationstechnik für Banken und und Sparkassen im 20. Jahrhundert (Editor: K. Ganzhorn)
Teil I: Wechselwirkungen zwischen Bedarf und Technik
- Von handschriftlicher Kontenführung zur Elektronik
- Neue Basis der Computertechnik: IBM/360 Weltstandard, Tor zu eigenständigen Software- & Geräte-Industrien
- Vollelektronische Dokument- Verarbeitung
- Neue Technologien und ihre Auswirkungen
- Online-Service, Kundennähe und Netzwerke
- Selbstbedienung als Instrument der Bankenstrategie
- e-Banking - Internet-Banking -
- Börsen- und Handelssysteme - Whole Sale Banking
- Chronologische Entwicklungen in der Kreditwirtschaft
- Zeittafeln der Informationstechnik in der Kreditwirtschaft
Teil II: Verfahren und Technologien
- Depot-Geschäftsvorgänge - Bearbeitung mit Magnetband-Anlage
- Leistungsorientierte Computer- Einrichtungen
- Ordnungsprozesse in der deutschen Kreditwirtschaft
- Online- und Datenfernverarbeitung
(Preis: 35 ?)
Volume X: Research and Development in IBM - History of Far Eastern Languages in Computing (K. Hensch)
Introduction - Prologue - Requirements, Phonetic Solutions in the 1960s - Characters Used in Far Eastern
Countries - Technologies for Handling Characters - Reduced Phonetic Character Set Implementation, 1960s -
Initial Efforts, Full Kanji Solution in Japan, Early 1970s - First Japan Kanji Capability Realization, Early
1970s - Newspaper Publishing Systems in Japan - IBM Japan Taking the Lead, Achievem. through 1990s -
East Asia Coded Character Set Standardization - Major Changes through New Technologies - Epilogue
(Preis: 39 ?)
Bezugsquelle (Order Placement):
Haus zur Geschichte der IBM Datenverarbeitung
E-mail: historyh AT de DOT ibm DOT com
Oral History and e-mail from MauricePapo 2008 March 27
Oral History 5.0 megabytes .pdf
The 1401 has been such a major event in my IBM life that I have too many recollections including pictures.. and I don't know where to start! Unlike Karl, I don't have the time or the motivation to embark in the task of writing books about it!
Let me try for a beginning to share some elements for thoughts
We will probably have some more exchanges in the future.. but I hope that the above may help as a beginning.
- The initial project, and the resulting competition between the 3 IBM labs came from a visit of the IBM France President Christian de Waldner to Watson complaining about the major inroads that the French computer company Bull was doing to IBM, mainly in tne banking areas for lack of IBM answers.. The competition decision was then taken at the highest IBM level.
- I do have, but so do you certainly, many 1401 and 1400 related patents bearing my name
- Back in 2002, there was a lot of Email exchanges coordinated by Van Snyder <vsnyder AT math DOT jpl DOT nasa DOT gov> on the subject of 1401 and "1401 nostalgia". If you have not already done so, you may contact him
- Many years ago (I was still active with IBM) I was asked to audio record my recollection of the 1401 history for an IBM "Oral history program". I am sure this should be readily accessible for you
- Concerning the "plug board approach" of the WWAM (which incidentally was also referred to at its beginning as VLAM: Variable Word Length Accounting Machine) we started with a stored memory program but found out that we could not afford the cost of memory at the beginning of the project. This is why we reverted to a plug board approach, trying to make it as compatible as possible with the stored program approach... until the cost of memory went down and the US team was able to considerably improve the project by going to a stored program...
- The WWAM wanted also to "prove" a new packaging "card on card" approach. The initial WWAM model was built this way, and I have of course numerous pictures of this design and the implementation team.
- 1401 was such a commercial success that IBM managed to have its private jet registered under the 1401 name. Among my many "old" pictures I have one of the jet landed in the Nice Airport, showing its registration name and the official visit (including Tod Groo IBM VP at the time)
- Among the WWAM team which worked in Poughkeepsie, the "chief product planner" was Jacques Maisonrouge well known who later became President of IBM Europe and Vice-President of IBM Corp
- Thre are still a number of engineers and techniciens who participated in the design and build up of the initial WWAM in the IBM La Gaude laboratory, still alive in the area.
- The intial WWAM model has been used in the La Gaude Laboratory for quite some time as part of the computing center. There was even a project to move it to the USA for patent purposes... We tried to preserve it... but I don't know what was its final fate
At 07:25 26/03/08, Robert Garner wrote:
I've learned through exchanging emails with Karl Ganzhorn about your involvement in the mid 1950's in IBM's WWAM task force.
I am working with a group of volunteers at the Computer History Museum in Mountain View restoring an IBM 1401 to complete operation.
May I ask if you might be interested or able to in share any of your stories, recollections or history of the WWAM / 1401 project?
p.s. I'm also trying to locate Francis Underwood, principle designer of the 1401 in Endicott.
And Ralph Mork, or any other surviving members of the WWAM / 1401 team.
Past Vice President Regional Activities
14, Bd Prince de Galles
06000 Nice, France
Email: m.papo DOT ieee DOT org
Phone: (33) 418.104.22.168
Fax: (33) 422.214.171.124
A collection of e-mails working to find answers
From Gary Mikotoff 4/4/2008
We have contacted Jack Palmer. His e-mail address is Jack Palmer [jackpalm AT Charter DOT net]. Part of the e-mail he sent Dave Macklin and me stated:
"But (1) his name was Francis O. Underwood, and I seem to recall that others have tried to find him in recent years, unsuccessfully. And (2) don't hesitate to communicate with me, particularly concerning am apparent 1401 COBOL inquiry. And (3) I remember an engineer Glen Nielsen in San Jose, I think a manager, I think on the 1620 project, no 1401 connection that I ever heard of. And (4) I remember Gary sending the 1403 in the T/L basement into a paper-consuming paroxysm, via a no-escape null print loop that had it executing endless skip-to-top-of page instructions. And (5), tying into Gary's story, I remember that the first index register arithmetic did not wrap around, making it impossible to decrement IR contents. They quickly redesigned it on our recommendation to wrap around at 16K, but, as always, it took forever to get our own machine fixed.
And finally, Dave, Gary's memory seems to be better than yours (and why not?!), for you have slightly garbled the history thing via capitalization and lack of mention of co-authors, etc., but you get credit for coming up with my e-mail address!
With best regards,
Q from Robert Garner, Answer by Maurice Papo 4/5/2008
I'm getting the impression that the 1401 was marketed fairly seriously across Europe.. (I'll forward pic of 1401 "datamobile" in Denmark and 1401 truck unloading in Finland.) YES but not only in Europe. The 1401 sales World Wide (!) including the US were more than any other IBM computer until the PC came along! I kept the sales figures for a while, and don't remember throwing them away... but god knows where they are! Along the years, I met many 1401 customers who remembered it with nostalgia.. Remember that the project started initially from tough competition by Bull Company which operated only in Europe, but it quickly found its market value in the US.
From an e-mail Garner to Branscomb dated July 16, 2013
Thanks for your gracious response to my questions (copied below).
This better explains the relationship between Dayger's tape-to-print project, SPACE, and your involvement!
I assume Bo was demanding tape attach in order that SPACE could be used as an I/O spooler for the 7000 series mainframes?*
> Things came together as I recall in the late spring of 1959 when I was satisfied and we had some big �guns� like Bob Evans demanding tape be included with the initial announcement.
Or was he presaging that tape would increase SPACE sales?
According to Shel, no-one realized that was going to happen; i.e., it was a surprise that so many customers (eventually over 50%) bought tape systems.
Besides the 1403, which items did you consider new?
> We set a rather aggressive schedule considering the fact that almost everything in the system was new.
From my perspective,
-the Germanium transistors and CDTL circuits were proven from the 7070,
-and SPACE's data paths were taken from the WWAM.
-Also, the tape controller circuits were borrowed from the 7070.
-(The magnetic core memory transformer X/Y driver was relatively new.)
That would be great! I'll be meeting with Shel and the woman responsible for the movie script, Jas, next Weds.
> Regarding your question about recording a brief piece on the 1401 goals and the overall Space program, I would be receptive if you feel it would be helpful.
I'll get back to you.
* Perusing a recent issue of the 1964 ACM Proceedings, a practical article, "The STL Integrated Computer Operating System" by TRW, pointed out that their average 7090 compute job took only 3.6 minutes (with 80% completed under 5 minutes), but then each job required an average 4.1 hours on their peripheral 1401s for tape-to-print and tape-to-card, due to the operator sequentially mounting tapes, etc.. (Their configuration was actually two 7090s and six 1401s.) By incorporating a 1301 disk with a custom OS using "job information blocks" to act as a I/O spooler buffer, the overall job throughput dropped to just 1.6 hours, which effectively could allow for one fewer 7094 (if the load remained constant)! Illustrates the importance of balancing computer and I/O, particularly during that era when CPU cycles were so damn expensive. I'm sure this is one reason that Fred insisted that all mid- to high-end 360s have a disk!
The following enclosed item may be Branscomb to Garner, dated ?
Sorry I overlooked responding to your question about goals for the 1401 program. Actually, I think you have a good understanding of our major goals and later I will comment on why I think Jonie Dayger called our goals �lofty�.
Stated simply, our goal was to move the thousands of �plugboard� (control panel) controlled customers to stored program control. To do that we had to focus upon several major areas:
FUNCTION/COST: For the larger customers, we wanted to make sure we could handle essentially all the applications that a customer with multiple accounting machines and a 604 could run. At the same time, we knew that entry price for the smaller customers was crucial so we established cost targets that would provide a complete small system for less than $2500/month. We were determined to meet this system target - not because it was necessary to have a successful product but because we wanted to move the maximum number of customers to stored program control.
SIMPLICITY: We realized that just function and cost would not move all those customers to a new totally different system. So we did whatever we could to make the 1401 a system people from the old system environment could relate to. Most early computers were �fixed word�. �Variable field length� in our system helped unit record people relate to it and very importantly gave us much more efficient use of memory which was crucial to making our cost targets. Our architecture also provided a very simple, easily understood instruction set.
PRINTING: This was and is a very important function in business data processing. �Lines per minute� always draws a lot of focus but print editing is also an important function and the 1401 was excellent here as well. The 1403 had excellent paper handling which significantly improve throughput in printing applications.
SCHEDULE: We set a rather aggressive schedule considering the fact that almost everything in the system was new.
TAPE TO PRINT: Jonie Dayger had a �tape to print� product under development when the 1401 was begun. Since there were only several hundred potential customers, his forecast was relatively small. This meant that his target rentals were quite high but that was not a problem in the tape to print market. Since we were pursuing large volumes, our target rentals were much more attractive. As a matter of fact, my recollection is that a tape to print version of the 1401 had a rental target lower than Jonie�s product even though it provided much more function. I think this is one reason for Jonie�s �lofty goals� comments. I think the other reason is that there had been several attempts to develop a new system to replace the IBM 402/405/407 and all had failed. The Transcriber (my first assignment in IBM) effort was terminated in about 1953. The MAC proposed out of Poky identified their low end machine and gave very little detail but did note that it would probably be a control panel machine since they could not afford stored program. I think at one point there was proposed a TAM (Transistorized Accounting Machine) but never really got off the ground. And of course the WWAM program was in great difficulty and was terminated with the acceptance of the 1401 into the product plan. So Jonie had some basis for his view and there were others skeptical that we could meet our objectives.
Even though Fran Underwood had architected attachment of tape and disks to the 1401, I refused to incorporate tape into the initial announcement until satisfied that we were going to meet our fundamental objectives. Things came together as I recall in the late spring of 1959 when I was satisfied and we had some big �guns� like Bob Evans demanding tape be included with the initial announcement.
I had a chuckle from your comment about �adopting or stealing� the 1403. Jonie very much wanted the big volume for the 1403 that would come from the 1401 but he wanted more urgently to get his printer into the market and he had a product that would do it. I felt he and I pretty much saw things in the same manner. I did not want to get diverted to tape prematurely and he did not want to get diverted from his product prematurely. By the way, I�m quite sure that Delco developed and made a special high voltage print magnet driver for the 1403.
From: Robert Garner to Charles Branscomb, June 16, 2013
From: Robert Garner
Subject: Re: Jonie's tape-to-print vs. 1401 and a new 1401 movie (was Re: IBM Corp Market Research's "67 Years of Progress" chart, 1960, missing the 1401 !
Date: June 16, 2013 11:48:05 AM PDT
To: Charles Branscomb
Good to hear from you, as always!
>> But Jonie was not convinced that we would meet what he called our "lofty" goals and he continued his product until the spring or summer of 1959.
My understanding had been that you had convinced Jonie to drop his standalone tape-to-print product, thereby deferring the tape-to-print market to SPACE. If he had been comfortable splitting the market (or perhaps he wasn't, and you outright "stole"/adopted the 1403 from him), he clearly must not have been convinced that SPACE was going to be successful. (Another indicator that you, Fran, and Shel were spot on in your predictions of how successful SPACE would be!)
Do you recall your "lofty" goals for SPACE?
>> But Jonie was not convinced that we would meet what he called our "lofty" goals and he continued his product until the spring or summer of 1959.
Do you recall what finally swayed him (or someone ordered him) to drop his tape-to-print machine in the "spring or summer of 1959".
Did his machine use the same high-power hammer driver transistors from Delco? (I assume he didn't get the transistor waver either.)
I understand that Delco had developed them (or perhaps the high-voltage tolerance) for electronic automotive spark plug distributors.
Do you recall if they make a special power transistor tailored for the 1403?
Best Wishes and Happy Father's Day!
On Jun 16, 2013, at 10:41 AM, Charles Branscomb wrote:
>> My recollection is that we did announce Model D for those customers that only wanted "tape to print" but I do not know how many were ordered without the 1402.
>> Also, our recollections are a little different about Jonie Dayger's program for a tape to print product. When his team initially could not find a solid state driver for the print magnet drivers, we joined them in appealing to Ralph Palmer (director of engineering for our division) for an exception to the edict. Ralph refused and then we joined Jonie's team in pursuing an outside vendor solution which fortunately was successful. But Jonie was not convinced that we would meet what he called our "lofty" goals and he continued his product until the spring or summer of 1959. It is unfortunate that Fred Demer (main inventor of the 1403 printing concept) and Jonie did not receive more recognition for a product that was a profound advance in computer output printing.
On Jul 21, 2013, at 1:31 PM, Charles Branscomb wrote to Robert Garner
The IEEE article is an excellent, informative, articulation of the early computers. Congratulations.
Im sure that Bob Evans push for tape with the initial 1401 announcement was to help the 7000 series. More specifically, the 7070 was in real trouble and Bob was moved up to Endicott to bail it out.
My thoughts on what was new in the 1401 program may be somewhat different from yours. My thinking starts with the people in Engineering, Manufacturing, and Field Service and their lack of experience in dealing with anything like the 1401. My guess is that three fourths of them had never been exposed to a stored program system, solid state technology, and the packaging, power supplies, cabling, etc. that support low level signals. The projected production volumes of the 1401 presented a severe training challenge in manufacturing and field service. The lack of experience showed itself in many ways. For example, we had great debates with manufacturing about product costs and with field people about service costs. Both of these had real effect on meeting our rental targets and frankly, we never did get what we thought were proper service estimates because of their inexperience with something like the 1401.
Back to engineering and people. Start with me - I had taken 2 or 3 night school electronic courses but I had zero experience in stored programming and solid state technology. Most of our logic designers had some tube circuit experience but not solid state. Also, we of course had essentially no design automation in those days to help identify timing races, hot spots, etc. In view of our focus on cost and our volume, we worked with the circuit designers on special SMS cards rather than just use the standard cards. I wish I could recall the specifics about power supplies, for example. I feel quite sure that we were not satisfied with the "standard" supplies and pursued supplies specific to us that were more reliable and cost effective. It is clear some of the internal data paths were from WWAM but I think you would find the input/output was handled quite differently and of course variable field length was handled with word marks. Further, implementation of the data paths in the most reliable and efficient manner was much more crucial to me. By ther way, I thought I was told by the memory people that the way they were able to meet our cost objectives was with new X/Y drivers.
I will not continue to "ramble" but will mention just one other example. It was the very early days of back panel wiring with Gardner-Denver machines and this was a real concern for some of us. To make changes in the field, service people had to use a manual tool and that led to some troubles.
Its difficult to put myself back into the second half of the 50's but I'm comfortable in saying that almost everything in the system was new to a large percerntage of the people who had to design, manufacture and service the 1401.
Regarding tape on the 1401 itself, we certainly did not focus upon a tape processing system. We did know that the 1401 was excellent at card to tape and tape to print functions. I am still not convinced that the 1401(not the 1460,1410, or 7010) with no operating system was an adequate tape processing system. My guess still is that most 1401's with tape were used for I/O operations but of course I do not have any data to support that view.
Two IBM 1401 Principle Founders found!!
from Robert Garner, April 16, 2008
I have some auspicious news to report: After several years of on/off searching, I've located the 1401's lead designer and "father", Mr. Francis O Underwood! He's living in Lockhart, TX. In addition, at Fran's urging, I soon found the 1401's lead marketing/planner, Sheldon Jacobs. He's closer, in Saratoga, CA. Both are now on our 1401_team alias and are cc'd on this email.
Fran is just delighted that we're interested in the seminal 1401, have restored one for the Museum/public and considers it an honor to be able to help us out. He has delightful, insightful stories to tell, and hopefully we'll get him on the phone soon. (btw, He also worked on the 1130 and 1800.) Shel is also quite excited about our project and is looking forward to participating and sharing his many period stories.
I've also been in email contact with two original members of IBM's World Wide Accounting Machine (WWAM) program that preceded the 1401/SPACE program: Maurice Papo (in Nice, France) and Karl Ganzhorn (in Germany), also cc'd on this email. (They went on to lead IBM's respective research labs in Europe.) Maurice has shared several fascinating 1401 pictures, including the "1401 plane" and a wooden model of a 1401 system (given to him by his staff): http://ed-thelen.org/1401Project/OutsidePhotos.html
All four are just delighted that we have so much interest in the 1401 and are honored to be associated with our restoration project. Fran and Shel are brimming over with 1401-era stories to share. What follows is my synopsis of some Shel shared on Thursday on the phone, and some from my correspondence with Fran. (Which we'll eventually record/put on site. See this link for some fascinating/incredible Fran narratives: http://www.ed-thelen.org/comp-hist/stories.html#Fran
First, from reading the oral histories and discussions with Fran & Shel, I've come to understand why 1401 sales literally exploded and also learned something new - that Fran & Shel had to overcome some significant early doubts about the viability of the 1401 in the business marketplace.
Although we take stored program computers for granted, several "status quo" factors were obscuring potential progress in the market: 1) the predominance of feature-ladened plug-board accounting machines, 2) certain Endicott planners not understanding the power of combining computing and card handling into one machine, and 3) little appreciation of the resulting software ecosystem waiting to be created.
It took Fran's vision of the suitability of a stored-program architecture for the high-volume, business market (and a particular circuit designer reducing the cost of core and other fortuitous events from engineering, e.g., SMS, chain printer, and 088) and Shel's doggedness to get his 1401 volume estimates (validated by his customer visits) accepted by Endicott planners that finally got the 1401 across its chasm. These factors all converged at the same time as a glaring market vacuum/black hole for an inexpensive stored-program machine had opened up. The 1401 just sweep away the unit-record accounting market. It was the market's first high-volume, stored-program computer.
This account from Fran is stirring:
About a week before the product announcement of the 1401
Tom Watson actually came to my office seeking some
confirmation about the current sales forecast. I could see
that he was finding it difficult to believe the 3500 unit
projection. I told him that not only was it true, but we actually
expected far more sales; at least 10000. I could see his eyes sparkle
with the thought of so much revenue.
There was just an enormous market demand for the 1401: According to Shel, IBM sales booked the entire five-year 1401 planning forecast--5,100 systems--in the first five WEEKS! Fran mentioned in his 1968 oral history that they booked 750 systems alone in the first week alone. Shel said that the eventual run rate for the 1401 was $30 million per month. If that happened in say 1966, in today's 6.5x inflated dollars that would be equivalent to $200 million per month!
My interpretation: If one assumes that a typical 1401 system rented for say $3,000 a month (Shel?), using a ratio of about 50x between monthly rental and purchase cost,* and 6.5x to get today's dollars (that's a 325x combined multiplier!), those 5,100 systems, if purchased, would have been come to $5 billion worth of sales in today's dollars! That's one billion a week for a complex computing system. (Of course, IBM likely earned more renting.)
The 1401 was so successful for IBM that Fran wrote that five key deserving members (and spouses) of the team were later granted three-week, first-class, fully-paid vacations to anywhere in the world! (Don't get any ideas! ;-)
Regarding the 1401 name, Shel said that the 1401 was NOT named after is minimum memory size, 1400 characters. Apparently IBM marketing had decided to grow to 4-digit names and "1401" was the first 4-digit designation, arbitrarily chosen. Shel had originally suggested the "450" (perhaps a merge of 400-series tabulators, 403/407, and the 650, since many saw the 1401 as a follow-on to the 407/650/604 market. Shel? Of course, Fran's internal project name, SPACE, (Sputnik-inspired) would have better, as sales rocketed there. ;-) I believe Shel said 15,000 total 1401's were sold. (Shel?)
More on 1400 characters: Fran hand-coded the "French banking application" in 1400 characters. (It was known as "le plus grand decouvert", handled well by the existing Bull Gamma III that was outselling IBM accounting machines in Europe, according to Maurice). Shel also had a sharp young engineer code up in 1400 characters an inventory/billing/commission application for a NJ correspondence school. Nevertheless, Shel noted that "almost no" 1,400-character machines were sold, as we've understood. 1400 characters seemed to be a demonstration that the pricing-sheet, entry-level $2,500/month config could handle any unit-record/accounting job, and cheaper.
More to come...
p.s. I also have copies of early 1401 and WWAM design documents and oral history transcriptions taken in the late 60's (thanks to the IBM archivists, Paul Lasewicz and Dawn Stafford.)
* Here's where I saw the rental vs. purchase cost figures from for the 1401: http://ed-thelen.org/comp-hist/BRL61-ibm1401.htmlCOST, PRICE AND RENTAL RATES Manufacturer Monthly Basic System Cost Rental IBM 1401 Processing Unit, $ 70,500 $1,200 Model A-1 IBM 1402 Card Read Punch, Model 1 24,800 550 IBM 1403 Printer, Model 1 30,300 725 Total $125,600 $2,475 Additional Equipment IBM 729 II Magnetic Tape unit $ 27,500 $ 700 IBM 729 IV Magnetic Tape Unit 59,000 1,100 (maximum of 6 tape units) Tape Input-Output Adapter Feature 22,700 980 IBM 1406 (Additional Core) Stor- 20,100 to 575 to age (3 models) 55,100 1,575 Maintenance contracts available for purchased or rented equipment.Which also indicates delivery time was 2 years!
inflation calculator: http://www.westegg.com/inflation/
WWAM Photos from MauricePapo - April 23, 2008 & Dec 19,2014
Most of the views are self evident.. The WWAM_all image ?[WWAM_Reader]? includes the connexions to the card punch and the reader printer which were required to answer the unique (?) European requiements for the whole system.
MauriceRobert, Here are a few more WWAM pictures.. One shows more clearly the plug board wiring of an actual application which proves that the amount of pluggable wires is more than reasonable The other shows what we used to call "yellow wires" for changes to the printed circuit board Maurice I think one is the WWAM core memory but I am not sure.. The second is just anothr view of the WWAM open. The last one is important to understand how the "large printed cards" swing out for servicing.
An earlier French Lab product - the 421 accounting machine
- After announcement:
- Status Symbol
I was just reading Karl Ganzhorn's History-02 email to you and something jogged my memory.
While the initial cost objectives could not fully be achieved the 1401 nevertheless found a huge market, admittedly not for the originally aimed smaller customers for which Jonny Dayger's 1403 printer was too powerful and expensive, but the only one available. This was solved with the follow-on systems IBM 1430, IBM /360 Model 20 and /360 Model 15 later.
After the first few systems were installed in Atlanta having a 1401 system turned into a "Status Symbol". Customers were building show place computer rooms with plate glass windows to view system from the hall. Some even had a button you could push and hear a recording describing the system. I remember two really nice ones at Avon (the Cosmetic company) and Retail Credit Corp. (later to become Equifax). This was before the race riots in Los Angeles(1) where a data center was burned down and the insurance companies demanded much more security. Before that you could just walk in off the street to any company and go where you wanted. There may be a receptionist in the lobby but no security guard and sign in desk.
These show case computer rooms reminded me of things that came later like, my lawn mower is bigger than yours, and still later my SUV can crush your Minivan. Everyone wanted one and some companies jumped the gap and were paying more for their 1401 than they did for their EAM installation. We ended up with 125 1401 systems in the Atlanta metro area.
Those thick cables? ;-))
Ed Thelen was asking/commenting/complaining about the thick cables and enormous plugs associated with the 1401 system. Maybe the cabling cost was not part of the cost target of a 1401 system. The signal cables were low frequency, but the tape system had amplified raw read head signals running back to the 1401 TAU in coax - ...
Robert Garner commented
Reading the WWAM docs, I believe the 1402 <-> 1401 connection came right from the WWAM design.
And the 729 <-> 1401 TAU connections are exact copies of the 7040 (or whichever it was) design.
The 1403 <-> 1401 connections I suspect came directly from the standalone print unit (forgot name at moment) and were morphed from tubes drivers to power transistors.
We can verify all this with the 1401 design team: Fran, John, Justin, and names to come!
Maybe thick cables is the trade-off for having the print and tape controllers in the 1401 cabinet, and the print hammer power supply in in the 1402.
Years later complex controllers are a small part of a hard drive, the cable receiving/driving electronics became "free", the cable speeds went through the roof, and serial became the name of the game. Think the various versions of USB (Universal Serial Bus). ;-))