In the same way that different readers can have utterly disparate evaluations of the same story – whether in terms of an author’s literary style, or, such fundamental elements as plot, theme, and setting – so and even more can different artists depict a story’s events and character by strikingly different visual styles. This is nicely epitomized in the illustrations created by Hubert Rogers and Edd Cartier to present the world imagined by A.E. van Vogt for his tale “Recruiting Station”. First published in the March, 1942 issue of Astounding Science Fiction…
… the story was reprinted by Fantasy Press as “Masters of Time” in their 1950 book by the same title, the publication also including van Vogt’s unrelated tale “The Changeling“, which originally appeared in Astounding in April of 1944.
Being far-too-far away in time from having read “Recruiting Station” (decades!) to remember the story’s precise details, suffice to say that though the tale doesn’t have the consistency of focus (emphatically not a hallmark of Van Vogt’s writing!) the author anomalously showed in his truly superb 1942 “Asylum”, it displayed the sense leaps of imagination coupled with creative-disconnectedness – of time, place, and sequence events – that made his story-telling fascinating, entrancing, perplexing (and yes, eye-rollingly maddening) at the same time, and, the presence of female protagonists central to the story, I think reflective of his early work as a writer of romances. MPorcius Fiction Log has a thorough evaluation of the story, aptly concluding with the following, “In my opinion, “Recruiting Station” is a good example of what van Vogt is all about. It is also interesting as a product of its time, as I have suggested, and feminist readers might find noteworthy its depiction of a college-educated professional woman who is given the responsibility of saving the universe but who at the same time has a man at the center of her psychological life, a man whose help she needs to succeed in her awful mission and to achieve personal happiness. Students of van Vogt’s long career may find his descriptions of the soldiers in the story as lusty, adventurous men unafraid of death, to be of a piece with his interest in “the violent male.” “Recruiting Station” gets a big thumbs up from this van Vogt aficionado.”
Fantasy Press’ 1950 publication has great cover and full page (just two in the whole book!) illustrations by Edd Cartier, while the chapters are headed by two alternating illustrations.
“Forty feet a day. In a blaze of wonder, Garson stood finally with his troop a hundred yards from that unnatural battle front. Like a robot he stood stiffly among those robot men, but his eyes and mind fed in undiminished fascination at the deadly mechanical routine that was the offense and defense.”
(page 69)
(Interesting contrast with Hubert Roger’s cover!)
“The Jeep caught him when he was still twenty feet from the fence. The cool-eyed women who operated it pointed the steadiest pistols Craig had ever faced. A few minutes later, at the house, Craig saw that the whole gang had been rounded up: Anrella, Nesbitt, Yerd, Shore, Cathcott, Gregory, all the servants; altogether forty people were lined up before a regular arsenal of machine guns manned by about a hundred women.”
(page 171)
(Though 1950 was well into the “jet age”, the aircraft above have very much of a WW II “vibe” to them. Otherwise, the lady is serious!)
Though the contrast is striking and the juxtaposition of black, violet, and shades of red vivid, Graves Gladney’s cover for the July, 1939 issue of Astounding Science Fiction, inspired by and presenting A.E. van Vogt’s story “Black Destroyer” doesn’t – in and of itself – attain anywhere near the same level of crisply conjectured dramatic realism as those of other Astounding cover artists of the late 30s and early 40s … such as Hubert Rogers and Charles Schneeman. Let alone, other covers by Gladney himself, like his fine work for Astounding’s March, 1939, issue, depicting “Cloark of Aesir”! Similarly, Gladney’s conception of the story’s non-human protagonist, the feline-like (but not really feline!) Coeurl is inconsistent with the the creature as actually described by van Vogt (where did Coeurl’s tentacles go?), while the diminutive winged-egg spaceship doesn’t quite convey the mystique of the mighty “Space Beagle”.
That being said, allowances can be made. In this case, vastly more important than the cover art itself is what this art signifies: This is the issue of Astounding that symbolizes the commencement of the “Golden Age of Science Fiction”, by virtue of “Black Destroyer” being A.E. van Vogt’s first published tale of science fiction, as well as the first story by Isaac Asimov (“Trends“) to be published in the magazine. Though not often mentioned in the context of this issue’s significance, the July ’39 issue also features C.L. Moore’s wonderful (and wonderfully told) story “Greater Than Gods“. Though this wasn’t her first effort at science fiction (or, science fiction / fantasy, if you prefer) that having been “Shambleau” in Weird Tales in 1933, “Greater Than Gods”, a story with overlapping themes of free will, destiny, parallel universes, and at a minor level romance, is solidly representative of her ability to create vivid worlds and settings, maintain a fast and gripping pace, portray the mental states of her characters, and particularly – in an almost Lovecraftian way – generate a mood that is almost physical in feeling. (Albeit “cosmicism” is certainly not at all central to her work.) I’d certainly include Moore, along with, for example, Cordwainer Smith, as having been among the “top ten”; “top five” science fiction writers of the Golden Age, and not just the Golden Age, though her oeuvre in the field ended by the late 1950s. As for Isaac Asimov? Well, despite the size and scope of his body of work, with the exception of “Pebble in the Sky” I have never been partial to his writing.
So. Having commenced this blog in 2016, I’ve at long last acquired a copy of the July ’39 issue of Astounding. (The impetus for this post!) Below, you’ll find all manner of links pertaining to the literary and cultural intersection between “Black Destroyer” and the Golden Age, specifically about “Black Destroyer”, and about A.E. van Vogt himself. It seems that he is still well remembered, and that is a very good thing.
For your enchantment, enlightenment, and entertainment…
Via “Battered, Tattered, Yellowed, & Creased ~ Adventures in Fiction”, here’s an image from Dungeons & Dragons of a “Displacer Beast”, obviously inspired by Coeurl. (Artist unknown.)
Gustavus Adolphus College A Guide to Speculative Fiction at Gustavus Library: 1926-1950: “The Pulp Era and the Golden Age” (Abe Nemon, August 24, 2023)
… The Guardian (Damien Walter: “Science-fiction’s Golden Age writers left a fantastic legacy”, September 13, 2013)
“Today’s SF writers stand on the shoulders of giants and their fans barely understand where the wealth of this genre has come from. A time when the exercise of imagination and wonder was considered perverse and childish, but which now fuels a multi-billion dollar a year industry that churns out so much crap that the rare treasures of the Masters are lost in a avalanche of dreck. Harlan performs a terrific service in this commentary and in his championing of work that challenges the imagination.” – @Zagadka42 – 10 years ago
A. E. van Vogt, 1912-2000 “SF Authors Remember A.E. van Vogt” Harlan Ellison Poul Anderson Ray Bradbury Sir Arthur C. Clarke, CBE Jack L. Chalker James E. Gunn David Langford Paul Levinson Richard Matheson Jerry Pournelle Mike Resnick Robert J. Sawyer Michael Swanwick Jack Williamson
A.E. van Vogt Interview…
“An A. E. van Vogt interview” (by Richard Wolinsky?), of February 23, 1980 on KPFA radio program “Probabilities”, at Charles Smyth’s YouTube channel.
Cover Artist Graves Gladney (James F.G. Gladney), at…
The impact of Ridley Scott’s 1979 film “Alien” in the worlds of horror and cinematography has surely been enormous, and, continues. Certainly the movie didn’t appear “out of nowhere”, and – consciously or otherwise, as in works of art of all genres – its creation is the result of numerous influences and cultural antecedents, both literary and cinematic. Among the influences that immediately came to my mind – at least, upon writing this post! – are the films “It! The Terror from Beyond Space” (1958), “Planet of the Vampires” (1965), and A.E. van Vogt’s 1939 Astounding Science Fiction short stories “Black Destroyer” and “Discord in Scarlet” both of which were incorporated into his 1950 fix-up novel The Voyage of the Space Beagle.
My supposition was confirmed through the (inevitably!) very lengthy entry for the film at Wikipedia, which discusses “Alien’s” origins in great detail. Specifically: “Alien‘s roots in earlier works of fiction have been analyzed and acknowledged extensively by critics. The film has been said to have much in common with B movies such as The Thing from Another World (1951). Creature from the Black Lagoon (1954), It! The Terror from Beyond Space (1958), Night of the Blood Beast (1958), and Queen of Blood (1966), as well as its fellow 1970s horror films Jaws (1975) and Halloween (1978). Literary connections have also been suggested: Philip French of the Guardian has perceived thematic parallels with Agatha Christie’s And Then There Were None (1939). Many critics have also suggested that the film derives in part from A. E. van Vogt‘s The Voyage of the Space Beagle (1950), particularly its stories “The Black Destroyer”, in which a cat-like alien infiltrates the ship and hunts the crew, and “Discord in Scarlet”, in which an alien implants parasitic eggs inside crew members which then hatch and eat their way out. O’Bannon denies that this was a source of his inspiration for Alien‘s story. Van Vogt in fact initiated a lawsuit against 20th Century Fox over the similarities, but Fox settled out of court.
Several critics have suggested that the film was inspired by Italian filmmaker Mario Bava‘s cultclassicPlanet of the Vampires (1965), in both narrative details and visual design. Rick Sanchez of IGN has noted the “striking resemblance” between the two movies, especially in a celebrated sequence in which the crew discovers a ruin containing the skeletal remains of long-dead giant beings, and in the design and shots of the ship itself. Cinefantastique also noted the remarkable similarities between these scenes and other minor parallels. Robert Monell, on the DVD Maniacs website, observed that much of the conceptual design and some specific imagery in Alien “undoubtedly owes a great debt” to Bava’s film. Despite these similarities, O’Bannon and Scott both claimed in a 1979 interview that they had not seen Planet of the Vampires; decades later, O’Bannon would admit: “I stole the giant skeleton from the Planet of the Vampires.”
But…! Another “key” to the origin of “Alien” can be found at CultureNC’s YouTube channel (“Culture NC est une chaîne qui regroupe des vidéos sur la culture calédonienne” ((“Culture NC is a channel that brings together videos on New Caledonian culture”)) in the video “Alien: Pulp Origins“, of September 5, 2022. Therein, along with mention of “It! The Terror from Beyond Space” and “Planet of the Vampires”, CultureNC touches upon Howard Hawks’ 1951 “The Thing From Another World”, the two aforementioned A.E. van Vogt stories, the anthology Strange Relations by Philip José Farmer, and, the 1953 short story “Junkyard” by Clifford D. Simak. Ultimately, however, CultureNC arrives at an even earlier short story as having either prefigured “Alien”: Clark Ashton Smith’s “The Vaults of Yoh-Vombis” from the May, 1932 issue of Weird Tales.
I find CultureNC’s discussion fascinating, While it’s unknown if Smith’s specific tale truly influenced the creators of “Alien” – that I doubt, given the tale’s time-frame and perhaps relative obscurity – what is remarkable (and correct) is that the story foreshadowed, if not anticipated, plot elements that emerged in the movie forty-seven years after its very Weird publication.
You can view Richard Corben’s adaptation of Smith’s story here. I’ve created PDF of the tale (by way of the Pulp Magazine Archive) which you can access (“yay! – free stuff!”), here.
For all its impact, and in spite of its obvious science-fiction tropes (space travel, cybernetics, suspended animation, and extraterrestrial life (of a gross and very deadly sort)), “Alien” unlike “Blade Runner” is emphatically not science fiction. It’s gothic horror; visual horror, which simply uses the idea (to be true, with marvelous effectiveness) – versus the reality – of “space” as a setting of emotional darkness, fear, and negative infinitude.
But yeah, it’s entertaining movie!
So, without further mouse clicking / scrolling delay, here’s Culture NC’s video:
There are two YouTube (audio) versions of “The Vaults of Yoh-Vombis”. Here they be:
“Things change, Kellon. They either change forward – or back -“
Hubert Rogers’ cover for the January, 1942 issue of Astounding Science Fiction is the perfect visual accompaniment to (or, symbolic reflection of!) Jack Williamson’s fine leading story, “Breakdown”. A fine example of Williamson’s literary skill, the story is a tale of the collapse of a solar-system spanning civilization due to a coalescence of social, economic, cultural, and technological factors, and the confrontation of the civilization’s de-facto leader – “Boss Kellon” (Harvey Kellon, to be specific) “Executive Secretary of the Union of Spacemen, Managers & Engineers” – with the realization that his preeminence and power have been rendered meaningless by changes that are beyond his control. Given that the plot and theme are universal in time and place, a story of this nature could in good literary hands be effectively told for any setting. But, Williamson’s adeptness in combining the theme of interplanetary travel with well-crafted characters, in fast-paced, engrossing prose, makes for an enjoyable read.
Though not nearly as powerful as the excellent “With Folded Hands” and “…And Searching Mind”, the story has thus far been anthologized nine times.
The “softness” of the objects in the illustration lend it a dreamlike quality. This extends from the grayish-blue torpedo-like spacecraft occupying pride-of-place in the center of the painting, to the two hazy, ill-defined moons (or, are they planets?!) floating in the background. The blues and grays contrast nicely with the red hills in the distance, and, the pale green building – the “Union Tower”? – to the right. I also like how Rogers created imaginary logos in red for both spacecraft and building, the latter as script.
If there is a single word for the cover, it would be pensive.
Every artist creates works that are memorable. Not necessarily in terms of technical expertise; not in terms of boldness of color; not in terms of clarity and detail. But instead, in terms of qualities that resonate with the human spirit: Ambiguity. Mystery. Symbolism. The unreal, in confrontation with the real.
One such artist was William Timmins, whose cover illustrations were featured on Astounding Science Fiction between December, 1942 and December, 1950. While some of his efforts were – “ho-hum” – adequate if unremarkable, others were striking in their power and boldness, embodying in a single painting a story’s animating concept and message. A particular example is his cover art for the December, 1942 issue of Astounding, for A.E. van Vogt’s “The Weapon Shop”, showing a five-hundred foot high five-tiered information center, from which leads an elevated ramp upon which stand pedestrians. Though a direct representation of a scene central to the story, the image has a dreamlike quality by virtue of the magnitude and distance of the building itself, which is viewed as if from a distance, from a vantage point below the ramp.
An even more memorable painting, for the cover of the January ’46 issue of Astounding, is a representation for Henry Kuttner’s story “The Fairy Chessmen”. (You can access parts one and two via the Pulp Magazine Archive.) In the image, as in the tale, Robert Cameron, civilian director of Psychometrics, is walking across a chessboard-as-hillside-landscape, dominated by human-size pawns, knight, queen, and king. With an autogiro in the distance (they were fashionable in the 40s!), he stands beneath a dark and gloomy sky, with a pattern of square, grayish yellow clouds above.
But, the chessboard is more than a mere chessboard, for the chess pieces are alive, watching, and waiting.
And so, I wondered: “How did Timmins’ painting actually appear, before it was a magazine cover?”
Assuming his painting was lost or destroyed decades ago – like the vast bulk of then unappreciated and now retrospectively invaluable pulp art – I thought I’d do an experiment: Using my scan of the original cover as a basis, I used Photoshop to repair defects, and, create elements of the cover art that were obscured or completely covered by title, logo, and other textual elements. This involved replicating the “fuzziness” and vagueness of some cover features, and at the same time trying to make these modifications consistent with the overall “feel” of Timmins’ painting. While I didn’t do so consciously, on completion, I realized that the square shape of the clouds – with gaps between them – mimicked that of the chessboard.
Perhaps Timmins’ original painting looked something like this.
Hope you like it. Other examples may follow.
The chessmen wonder.
“What is it about SF that draws us? What is sf anyhow? It grips fans; it grips editors; it grips writers. And none make any money. When I ponder this I see always in my mind Henry Kuttner’s FAIRY CHESSMEN with its opening paragraph, the doorknob that winks at the protagonist. When I ponder this I also see – outside my mind, right beside my desk – a complete file of UNKNOWN and UNKNOWN WORLDS, PLUS Astounding back to October 1933 … these being guarded by a nine-hundred-pound fireproof file cabinet, separated from the world, separated from life. Hence separated from decay and wear. Hence separated from time. I paid $390 for this fireproof file which protects these magazines. After my wife and daughter these mean more to me than anything else I own – or hope to own.”
“Notes Made Late at Night by a Weary SF Writer”, by Philip K. Dick written 1968 in Eternity Science Fiction, July, 1972
Here’s is Gnome Press’s 1951 edition of Tomorrow and Tomorrow / The Fairy Chessmen, featuring art by Harry Harrison. The dystopian theme of devastation by nuclear war is obviously implied by the presence of a mushroom cloud, a not uncommon visual trope in 50s science fiction art. I had absolutely no idea – until creating this post – that Harrison began his career as an illustrator, his first story appearing in 1951.
Another edition of The Fairy Chessmen, this time published in 1956 as Galaxy (Science Fiction) Novel # 26 under the title Chessboard Planet, with cover illustration by Edmund Emshwiller. The background has somewhat of a Richard Powers-ish feel to it…
(?) (!)
This Will Read You
Sutin, Lawrence, Divine Invasions – A Life of Philip K. Dick, Harmony Books, New York, N.Y., 1989 (page 35)
Known primarily for his collection of “Venus Equilateral” stories, Golden Age science-fiction writer George O. Smith’sbody of work comprised nine novels, many short stories, and, a number of reviews. I’ve only read a few examples of his work, these comprising a few Venus Equilateral tales. To be honest, I found these stories – which I think fall into the continuum of “hard science fiction” – to be straightforward, middling, and serviceable; neither bad nor exceptional. I’m glad I read them, but have no impetus to revisit them for another reading (or two, or three) as for example the stories of A.E. van Vogt (the early van Vogt!), Philip K. Dick, Cordwainer Smith, or Catherine Moore.
Among Smith’s novels was Nomad, which originally appeared as a three-part series in the December, 1944, and January and February 1945 issues of Astounding Science Fiction. For the December issue, William Timmins’ somewhat bland cover art is cast in muted tones of green, gray, and red.
____________________
Paul Orban’s interior illustrations do the story greater artistic justice. Here’s the opening illustration, on page 7. Note how the spacecraft has the general appearance of a submarine (a one-man submarine?!) – down to entry hatch, typical of many such illustrations from the period.
____________________
This illustration of a disintegrating spaceship appears on page 27. The nautical design theme is evident here, also.
____________________
The Nomad series was published in novel form by Prime Press in 1950, in a run of 2,500 copies. The cover illustration is by L. Robert Tschirsky, whose illustrations were featured on (and in) several works of science fiction, fantasy, and pseudoscience (Atlantis and all that) in the late 1940s.
My recent posts about the reality of space warfare – as imagined in Astounding Science Fiction in 1939 – present articles by Malcolm R. Jameson and Willy Ley, and, readers’ responses. That Willy Ley would figure so prominently in this topic is hardly surprising, for by profession he was a science writer with a lifelong focus in rocketry and space exploration, though his interests did extend further, encompassing the pseudoscience of – *ahem* – cryptozoology. The true scope of his enormous output can be fully appreciated by even the quickest glance at his biographical profile at the Internet Speculative Fiction Database. He body of work was quite multi-faceted, for it comprised a novel, letters, book reviews, interior art (primarily in 1948 issues of Astounding), twenty-two perhaps-more-better-known non-fiction books, as well as – well, primarily! – essays and articles for mid-twentieth-century science fiction pulps. An example of the latter is his oeuvre for Galaxy Science Fiction, whichbetween 1952 and 1969 published over 150 of his articles under the heading “For Your Information”.
His straightforward science journalism was accompanied by four (or five, depending on how you count?!) works of fiction. The “first” four are…
“At the Perihelion” (1937) “Orbit XXIII-H” (1938) “Fog” (1940) “The Invasion” (1940)
…the first three of these having been published in Astounding, and “The Invasion” in Super Science Stories.
Having read “Fog” (while preparing this post, and my posts about Space Warfare), I have to confess that I found it to be utterly underwhelming. Except for being placed in a metropolitan setting in post-1940s America, it’s much more a tale of totalitarian surveillance (hmmm…!) and political chaos (hmmm…?) in a dystopian future, I think inspired by Ley’s own experiences in Nazi Germany, from which he fled in early 1935. So, the simple title – it is apropos! – connotes the constant sense of uncertainty that pervades daily life in such a situation. (Once again, hmmm…!!) Otherwise, Charles Schneeman’s two illustrations for the story were better than the mere story itself!
Given Willy Ley’s huge body of work and influence in popularizing rocketry and space exploration, the abundance of information about him is entirely unsurprising. However, while delving into his biography amidst my posts on space warfare, I came across the following poignant news item by New York Times science writer Walter Sullivan: It’s Willy Ley’s obituary, published after his passing on June 24, 1969. While the obit doesn’t necessarily present information not already known and available elsewhere, it’s still of historical interest in terms of the details of Ley’s personal life, and, how a figure so significant in the worlds of science and journalism (like Walter Sullivan, himself!) was perceived in the popular press.
Here it is:
____________________
Willy Ley, Prolific Science Writer, Is Dead at 62
Prophesied Travel in Space in Book Issued in 1926 Fled Germany in ‘35 – Tested Rockets in Westchester
By WALTER SULLIVAN
The New York Times June 25, 1969
Willy Ley, who helped usher in the age of rocketry and then became perhaps its chief popularizer, died yesterday morning at his home In Jackson Heights, Queens. His age was 62.
Mr. Ley, the author of more than 30 books in English and German, was a frequent lecturer as well as teacher and industrial consultant.
His death, apparently from a heart attack, came suddenly. About a week ago a medical checkup had disclosed a circulatory disorder and he was taking digitalis.
Earlier in the day, in a telephone conversation with a book publisher, Mr. Ley spoke of the possibility that he might have to follow man’s first flight to the moon by television from his home, instead of from the Manned Spacecraft Center in Texas. It was a disappointing prospect, for Mr. Ley had been one of the earliest protagonists of such a flight.
He was born in Berlin in 1906 and his early studies, at the Universities of Berlin and Konigsberg, were in astronomy, physics, zoology and paleontology (the study of fossils). Some of his most successful books were on exotic beasts of fact and myth.
However, in 1927 he and his German colleagues were inspired by the writings of Hermann Oberth to found the Society for Space Travel. A punctilious registrar in Breslau at first refused to permit the group to incorporate under the title Verein fur Raumschiffahrt because, he said, the last word of the title (meaning “space travel”) did not exist in the German language.
Collaborated on Films
Mr. Ley’s first book on space travel appeared in 1926 and during that period he collaborated with Fritz Lang in several German science-fiction films, including one entitled “Frau im Mond” (“Woman in the Moon”).
(Here’s “Frau im Mond”, from Daily Motion.)
(And, a sort-of-counterpart to Lang’s film, from a decade later: Vasili Zhuravlov’s “Cosmic Voyage” (Космический Рейс – Kosmicheskiy reys) from 1936.
Among those whom he recruited into the Society for Space Travel was a young man named Werner Von Braun who ultimately became a leader in German military rocket development. After World War I, when Dr. Von Braun had begun working with the American rocket program, he and Mr. Ley collaborated on several books including “The Exploration of Mars.”
As the Nazis rose to power they were determined to take over rocket research from the society. The latter, through a series of flights with primitive liquid-fueled rockets from an abandoned ammunition dump on the outskirts of Berlin, had shown that rockets could be used to circumvent provisions in the Versailles Treaty forbidding German development of artillery.
In 1935, Mr. Ley got word to Dutch and British friends that he was in trouble with the Gestapo. He had been ordered to cease writing on rocketry for foreign publications and did so, but some of his earlier articles being held in reserve by British newspapers appeared after this edict.
Mr. Ley left for Britain and then was brought to the United States under the auspices of the American Interplanetary Society (which about this time changed its name to the American Rocket Society). Members of this group put up bond to permit his entry into the country.
Built Test Stand
Mr. Ley lived for half a year with G. Edward Pendray, head of the American Rocket Society, and the two men built a test stand for small rockets near Mr. Pendray’s home in Crestwood, N.Y. It was in a swamp between Scarsdale and Bronxville.
Mr. Pendray recalled yesterday the alarm of neighbors at the roaring of rockets on their test stand. However Mr. Ley’s activities as an experimenter gave way to concentration on writing.
He turned out a steady stream of books and articles. Interest in rocketry and space travel was low at the time and his titles ran to such subjects as “Salamanders and Other Wonders,” “Dragons in Amber” and “The Lungfish, The Dodo and the Unicorn.”
However when the rockets developed by his former colleagues in Germany began flying across the English Channel, there was a dramatic change. The demand for expert writing on rocketry became insatiable.
Meanwhile, Mr. Ley in 1940 joined the newspaper PM as science editor and soon met a Russian-born ballet dancer, Olga Feldman [Feldmann], who was writing a column on physical fitness for the newspaper. They were married in 1941.
Soon afterward, Mrs. Ley was doing research for her husband at a public library and read to him, over the phone, certain information on rockets that she had uncovered there. Someone in the next phone booth overheard transmission of this information in a Russian accent and reportedly notified the Federal Bureau of Investigation.
It took a certain amount of explaining to convince the Federal authorities that nothing untoward was going on.
In 1944 he became a United States citizen and left PM. He became further identified with space travel with such books as “Watchers of the Skies,” “Conquest of Space” and “Rockets, Missiles and Men in Space.” He also developed a powerful lecture style.
One close acquaintance noted yesterday that Mr. Ley’s big frame and German accent conspired to give him an impressively authoritative manner. Perhaps, he suggested, that was why Mr. Ley unconsciously retained the accent, even though he became fluent in his spoken and written English.
One of those who knew him well said he was a natural lecturer, “not only on the platform, but in private.”
“If you asked him a question you got a lecture,” he said, adding that Mr. Ley’s knowledge was “encyclopedic.”
Mr. Ley enjoyed good food, good drink and good conversation and belonged to a small convivial group of writers and scholars known as the “Trap Door Spiders,” who met once a month. The name, members say, is based on the practice of such spiders in closing a trap door to escape their mates.
He was a great admirer of Wagner operas and could accompany himself on the piano as he sang- Wagnerian arias.
Publishing associates said yesterday that Mr. Ley had at least six books under contract. He had told Scribners that next Monday he would deliver the final section of “Man and the Moon,” a major work, in preparation for five years. It deals with the role of the moon in music and literature.
Mr. Ley, one of his book editors said, was “like those 19th-century natural scientists who were up on every field of science.” He had been on the faculty of Fairleigh Dickinson University for many years.
While Mr. Ley was an ardent promoter of trips to Mars and other distant bodies, his earliest passion was for the moon.
“The moon is still silvery in the night sky,” he wrote in The New York Times last year, “but it is no longer unreachable.”
“In 1930 I introduced a number of aeronautical engineers in Berlin to the first liquid fuel rocket they had ever seen,” he said. “It stood about 5 feet tall and, even when fueled, was light enough to be lifted with one hand. It could climb about 1500 feet and was brought back by parachute.
“What, the engineers wanted to know, was the aim of all this? Eventually, I replied, rockets of this type will carry men to the moon.”
Mr. Ley lived to within one month of the scheduled fulfillment of his prophecy.
Besides his widow, he is survived by two daughters, Sandra Ley and Mrs. Xenia Parker of 252 East 61st Street. Since World War II Mr. Ley had lived at 37-26 77th Street in Jackson Heights
The funeral will take place ‘tomorrow at 1 P.M. at the Walter B. Cooke funeral home, 1504 Third Avenue.
____________________
Despite Willy Ley’s prominence in the history of science journalism, oddly, no information is available about his place of burial. However (!), if we’re talking biographical details, here’s the Declaration of Intention for American citizenship that he filed on June 22, 1937, five months after he reached Miami – from Havana – on February 2 of that year. Note that, appropriate to his current and future career, he listed his profession as “Scientific Research Writer”. (This document’s from Ancestry.com.)
A Reference or Two, or Three, and More, for Willy O.O. (Otto Oskar) Ley, at…
…Archive.org – Publications (262 scanned works – includes monographs, but primarily comprised of issues of science-fiction pulps featuring his articles.)
…Project Gutenberg (7 books. These appear to be juvenile or young adult fiction, all authored by Carey Rockwell, with Willy Ley as “Technical Advisor”.)
From Astounding Science Fiction of May, 1949, the article “Electrical Mathematicians,” by Lorne Maclaughlan, focuses on the the use of computers – specifically, electronic as opposed to mechanical computers – as devices to perform mathematical calculations. It’s one of the four non-fiction articles pertaining to cybernetics and computation published by the magazine that year, the other three having been:
“Modern Calculators” (digital and analog calculation), by E.L. Locke; pp. 87-106 – January
“The Little Blue Cells” (‘Selectron’ data storage tube), by J.J. Coupling; pp. 85-99 – February
“Cybernetics” (review of Norbert Wiener’s book by the same title), by E.L. Locke; pp. 78-87 – September
The identity and background of author Maclaughlan remain an enigma. (At least, in terms of “this” post!) The Internet Speculative Fiction Database lists only two other entries under his name, both in Astounding (“Noise from Outside” in 1947, and “Servomechanisms” in 1948, while web searches yield a parallel paucity of results. This absence biographical information, especially in light of the over seven decades that have transpired since 1949, coupled with the author’s distinctive writing style – combining clarity and economy of expression, and easy familiarity with the language of technology – leads me to wonder if that very name “Lorne Maclaughlin” (note the lack of a middle initial?) might actually have been a pen-name for an engineer or academic. Given the somewhat ambiguous reputation of science-fiction in professional and credentialed circles (albeit a reputation by the 1940s changing for the better) maybe “Maclaughlan” – assuming the name was a pseudonym – might have wanted to maintain a degree of anonymity.
Well, if so (maybe so?!) that anonymity has successfully persisted to this day!
Anyway, the cover art’s cool.
Depicting a scene from the opening of Hal Clement’s serialized novel Needle (the inspiration for the 1987 Kyle MacLachlan film The Hidden?), it’s one of the three (color, naturally) AstoundingScience Fiction cover illustrations by Paul Orban, an illustrator primarily known for his fabulously imaginative interior work, whose abundant output was only exceeded by his talent.
As for Maclaughlan’s article itself, it begins with a brief overview of the implications of the increasing centrality of calculating devices in contemporary (1949 contemporary, that is!) society, and the future.
This is followed by a discussion of the very nature of calculation, whether performed by mechanical or electronic devices, which then segues into a comparison of the similarities and differences between binary and decimal systems of counting and computation, and an explanation of the utility of the former in computing devices.
Next, a lengthy discussion of memory. (We’ve all heard of that…) note the statement, “Not only must we “teach” the machine the multiplication table – by the process of wiring in the right connections – but it may also be necessary to provide built-in tables of sine and cosine functions, as well as other commonly used functions. This is a permanent kind of memory – a fast temporary kind of memory is also needed to remember such things as the product referred to above until it is no longer needed. This memory has not been easy to provide in required amounts, but recently invented electronic devices seem to offer some hope that this difficulty can be overcome.” In this, author Maclaughlan is anticipating what we know today as ROM (read-only-memory) and RAM (random-access-memory), respectively.
From this, we come to computation in terms of the technology and operation of then-existing computers. This encompasses ENIAC (Electronic Numerical Integrator and Computer), EDVAC (Electronic Discrete Variable Automatic Computer), and MANIAC (Mathematical Analyzer Numerical Integrator and Automatic Computer Model I), and briefly touches upon the Selectron tube, the latter device the subject of J.J. Coupling’s article in the February 1949 issue of Astounding.
The final part of Mclaughlan’s article is a discussion of the nature, advantages, and use of “analyzers” – Differential Analyzers, and Transient Network Analyzers – in computation: Specifically, in the solution of differential equations pertinent to scientific research, such as, “…the flow of microwave energy in wave-guides, the flow of compressible fluids in pipes, and even the solution of Schrodinger’s Wave Equation,” or military applications, such as aiming anti-aircraft guns or determining the trajectory of nuclear weapons, noting, “These latter-day buzz-bombs will be sufficiently lethal to warrant their carrying along their own computers.”
Prescience, or, inevitability?
And finally, the article concludes with a photograph.
And, so…
ELECTRICAL MATHEMATICIANS
“The differential analyzer is more versatile than the network analyzer discussed above because it can integrate, differentiate – in effect – and multiply, and thus solve rather complicated differential equations. These functions are performed by mechanical or electro-mechanical devices in the differential analyzer. If these things could be accomplished by purely electrical means, we would expect a great increase in speed, and some decrease in size and weight.”
To an extent none of us today can realize, these rapidly growing electrical calculators will become more and more important factors in ordinary life. So far, they are handling only simple, straight-arithmetic problems. They are brains, but so far they think only on low levels. Give them time; they will be planners yet!
In this machine age no one is surprised at the announcement of some new or improved labor-saving device. The scientists and technologists who design our new electronic rattraps, microwave hot-dog dispensers and atomic power plants have succeeded so well that they have created a serious manpower shortage in their own professions. This shortage, which is chiefly in the field of analysis has recently forced them to put an unprecedented amount of effort into the design of machines to save themselves mental labor. The results of their efforts are an amazingly variegated collection of computing machines, or “artificial brains” as they are called in the popular press:
The development of such machines took a tremendous spurt during the war, and today we can scarcely find a laboratory or university in the land which is not devoting some part of its efforts to work of this kind. Progress is so rapid that the machines are obsolete before they are completed, and thus no two identical machines exist.
We cannot say that the computing machine is a new invention – the unknown Chinese originator of the abacus provided man with his first calculating machine in the sixth century B.C. This would seem to make the machine nearly as old as the art of calculating, but man is equipped with fingers and toes which have always provided a handy portable computing device. In fact, as we shall see, the simple fact that we have ten lingers has a definite bearing on the number of tubes and the kind of circuits required in electronic digital computers.
__________
Kelvin Wheel-and-Disk integrator. This device, which gives the integral of a radial distance with respect to an angle, is the most important unit in a differential analyzer of the electromechanical type.
__________
It should be pointed out that there are two distinct types of computing machines in common use today. One type deals with discrete whole numbers, counting them off with the aid of teeth on a wheel, or electrical pulses in vacuum tube circuits. These numbers represent quantities, and they are added and multiplied just as numbers are on paper, but at a much higher speed. These machines called digital computers, range from the simple cash-register adding machine to the complex all-electric ENIAC, with its eighteen thousand radio-type vacuum tubes.
The other type of machine is the analogue type of computer, in which the number to be dealt with is converted into some measurable quantity, such as length along a slide rule, or angle of rotation of a shaft. The operations are performed electrically or mechanically, and the answer appears as a length, an angle, a voltage or some other quantity which must be converted back to a number. The ideal machine of the analogue type will accept mathematical functions, empirical curves and directions for mixing and stirring, and turn out results in the form of curves automatically.
The digital computer is much more accurate than the analogue type for the simple reason that is easy to extend the number of significant digits in such machines to something like thirty or forty. It is impossible to measure a point on a curve to anything approaching one part in 1040. However, the analogue computers are in many ways faster and more versatile, because they can perform certain difficult mathematical operations directly, while digital machines require that these operations be reduced to addition and multiplication.
One of the first things we must do to understand modern digital computing machines is to disconnect our minds from the decimal number system, and get a more basic concept of number representation. The decimal system of numbers is a natural choice, based on the fact man has that ten fingers. We would perhaps be more fortunate had evolution given us twelve, for then our number system would be the more convenient duo-decimal system. Let us examine this system as a starting point, by studying the table of numbers below.
1
2
3
4
5
6
7
8
9
*
t
10
11
12
13
14
15
16
17
18
19
1*
t
20
21
22
23
24
25
26
27
28
29
2*
t
30
The six-fingered man would count to six on one hand, and then continue, seven, eight, nine, star, dagger, ten on the other. His ten would be our twelve, of course, but it would be a resting point for him while he got his shoes off to continue to his twenty – our twenty-four – on his twelve toes.
If we continue the table for twelve lines of twelve numbers each we will get to his one hundred, which corresponds to our one hundred forty-four. This number is his ten squared – our twelve squared – as it would be, and is preceded by his daggerty-dagger, ††. This duodecimal system has the advantage that ten can be divided by 2, 3, 4 and 6, giving in each case whole numbers – 10/4 = 3, 10/6 = 2, et cetera – while our ten is only divisible by 2 and 5. The ancient Babylonians were fond of this system, and also used sixty as a number base. These systems remain today as the bases of our measurement of time in seconds, minutes and hours.
Now let us examine the binary system, based on two. In this system all numbers are made up of combinations of just two digits, one and zero. The simplicity of this system makes it possible to use simple devices such as electromagnetic relays to represent numbers. The simple relay has two possible positions, open and closed, and we can represent zero by means of the open position, and one by the closed position, and then build up any number as shown in the table below.
Decimal System
Binary System
1
1
2
10
3
11
4
100
5
101
6
110
7
111
8
1000
9
1001
10
1010
Computation is easy with this system, once we get the hang of it. Thus our two cubed becomes, 1011 = 10 x 10 x 10 = 1000, and our two times three becomes 10 x 11 = 110, which is our six, as it should be.
With our minds cleared for action on any number base let us consider the capabilities which are necessary in a digital computer. Digital computation requires that all operations be reduced to those of addition, subtraction, multiplication and division whether a machine is used or not. These operations involve certain reflex actions, such as the response “six” when presented with the numbers “two” and “three” and the idea “multiply.” The trained human mind possesses such reflex actions, and the machine must also possess them, as a first requirement. Simple computing devices such as the commercial accounting machine possess a few reflexes. It is necessary to build many rapid reflexes into mathematical computing machines.
The next “mental” capability the machine must possess is that of memory. When we must multiply two numbers together before adding them to a third, memory is needed to preserve the product until the second operation can be performed. Commercial calculating machines have limited memory – after multiplication, for example, the number appears on the output wheels, and the third number can easily be added. The memory requirements in a good mathematical machine are much, much more stringent, and provide some of the toughest problems in design. Not only must we “teach” the machine the multiplication table – by the process of wiring in the right connections – but it may also be necessary to provide built-in tables of sine and cosine functions, as well as other commonly used functions. This is a permanent kind of memory – a fast temporary kind of memory is also needed to remember such things as the product referred to above until it is no longer needed. This memory has not been easy to provide in required amounts, but recently invented electronic devices seem to offer some hope that this difficulty can be overcome.
There are still two capabilities left. These are choice and sequence. The computing machine should be able to choose between two numbers, or two operations it can perform, in accordance with certain rules. Sequence involves, as the name implies, the proper choice of order of numbers or operations according to some rule which applies in the particular problem being solved.
These last two capabilities are not found to any great extent in any but the most modern mathematical computing machines. On the other hand there are a multitude of other mental capabilities found in humans which are undesirable in mathematical machines. Emotion, aesthetics, creative ability and so forth are not desirable, for these help to make humans unfit for much routine computing work. What we want is perfect slave, fast, untiring and industrious, who will never embarrass or disconcert us with unexpected response. (Of course the engineers in charge of some of the complicated modern mathematical machines are quick to accuse them of temper tantrums and other undesirable emotions.)
Perhaps the fanciest digital computing machine today is the IBM Automatic Sequence Controlled calculator at Harvard. The letters IBM International Business Machines Corporation, which has developed a series of machines intended for use in accounting work. These machines use a punched card – a device with quite a history, as histories go in the computing field. It would seem that weaving machines which could be used to more or less automatically weave patterned cloth excited the imagination of a good many inventors in in the early eighteenth century. In such weaving it was necessary to sequence automatically the “shredding,” or controlling of the warp threads so that weft threads could be passed through them to weave a pattern. Punched tape and punched cards had already been by 1727. The punched cards we use today get the name Jacquard cards from the name of the inventor of an improved weaving machine around the year 1800.
This basic idea was good enough to attract the attention of Charles Babbage, an English actuary, who is regarded as the lather of the modern computing machine. His “difference engine” was designed, in his words, “to perform the whole operation” – of the computing and printing of tables of functions – “with no mental attention when numbers have once been fed in the machine.” When this “engine” was nearly complete the government withdrew its support of the Project, and Babbage began the construction of an analytical machine on his own. This machine, a wholly mechanical device, was to use punched Jacquard cards for automatic sequencing. In 1906 his son successfully completed a machine with which he calculated pi to twenty-nine significant figures.
Hollerith, in this country, made a great advance in the use of punched cards when he invented a card sorter to aid in classifying the results of the 1880 census. Most people today are familiar with the kind of things that a sorter can do. Thus if we have a sorter and a stack of cards with personal and alphabetical information punched thereon we can request the machine to pick out all left-handed individuals with cross-eyes and Z for a second initial, and bzzzzt, bzzzzt, bzzzzt – there they are.
The IBM Company, by catering to the needs of organizations which handle – and have – a good deal of money, was able to put the manufacture of computing machines on a paying basis. It need not be pointed out that it is much more difficult to produce profitably machines which will only be used for such tasks as the calculation of pi to umpteen places. However the punched card machines built for accountants have found their way into scientific computing laboratories, and the IBM Company has a research laboratory which is actively developing new machines for scientific use as well as for accounting.
A punched card machine operating on the Hollerith principle interprets numerical and operational data according to the positions of holes punched on cards, and then perform various mathematical operations. The cards, which are familiar to most people – postal notes, government checks, et cetera – have twelve vertical positions in each of eighty columns. The vertical positions are labeled y, x, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9. Thus an 80 digit or two 40 digit numbers can be set up on one card, and the y space, for example, may be used to indicate sign.
The cards are read for purposes of sorting et cetera by a simple mechanism involving a metal cylinder and sets of electrically conducting brushes. As the card moves between the rotating cylinder and the eighty brushes, one for each column, an electrical contact is made whenever a punched hole passes under a brush. The position of the cylinder at the time that the brush makes contact indicates the number, or letter, represented. Any number system could be used, but the decimal system is selected because of its familiarity. The various IBM machines now on the market include Card Punchers, Card Interpretaters [sic], Card Sorters, Collators and others, all operating on the same basic principles. The most useful machine to scientific workers is the Automatic Multiplying Punch. This machine will multiply factors punched in cards, and will automatically punch the product in a card, or even add and punch out products.
The computer lab at Harvard, mentioned above, uses a combination of these machines and a device for sequencing their operations – whence the name IBM Automatic Sequence Controlled. This calculator is one of the half-dozen large machines in this country which can be used to tear into a tough problem and quickly reduce it to a neat column of figures – or a stack of cards, in this case. Since it is a digital type of computer capable of great accuracy, but because it is partly mechanical in operation it is slow compared to the newer all electronic machines. The automatic sequencing apparatus is not easy to set up, and thus type of machine is best suited to the solution of repetitive types of problems, such as the calculation of tables. The punched card is a convenient form in which to store tables of simple functions, e.g. Sin x, Log x, which are often needed in computation of tables of more complicated functions.
Of course, if you want to prepare a table umpteen places Bessell Functions, or evaluate some determinants, or make some matrix algebra manipulations you will have to wait s time for your turn on this or any similar machine. You will have to have a pretty good story too, for these machines are at work today, and sometimes night as well with important problems. It must be realized too, that a problem be rather important and complex before it is even worthwhile to the labor of setting it up for solution in such a complicated machine.
Punched cards are often used to store scientific data other than tables with the advantages of machine sorting et cetera possible with IBM machines. Thus at the Caltech wind tunnel data from instruments is punched directly on cards. Astronomers locate star images by pre-computed co-ordinates on punched cards, and then measure the star positions accurately and record the new information on new cards. The Census Bureau makes a great deal of use of punched cards at present, but plans are being made to go over to the faster electronic computers for this work.
__________
Basic flip-flop vacuum-tube circuit used in the ENIAC and in other digital computers. Tube number 2 – shaded – is conducting, and tube number 1 is “cut-off”, in the diagram above. A positive pulse on tube 1 will cause it to conduct and the resultant drop in its plate voltage will cause tube 2 to cease conducting. This condition is stable until another pulse arrives, on the grid of tube 2.
__________
Shortly before the war, G.R. Stibitz and others at the Bell Telephone Laboratories developed a relay type of computer which could handle not only real numbers but complex numbers as well. The binary number system is convenient in a relay computer as we have pointed out. There is some difficulty entailed in the process of getting from a number expressed in the ordinary decimal system to the binary system and back again. For this reason Stibitz likes what he calls a bi-quinary system, which uses base 2 to tell if a number is between 0 and 4, or 5 and 9, and base 5 to tell which digit it is of the five. Early in the war the Army and Navy each ordered one of these relay computers, and machine computation was off to a flying start.
Dr. H.H. Aiken, who had built the IBM computer at Harvard has recently gone over to the relay type of computer, and his “Mark II” will soon be in operation on the complicated guided missile ballistics problems being studied at the Dahlgren Proving Ground. IBM has also been playing around with relay computers, and has delivered two sequence controlled machines of this type for ballistic research workers. Aiken does his sequencing with standard teletype tape, while some of the IBM jobs use plugboards.
An interesting example of a similar parallel development is the Zuse computer, named after its designer Conrad Zuse, who developed his machine in Germany during and since the war. Like the Bell Laboratories machine it uses a keyboard to feed numbers into its relays. The sequence is prepared in advance by an operator who punches instructions into a strip of film.
The art of machine computation took a tremendous jump ahead when in the fall of 1946 the ENIAC, the first electronic digital machine, was placed in operation. This machine was built for Army Ordnance at the Moore School of Engineering by J.W. Mauchly, J.P. Eckert and others. The ENIAC – Electro Numerical Integrator and Calculator – with its eighteen thousand tubes is over a thousand times faster than the relay machines, which in turn were twelve times faster than the original punched card machine at Harvard. This tremendous increase in speed is the result of shifting over from the use of one gram relay armatures to the use of 10-31 gram electrons as moving parts. Of course a number of new problems appeared when this one limitation was removed. They are being cleared up one by one, chiefly by electronic means.
The ENIAC, despite the light weight of its moving parts, is no vest-pocket machine, as the number of vacuum tubes would indicate. The filaments of these tubes alone require eighty kilowatts of power, and a special blower system is needed to take away the heat. The whole machine occupies a space about 100 feet by 10 feet by 3 feet. Tube failures were a source of a good deal of trouble, because for while at least one of the eighteen thousand tubes burned out each time the power was turned on. This trouble was reduced by leaving the filaments of the tubes on, night and day, to eliminate the shocks involved in heating and cooling, so that now the ENIAC burn-outs at only about one per day, which take on the average of only fifteen minutes to repair. Experience with this machine has aided the design of a series of successors, such as the EDVAC, the UNIVAC, and the MANIAC – inevitable name.
The most important type of unit in the ENIAC is a device which uses two triode tubes, called a flip-flop circuit. These tubes will do electrically what the relay does mechanically. Normally one of the two tubes is conducting current, and the other is “cut off.” A very short – 0.000001 seconds long – pulse of voltage can cause this tube to cut off or cease to conduct, and the other to begin to conduct. Since only these two stable states are possible, we have the beginning of a binary computer. We must add a small neon bulb to indicate when the second tube is conducting, and then add as many such units in series as there are binary digits in the number we wish to handle. These circuits are used as a fast memory device. The ENIAC has a fast memory of only twenty ten-digit numbers, a serious limitation which can only be overcome by adding to the already large lumber of tubes, or by going to other types of fast memory.
Adding is accomplished by connecting flip-flop circuits in tandem so that they can count series of electrical pulses. This counting works in the same way that the mileage indicator works in a car, except that the scale of two is used. Thus, suppose that initially all our flip-flop circuits are in one condition – call it flip. The first pulse causes the first circuit to go from flip to flop. The next one will return it to flip, and this causes the first circuit to emit a pulse which sends the second circuit to flop. This continues on throughout the chain of circuits, all connected in tandem, as long as pulses are fed into the first circuit. When two series of pulses have been fed in we can get our number by noting which circuits are on flip – binary zero – and which on flop – binary one. The result may be converted back to pulses for use elsewhere. The speed per digit in the adding operation is a comfortably short ten microseconds.
The description of the adding scheme above has omitted one added complication in circuit design which gives a considerable simplification in reading of numbers. The binary system is used to count only to ten in the ENIAC and the number is then converted to a decimal number. This is a bit of a nuisance, circuit-wise, but handy – the decimal system is familiar.
The ENIAC also has electronic circuits for multiplying, dividing, square-rooting and so forth. The multiplier uses a built-in electrical multiplication table to aid it in its high-speed, ten digit operation. One very important unit in the ENIAC is the master programmer, which changes the machine from one computing sequence to another, as a complex computation progresses, in accordance with a pre-set plan. The master programmer even makes possible connections which enable the machine to choose the proper computing sequence when faced with the necessity for a choice. Thus it would almost seem that the machine does possess a kind of built-in judgment, and that there is some reason for the term “electrical brain.”
It was mentioned that the fast memory of the ENIAC was limited. The slow memory, using punch cards, and IBM machines causes a great reduction in speed when it must be used. Also, although computation is all-electronic, data is fed in and results are taken out by electromechanical means – punch cards again. The limitations incurred may best be realized if we compare the time for a punch, about half a second, with the unit time of a flip-flop circuit, ten microseconds. The ratio is fifty thousand times.
Even more serious is the problem common to all digital machines, namely the difficulty of setting up a problem. These machines are not easy to use, and the sequence of operations for an easy problem may be very involved. If the problem is difficult, then, of course, the sequence gets more difficult, but the use of machine methods is mandatory. So, when faced with a real stinger of a problem, the scientist gets down to work, perhaps for months, just to figure out how to set up the machine. Considerable time is needed for the physical setting up of sequence connections too, but after that – brrrrrrrrrrrrrp, and a solution which would take years by former methods begins to roll out in a matter of minutes.
Professor D.R. Hartree of England, who recently worked with the ENIAC, describes the solution of problem in which this machine had to handle two hundred thousand digits. Now try writing digits as fast as possible. At a rate which will lead to errors and writer’s cramp you may put down ten thousand digits in an hour. Even at this speed it will take twenty hours just to write down two hundred thousand digits – and no computation has been performed. The machine handled the numbers and performed the computation in this example in four minutes flat. It is not surprising that Professor Hartree is impressed by such speeds – he once spent fifteen years on the computation of the electron orbits of atoms. This is the kind of job that a machine calculator can be coerced into doing in a few hours, or days at most.
Their utility to science is obvious!
The ENIAC is only the first of its kind. The EDVAC – Electronic Discrete Variable Computer – is an improved machine, also built Army Ordnance at the r of Pennsylvania. One of the chief improvements is a larger capacity memory device, made possible use of acoustical delay lines for storage of numbers. Numbers get stored as trains of compression pulses is bouncing back and forth in a two-inch column of mercury. Each delay line of this type does the work of five hundred fifty electronic tubes in the ENIAC, so that a substantial saving results.
The MANIAC – Mechanical and Numerical Integrator and Computer – is another Army Ordnance computer. It is being built at the Institute of Advanced Study at Princeton under the direction of Dr. J. von Neumann and Dr. H.H. Goldstine. This machine is to use a new type of fast memory tube which is being perfected by Dr. Jan Rajchman of RCA. This tube, called the Selectron, is a kind of cathode ray tube which is designed to store four thousand ninety-six off-on or binary signals – equivalent to about twelve hundred decimal digits. The binary digits are to be stored as charge on points on a cathode screen which are behind the interstices of two orthogonal sets of sixty-four wires each. An ingenious method of connecting certain of these wires together will enable electric signals to be fed in to pull the electron beam to any position for purposes of reading” or “writing” with just thirty-two leads brought out. Even so a pre-production model of the tube looks a bit formidable, but it is phenomenally small for the memory it possesses.
Among some of the other schemes for digital memory being worked on are delay networks using loops of wire in wire recorders. This scheme may not be as fast as the acoustical delay line used in the EDVAC, but it has the advantage that the pulses do not have to be periodically removed for reshaping. One practical difficulty here is the necessity of waiting for the right point on the wire to come around before reading begins. Of course all memory of a number can easily be erased when need for it is finished, and the wire loop is ready to be re-used.
It seems that the Selectron is one of the best bets to speed up the operation of all-electronic computers. With its aid it should be possible to multiply two twelve-digit numbers in one hundred millionths of a second.
Such speeds may seem fantastic, but problems have been formulated and shelved because even the fastest present-day computing machines could not complete the solution in thousands of years.
The Bureau of Standards, aided by Mauchly and Eckert of ENIAC fame and others, is now constructing some new machines of a general purpose type. This new digital computer is called the UNIVAC – Universal Automatic Computer – and is to be of a general purpose type suited for Bureau of Census work as well as, Army and Navy ballistics and fire control research. The UNIVAC is to be very compact, using only about eight hundred tubes, and occupying only about as much space as five file cabinets.
It is rather interesting that one of the limitations of this and other digital machines is the slow rate at which numbers are printed at the output. This limitation may be overcome in future machines by the use of a device called the “Numero-scope,” recently announced by the Harvard Computation Lab. This device is nothing but a cathode-ray oscilloscope, which can trace the outline of any number, if the right signal is fed into its deflecting plates. This is no mean trick – it takes six vacuum tubes to make the numeral 2, for example, but it has been done, and numbers may now be flashed on the screen of a cathode-ray tube and photographed with exposures as short as one five-hundredth of a second.
The analogue computer, as we have stated works with analogous quantities rather than with whole numbers. Thus we may represent quantities by lengths, angles, voltages, velocities, forces and so on. Thus an electrical or an hydraulic circuit problem may be solved on a mechanical device, while an electrical problem may be solved on a mechanical device. One simple example of an analogue computer is the slide rule. Here quantities of any sort are converted into lengths and since a logarithmic scale is used it is possible to multiply by adding lengths. If a linear scale is used we can add by adding lengths. Division and subtraction are possible by simply subtracting lengths in each case.
__________
The basic mechanism in the punched-card machine is the brush and roller combination shown. As the card passes over a steel roller, metallic brushes make an electrical connection – between A and B in the diagram – and a signal can be produced to reject the card, or set a counter wheel, et cetera.
__________
If we use angles, or angular displacements, to represent quantities successive displacements readily add to give a total. We can also use a differential like the one in the rear end of a car to add the angular displacements in two different shafts. The answer in this case, or a constant factor – gear ratio – times the answer appears on a third shaft. Direct voltages add conveniently, and alternating voltages add like vector or directed quantities, and so are convenient in the solution of problems involving directed lengths or forces.
Before going any further into discussion of the specific details or these devices it might be well to examine the relative advantages and disadvantages of the analogue type of computer. In the digital computer the accuracy can usually be increased at the expense of speed, so that if we want to go from 10 digit to 20 digit accuracy we must suffer a decrease to half the original speed.
With the analogue type of computer it is only possible to increase accuracy if the lengths – or angles, or voltages, or whatnot – are measured with greater percentage accuracy. This may call for watchmaker techniques unless we can afford lengths or other analogous quantities. The difficulties encountered in any case are such that the accuracy is always much less than in any digital machine.
There are several advantages possessed by the analogue computer which tend to offset the decreased accuracy. One of these is its greater speed, which results partly from the fact that most problems are more easily set up for solution by analogue methods. Sometimes the analogue computer is used for a quick look at a problem, to narrow down the field which must be investigated with greater accuracy by the more involved digital computer. Another advantage possessed by the analogue computer is its ability – if the ability is built in – to perform certain mathematical operations in direct fashion. Thus, for example, a pivoted rod can be used to give the sine of an angle. This ability also accounts in part for the greater speed by the analogue method. Still another advantage is ease with which empirical data in the form of curves may be fed into an analogue machine.
The first successful large-scale analogue computer was the Differential Analyzer designed by Dr. Vannevar Bush and others at M.I.T. The same type of machine has also been built by General Electric for its own use and for use in various Universities. The latest and most highly improved of these machines was recently installed at the new engineering school at U.C.L.A.
Note that this YouTube clip shows the incorporation of the differential analyzer in the movies When Worlds Collide, from 2:00 to 4:13 (full length version here), and, Earth Versus the Flying Saucers, from 4:36 to end (in full-length version at Archive.org, from 59:28 to 107).
____________________
The differential analyzer is used chiefly for the solution of differential equations. In view of this fact it is rather strange that the machine cannot differentiate. However it can integrate, and since this is the inverse of differentiation its mastery over the calculus is quite complete. (The inverse of an arithmetical process is commonly used by clerks in stores who count back our change, and thus use addition in place of subtraction). The integrators in a differential analyzer are of the Kelvin wheel-and-disk type in which an integrator wheel rides on a rotating disk, and is turned when the disk turns. The amount of angular rotation of the integrator wheel depends on its distance, R, from the center of the disk, and the angle the disk turns through, θ. This, by definition, is the integral of R with respect to θ.
The integrator is the most important device in the differential analyzer, and as such has received a great deal of attention. In 1944 G.E. engineers came up with a device in which troubles caused by slipping of the integrator wheel on the disk were virtually eliminated. This device was essentially a servo follow-up system in which light beams were passed through a polaroid disk attached to a very light integrator wheel. These light beams then went through other polaroid disks, then to phototubes, to an amplifier and a motor. The motor then caused the second and third polaroid disks to follow the disk on the integrator wheel with the customary boost in power level, or torque level.
Among other important components in the differential analyzer are the input tables. At these tables, in the older machines, operators followed plotted curves of functions which were to be fed into the machine with pointers, and thus converted distances on the curve sheets to angular rotations. In the newer machines light beam photocell servo-mechanisms accomplish the same thing without the aid of skilled operators. Known functions, of course, are generated by other and simpler means.
Because the differential analyzer handles quantities in the form of angular displacements the process of adding is accomplished by the use of differential gearing. To solve a differential equation the machine must first be set up so that the right shafts are connected together by the right gear ratios. When all is ready the data in the form of curves is fed into the machine at the input tables, the known functions are fed in from function generators, and the output pens are moved from left to right, all in synchronism. Adding wheels, integrators, input table lead-screws and so forth all begin to move and perform the operations required by the equation being solved. The totals of the quantities on each side of the equation are held equal by a servo-mechanism and the shaft which will give the function which is the desired answer moves the output pen up and down as it is pulled across a sheet of graph paper. Thus the answer appears as a curve, or a set of curves.
The accuracy of these results depends not only upon the accuracy with which these final curves can be read, but also upon the accuracy of the original data, and the accuracy of the various servos involved in the solution. Typically, about one-tenth of one percent, or three digit accuracy can be expected. If some of the components have been forced to accelerate too rapidly because of a poor choice of gear ratio, or if a lead screw has been forced to the end of its travel, the solution may be completely wrong – the analyst still has his headaches. These troubles are ordinarily avoided by making preliminary runs to determine the proper ranges of operation of all components.
Among the other types of analogue computers commonly used engineering work are the various kinds of network analyzers. A large electrical power network may be exceedingly complex, due to the more or less random geographical distribution of loads and generating plants. The effect of short circuits, arc-overs due to lightning, and load distribution must be studied with the aid of models, so that the design of circuit breakers, lightning arresters and so forth can proceed intelligently. Tests cannot be made on the actual power network, as they can on communication networks, because of the possibility that damage to large and expensive equipment might result.
The earliest type of power network model was the D-C Network Analyzer. The representation of three-phase alternating current systems by direct-current models of this kind has definite limitations, and the next step was the development of A-C Network Analyzers. These models, although they represent a three-phase system by a single system are much more versatile than the D-C Analyzers.
We may ask if such models should really be classed as computers. Fundamentally, these analyzers are merely models of systems which are too complicated for direct analysis, and too large for direct measurement of variables under all possible conditions. Much the same kind of model-making is carried on in the study of aircraft antennas using model planes and microwaves in place of short waves. However, if we examine some of the uses to which Network Analyzers have been put, it seems safe to class them as computers. Because of the use of electrical quantities in these devices and because of the flexibility of interconnections possible, they have been used for the solution of such problems as the flow of microwave energy in wave-guides, the flow of compressible fluids in pipes, and even the solution of Schrodinger’s Wave Equation.
Another type of network analyzer is the Transient Network Analyzer, which can show more clearly what happens in a power network when short circuits and overloads occur. This device may also be used to study analogous problems such as the amplitude of transient vibrations in mechanical systems when sudden shocks or overloads occur. The inverse of this kind of thing is the mechanical model used to study what goes on in a vacuum tube. In these models stretched sheets of dental rubber are used to represent electrostatic fields, and ball bearings serve as electrons.
The differential analyzer is more versatile than the network analyzer discussed above because it can integrate, differentiate – in effect – and multiply, and thus solve rather complicated differential equations. These functions are performed by mechanical or electro-mechanical devices in the differential analyzer. If these things could be accomplished by purely electrical means, we would expect a great increase in speed, and some decrease in size and weight. Such machines have been built by Westinghouse and Caltech, and seem to promise a fair increase in speed over the old differential analyzer. It seems inevitable that the use of many vacuum tubes will lead to somewhat lower accuracy and less dependability. Another difficulty with present types of electronic differential analyzers is that integration can only be performed with respect to time as the independent variable, so that the solution of certain problems is not easily possible.
Many other kinds of analogue computers have been perfected in the last few years – the field is definitely “hot.” Completed designs include such gadgets as the Bell Telephone M-4 Director, which used radar signals to figure out in a twinkling where an antiaircraft gun should be aimed so that the shell and a plane might meet. Undoubtedly work is in progress on computers which will make possible solution the “problem of delivery” of the modern atomic warhead. These latter-day buzz-bombs will be sufficiently lethal to warrant their carrying along their own computers.
Many scientists are disconcerted by the fact that by far the greater part of the computer research being carried on today is under the auspices of the Armed Forces. To be sure, we in the United States seem to be far ahead of anyone else in the world in computers. This may augur well for National Security if some desperate bludgeoning struggle is soon to occur. From the longer range point of view it seems that it is particularly desirable that the scientist whose pure research may lead him to yet undiscovered fundamental truths be also equipped with this new and powerful tool.
__________
Three types of computers. Top: General Electric’s A.C. Network analyzer. Middle: The differential analyzer – of the analogue computer group – at General Electric. Bottom: The Bell Laboratories relay-operated digital computer.
The preeminent science-fiction magazine of the mid-twentieth century was Astounding Science Fiction, which rose to prominence under the editorial reign of John W. Campbell, Jr. First published in January 1930 as Astounding Stories of Super Science, the magazine has continued publication under the leadership of several editors and through various title changes, now being known as Analog Science Fiction and Fact.
Though by definition and nature a science fiction publication, Astounding (akin to its post-WW II counterparts and rivals Galaxy Science Fiction, and, The Magazine of Fantasy and Science Fiction (“F&SF”), let alone a host of otherpulps which had a lesser degree of literary and cultural (as opposed to artistic!) impact) also published non-fiction material. This comprised leading editorials, book reviews, and letters, as well as articles – typically, one per issue – about some aspect of the sciences. As for any serial publication, the nature of this content reflected the opinions and interests of the magazine’s readers, and, the intellectual and cultural tenor of the times.
A perusal of science articles in Astounding from the late 1940s reveals a focus on aerodynamics, astronomy, atomic energy, chemistry (organic and inorganic), computation, cybernetics, data storage, electronics, meteorology, physics, and rocketry. (Biology it seems, not so much!) Viewed as a whole, these subject areas – in the realm of the “hard sciences” – reflect interests in space travel (but of course!), the frontiers of physics, information technology, and the creation and use of new energy sources.
Let’s take a closer look.
Here are the (non-fiction) science articles that were published in Astounding Science Fiction in 1949:
January: “Modern Calculators” (Digital and analog calculation), by E.L. Locke; pp. 87-106
February: “The Little Blue Cells” (The “Selectron” data storage tube), by J.J. Coupling; pp. 85-99
March: “The Case of the Missing Octane” (Chemistry of petroleum and gasoline), by Arthur Dugan; pp. 102-113 (Great caricatures by Edd Cartier!)
April: “9 F 19” (Hydrocarbons), by Arthur C. Parlett; pp. 46-162
May: “Electrical Mathematicians” (Machine (electronic) calculation), by Lorne MacLaughlan; pp. 93-108
______________________________
I like this cover: June: “The Aphrodite Project” (Determining the mass of the planet Venus), by Philip Latham; pp. 73-84. (Art by Chesley Bonestell)
______________________________
July: “Talking on Pulses” (Electronic transmission of human speech and other forms of communication), by C. Rudmore; pp. 105-116.
August: “Coded Speech” (Electronic speech; noise reduction), by C. Rudmore; pp. 134-145
September: “Cybernetics” (Review of Norbert Wiener’s book by the same title), by E.L. Locke; pp. 78-87
October – First article: “Chance Remarks” (Communication research), by J.J. Coupling; pp. 104-111
October – Second article: “The Great Floods” (Review of great floods in human history), by L. Sprague de Camp; pp. 112-120
November: “The Time of Your Life” (Time; Determining the length of the earth’s day), by R.S. Richardson; pp. 110-121
December – First article: “Bacterial Time Bomb“, by Arthur Dugan; pp. 93-95
December – Second article: “Science and Pravda“, by Willy Ley; pp. 96-111
Regardless of the topic, a notable aspect of the non-fiction science content of Astounding (likewise for Galaxy and F&SF) is that mathematics – in terms of equations and formulae, let alone Cartesian graphs – was kept to a minimum, if not eschewed altogether. Science articles largely relied upon text to communicate subject material, and often included photographs (especially for issues published during the latter part of the Second World War) and diagrams as supplementary illustrations.
One such example – from February of 1949 – is presented below, in the form of J.J. Coupling’s article “The Little Blue Cells”.
______________________________
This issue features great cover art by Hubert Rogers for Jack Williamson’s (writing under the pen-name “Will Stewart”) serial “Seetee Shock“. The cover symbolizes adventure and defiance in the face of danger, by incorporating a backdrop of warning and admonition (“YOU WERE NOT EVOLVED FOR SPACE”; “BACK ADVENTURER”, and more) around the figure of a space-suited explorer, while cleverly using extremes of light and dark and a sprinkling of stars to connote “outer space”. Like much of Rogers’ best work, symbolism is as important as representation.
______________________________
Coupling’s article is notable because it pertains to a subject frequently addressed by Astounding, with continuing and likely indefinite relevance: recording, storing, preserving, and accessing information: computer memory.
The article focuses on Dr. Jan A. Rajchman’s – then – newly developed “Selectron Tube”, which was developed in the late 1940s at RCA (Radio Corporation of America) and about which extensive and rich literature is readily available, particularly at Charles S. Osborne’s wesbite. As implied and admitted by Coupling’s article, even at the time of the device’s invention there was uncertainty about its long-term economic and technical viability, despite its functionality and innovative design.
An image of a Selectron Tube, from Giorgio Basile’s Lamps & Tubes, is shown below. (Scroll down to end of post for a photograph showing a Selectron Tube in the hands of its inventor, illustrating its size. The thing’s big!)
Eventually, the initial, 4,096-bit storage capacity Selectron Tube proved to be more difficult to manufacture than anticipated, and the concept was re-designed for a 256-bit storage capacity Tube. To no avail. Both tube designs were superseded by magnetic core memory in the early 1950s.
As for J.J. Coupling? Well…(!)…this was actually the nom de plume of Dr. John R. Pierce, a CalTech educated engineer, who had a long and rich literary career, writing for Astounding, Analog, and other publications. His lengthy oeuvre is listed at The Internet Speculative Fiction Database.
Today, Dr. Pierce’s “The Little Blue Cells” opens a window onto the world of information technology and scientific literature – for the general public – from over six decades gone by. His article, with accompanying illustrations, is presented below.
______________________________
______________________________
THE LITTLE BLUE CELLS By J.J. COUPLING
The most acute problem in the design of a robot, a thinking machine, or any of the self-serving devices of science-fiction is memory. We can make the robot’s body, its sensory equipment, its muscles and limbs. But thinking requires association of remembered data; memory is the essential key. So we present the Little Blue Cells!
Most of the robots I have met have been either man-sized androids with positronic brains to match, or huge block-square piles of assorted electrical junk. The small, self-portable models I admire from a distance, but I feel no temptation to speculate about their inner secrets. The workings of the big thinking machines have intrigued me, however. It used to be that I didn’t know whether to believe in them or not. Now, the Bell Laboratories relay computer, the various IBM machines and the Eniac are actually grinding through computations in a manner at once superhuman and subhuman. With the other readers of Astounding I’ve had a sort of inducted tour through the brain cases of these monsters in “Modern Computing Devices” by E.L. Locke. I’m pretty much convinced. It’s beginning to look as if we’ll know the first robot well long before he’s born.
Perhaps some readers of science fiction can look back to the old, unenlightened days and remember a prophetic story called, I believe, “The Thinking Machine.” The inventor of that epoch had first to devise an “electronic language” before he could build his electrical cogitator. The modern thinking machine of the digital computer type comes equipped with a special electronic alphabet and vocabulary if not with a complete language. The alphabet has the characters off and on, or 0 and 1, the digits of the binary system of enumeration, and words must certainly be of the form 1001-110—and so on. We may take it from Mr. Locke that somewhere in the works of our thinking machine information will be transformed into such a series of binary digits, whether it be fed in on paper tape or picked up by an electronic eye or ear. The machine’s most abstruse thought, or its fondest recollection – if such machines eventually come to have emotions – will be stored away as off’s and on’s in the multitudinous blue cells of the device’s memory.
I’m sure that I’m right in describing the memory cells of the machine as multitudinous and little – that is, if it’s a machine of any capabilities at all. To describe them as blue is perhaps guessing against considerable odds, but there are reasons even for this seemingly unlikely prognostication.
The multitudinous part is, I think, obvious. The more memory cells the machine has, the more the machine can store away – learn – the more tables and material it can have on hand, and the more complicated routines it can remember and follow. The human brain, for instance, has around ten billion nerve cells. It may be that each of these can do more than store a single binary digit – a single off or on, or 0 or 1. Even if each nerve cell stored only one digit, that would still make the brain a lot bigger than any computing machine contemplated at present. Present plans for machines actually to be built call for one hundred thousand or so binary digits, or, for only a hundred-thousandth as many storage cells as the brain has nerve cells. Mathematicians like to talk about machines to store one to ten million binary digits, which would still fall short of the least estimated size of the brain by a factor of one thousand to ten thousand. But, if one hundred thousand and ten million both small numbers as far as the human brain is concerned, they’re big numbers when it comes to building a machine, as we can readily see. It is because of the size of such numbers that we know that the memory cells of our thinking machine will have to be small, and, we might add, cheap.
For instance, some present-day computers use relays as memory cells. Now, a good and reliable relay, one good enough to avoid frequent failure even when many thousands of relays are used, costs perhaps two dollars. If we wanted a million cells, the cost of the relays would thus be two million dollars, and this is an unpleasant thought to start with. Further, one would probably mount about a thousand relays on one relay rack, and so there would be a thousand relay racks. These could perhaps be packed into a space of about six thousand square feet – around eighty by eighty feet. Then, there would have to be quite a lot of associated equipment, for more relays would be needed to make a connection to a given memory cell and to utilize the information in it. This would increase the cost and the space occupied a good deal. The thing isn’t physically possible, but it seems an unpromising start if we wish to advance further toward the at least ten sand-fold greater complexity of the human brain.
Fortunately, at just the time it as needed, something better than the relay has come along. That something, the possessor of the little blue cells, is the selectron. It is a vacuum tube which can serve in the place of several thousand relays. It promises to be reliable, small and, dually, at least, cheaper than relays, and in addition it is very much faster – perhaps a thousand-fold. The selectron was invented by an engineer, Dr. Jan A. Rajchman – pronounced Rikeman – for the purpose of making an improved computer and so its appearance at the right time is, after all, no accident. Instead, it is a tribute to Dr. Rajchman’s great inventive ability. Lots of people who worked on computers knew what the problem was, but only he thought of the selectron.
You might wonder how to go about inventing just what is needed, and if Dr. Rajchman’s career can cast any light on this, it’s certainly worth looking into. Did he, for instance, think about computers from his earliest technical infancy? The answer is that he certainly didn’t. I have a copy of his doctoral thesis, “Le Courant Résiduel dans Les Multiplicateurs D’ElectronsElectrostatiques,” which tells me that he was born in London in 1911, that he took his degree at Le Ecole Polytechnique Federale, at Zurich and thereafter did research on a radically new type of electrically focused photo-multiplier – see “Universes to Order,” in Astounding for February, 1944. I am not sure how many different problems he has worked on since, but during the war he did do some very high-powered theoretical work on the betatron, as well as some experimental work on the same device. It would seem that the best preparation for inventing is just to become thoroughly competent in things allied to the field in which something new is needed.
What was needed in connects with computers was, as we have said, a memory cell, or, rather, lot, of them. What do these cells have to do? First of all, one must be able to locate a given cell in the memory so as to put information into it or take information out. Then, one must be able to put into the cell the equivalent of a 0 or a 1. One must have this stay there indefinitely, until it is deliberately changed. Finally, one must be able to read off what is stored in the cell; one must be able to tell whether it signifies 0 or 1 without altering what is in the cell. The selectron has these features.
You might be interested in some of the earlier suggestions for using an electron tube as a memory in a computing machine. The electron beam of a cathode ray tube sounds like just the thing for locating a piece of information, for instance. One has merely to deflect it the right amount horizontally and vertically to reach a given spot on the screen of the tube. One wishes, however, to store a particular piece of information in a particular place and then to find that same place again and retrieve that same piece of information. This would mean producing the exact voltages on the deflecting plates when the formation was stored, and that is by no means easy. Further, if the accelerating voltage applied to the tube changes, the deflecting voltage needed to deflect the beam to a given place changes, and this adds difficulty. When we realize further that our memory simply must not make mistakes, we see that there are real objections -to locating and relocating a given spot by simply deflecting an electron beam to it. The selectron has a radically different means for getting electrons to a selected spot – the selectron grid.
The features of the selectron which Dr. Rajchman holds in his hand – page 163 – are illustrated simply in Figure 1. There is a central cathode and around it a concentric accelerating grid. When this grid is made positive with respect to the cathode, a stream of electrons floods the entire selectron grid, the next element beyond the accelerating grid. The selectron grid, is made up of a number of thin bars located in a circular array, pointing radially outward, and a number of thin rings, spaced the same distance apart as are the bars. Figure 2 shows a portion of the selectron grid formed by the rings and bars. The rings and bars together form a number of little rectangular openings or windows.
____________________
Figure 1. A cross-section view of the selectron. Information is stored as a voltage on the inner side of the insulating storage surface. The voltage is established by electron streams flowing through the “windows” formed by the bars and rings of the selectron grid. Such electron streams are also used in reading the information off.
____________________
Figure 2. A perspective view showing the arrangement of bars and rings forming the selectron grid and its windows.
____________________
Now, in operation each bar and ring of the selectron grid is held either several hundred volts positive with respect to the cathode, or else a little negative with respect to the cathode. After a definite pattern of ages has been established on the selectron grid, the accelerating grid is made positive and the selectron grid is flooded with electrons. What happens? Let us consider first the bars of the selectron grid. Figure 3 tells the story. If two neighboring bars are negative, the approaching electrons are simply repelled and turned back. If an electron enters the space between a positive bar and a negative bar, it is so strongly attracted toward the positive bar that it strikes it and is lost. Only if the bars on both sides of the space which the electron enters are positive does the electron get through. At the rings, the story is the same; an electron can pass between two rings only if both are positive; it is stopped if either one or both are negative. Thus we conclude that electrons can pass through a little window formed by two bars and two rings only if both bars and both rings are positive. If both bars and both rings forming a window are held positive, the window is open; if one or more of the bars or rings are negative, the window is closed. Thus, we have a means for letting electrons through one window at a time.
____________________
Figure 3. Electrons can pass between two bars or rings only if both are positive. If both are negative, the electron is turned back. If one is negative, the electron is deflected and lost on the positive bar or ring.
____________________
In the early model selectrons there were sixty-four apertures between bars around the tube, and sixty-four apertures lengthwise, giving four thousand ninety-six windows in all, and any one of these could be selected for the passage of electrons by applying proper voltages to the bars and rings. Does this mean that we must have one hundred twenty-eight leads into the tube for this alone, one for each bar and one for each ring? The tube would certainly work if it had one hundred twenty-eight leads to the selectron grid, but Dr. Rajchman’s ingenuity has cut this down instead to thirty-two, a saving by a factor of four. How is this done? The table of Figure 4 tells the story. Here we have in the top row the numbers of the bars, in order, sixty-four in all. These bars are connected to two sets of eight leads. The second and third rows show to which lead of a given set a bar is connected. Thus, Bar 1 is connected to Lead 1 of Set I. Bar 2 is connected to Lead 1 of Set II, while Bar 64 is connected to Lead 8 of Set II. To save space, some of the bars have been omitted from the table.
____________________
Figure 4. The sixty-four bars are connected to two sets of eight leads in the fashion shown. By making one lead of each set positive and the others negative, it is possible to make any pair of adjacent bars positive and at the same time have no other adjacent pair positive.
____________________
You will observe that if we make Lead 7 of Set I positive, and all the rest of the leads of Set I negative, Bars 13, 29, 45 and 61 will be positive. Then, if we make Lead 2 of Set II positive and all the other leads of Set II negative Bars 4, 8,12 and 16 will be positive. All the bars which do not appear in either of the above listings will be negative. Now, the only adjacent bars listed are 12 and 13, which have been written in italics. Hence, when Lead 7 of Set I and Lead 2 of Set II are made positive and all the other leads negative, electrons can pass between the two adjacent positive bars 12 and 13, but not between any other bars. Thus, by selecting one lead from Set I and one lead from Set II, we can select any of the sixty-four spaces between bars.
The thoughtful reader will have noticed, by the way, that there are only sixty-three spaces between sixty-four bars. This, however, omits the space out to infinity from Bar 1 and back from infinity to Bar 64. We can in effect shorten this space by adding an extra bar beyond the sixty-fourth and connecting it to Bar 1.
The same sort of connection used with the bars is made to the ring so that by selecting and making positive one lead each in two sets of eight leads we can select any of the sixty-four spaces between rings. Thus, in the end we have four sets of eight leads each, two sets the bars and two for the rings. We make positive one wire in each set at a time. The number of possible combinations we can get this way is four thousand ninety-six, and each allows electrons to go through just one window out of the four thousand ninety-six formed by the bars and rings of the selectron grid. The action is entirely positive. A given window is physically located in a given place. Small fluctuations in the voltages applied to the bars and rings will not interfere with the desired operation. This is a lot different from trying to locate a given spot by waving an electron beam around.
The selectron grid and its action are- of course, only a part of the mysteries of the selectron. They provide a means for directing a stream of electrons through one of several thousand little apertures at will. But, how can this stream of electrons be used in storing a signal and then in reading it off again? Part of the answer is not new. For some time electronic experts have n thinking of storing a signal on an insulating surface as an electric charge deposited on the surface by means of an electron stream. Thus, by putting electrons on a sheet of mica, for instance, we can make the surface negative, and by taking them off we can make it positive. It is easy enough to do either of these things, as we shall see in a moment.
There are two very serious difficulties with, such a scheme, however. First, how shall we keep the positive or negative charge on the insulating surface indefinitely? It will inevitably tend to leak off. Second, how can we determine whether the surface is charged positively or negatively without disturbing the charge? The logical exploring tool is an electron beam, but won’t the beam drain the charge off in the charge off in the very act of exploration? Both of these difficulties are overcome in the selectron. To understand how, we must know a little about secondary emission.
Beyond the accelerating and selectron grids of the selectron, as shown in Figure 1, there is a sheet of mica indicated as “storage surface.” This has a conducting backing. We are interested in what happens when electrons pass through an open window in the selectron grid – one made up of four positive bars and rings – and strike the mica. The essential ingredients of the situation are illustrated in the simplified drawing of Figure 5. Here the accelerating grid and the selectron grid are lumped together and shown as positive with respect to the cathode. Electrons are accelerated from the cathode, pass through the accelerating grid and the open window of the selectron grid, and shoot toward the mica storage surface. What happens? That depends on the potential of the storage surface with respect to the cathode.
____________________
Figure 5. When a window in the selectron grid is open – the bars and rings on all sides positive – electrons shoot through it toward the storage surface. What happens to the electrons depends on the potential of the storage surface with respect to the cathode. The potential of the storage surface is controlled by the flow of electrons to and from it, and by the potential of the conducting backing plate.
____________________
In Figure 6 the current reaching the part of the storage surface behind an open window is plotted vs. the potential of that part of the storage surface with respect to the cathode. Potential is negative with respect to the cathode to the left of the vertical axis and positive with respect to the cathode to the right of the vertical axis. Current to the storage surface is negative – electrons reaching the surface and sticking below the horizontal axis and positive – more electrons leaving the surface than reaching it – above the horizontal axis. The curve shows how current to the surface varies as the potential of the surface is varied.
____________________
Figure 6. If the storage surface is negative with respect to the cathode, no electrons each it – 0 current. If it is a little positive, electrons reach it but few leave – negative current. If it is more positive, secondary electrons leave, and if it is more positive than some potential V0 more secondaries leave than primaries strike – positive current. If the storage surface is at s higher potential than VS the potential of the selectron grid, the secondaries which leave are turned back – negative current to the storage surface.
____________________
If the surface is negative with respect to the cathode, the electrons shot toward it are turned back before they reach it and the current to the surface is zero. If the surface is just a little positive, the electrons shot toward it are slowed down by the retarding field between the very positive selectron grid and the much less positive storage surface, and they strike the surface feebly and stick, constituting a negative-current flow to the surface, and tending to make the surface more negative. If the potential of the storage surface is a little more positive with respect to the cathode, the electrons reach it with enough energy to knock a few electrons out of it. These are whisked away to the more positive selectron grid. These negative electrons leaving the surface are equivalent to a positive current to the surface. There are now as many electrons striking as before, but there are also some leaving, and there is less net negative current to the surface. Finally, at some potential labeled V0 in Figure 6, one secondary electron is driven from the surface for each primary electron which strikes it, and the net current to the surface is zero. If the potential of the storage surface is higher than V0, each primary electron releases more than one secondary and there is a net flow of electrons away from the surface, equivalent to a positive current to the surface. This tends to make the storage surface more positive.
As the potential of the storage surface rises further above V0, current for a time becomes more and more positive. Then, abruptly the neighborhood of the potential VS of the selectron grid itself, the current becomes negative again and stays negative. Why is this? The the primary electrons still strike the storage surface energetically and drive out more than one electron each. The fact is that these secondary electrons leave the surface with very little speed. When the storage surface is more positive than the selectron grid, there is a retarding field at the storage surface which tends to turn the secondaries back toward the storage surface. Hence, there, is still a flow of primaries – a negative current – to the surface, but the secondaries are turned back before reaching the selectron grid and fall on the storage surface again. Thus, the current to the storage surface is again negative.
Our mechanism for holding the storage surface positive or negative is immediately apparent from Figure 6. If the surface is more positive than Vs, the current to it is negative and its potential will tend to fall. If the surface has a potential between V0 and Vs, the current to it is positive and its potential will tend to rise. Hence, if the storage surface initially has any potential higher than V0, current will flow to it in such a way as to tend to make its potential VS, the potential of the selectron grid. If, on the other hand, the potential is between 0 and V0, the current to the surface will be negative and the potential of the surface will tend to fall to 0. If the potential of the surface is negative with respect to the cathode – less than 0 – there is no current to it from the electron stream and hence no tendency for the potential to rise and fall. Actually, some leakage would probably result in 3 very slight tendency for the potential to rise.
We see, then, that when it is bombarded by electrons, a part of the storage surface tends naturally to assume one of two potentials, or VS or 0. If it has initially any other potential, it tends to come back to one of these. Which potential it assumes is determined by whether the initial potential is greater or less than V0. Thus, if we store information on the part of the storage surface behind a particular window by making this area have a potential Vs with respect to the cathode – meaning, say, 1 – or 0 – meaning, 0 – and if this potential changes a little through electrical leakage, perhaps adjacent portions at a different potential, we can recover or regenerate the original potential merely by opening the window of the selectron grid and flooding the area with electrons. In fact, we can periodically regenerate the potentials behind all windows by opening all windows at once and flooding the whole surface with electrons. This is what is done in the operation of the selectron, and this regenerative feature, which makes it possible to retain the stored information indefinitely despite electrical leakage, is one of the most ingenious and important features of the selectron.
How do we get the information on the portions of the storage surface behind the various windows? That is, how do we initially bring some portions of the surface to the potential Vs and others to the potential V0? In this process of writing inflation into the tube, we first open the particular one of the four thousand ninety-six windows behind which we wish to store a particular piece of information, thus flooding a little portion of the surface with electrons. Then, to the terminals T of Figure 5, between the cathode and the conducting backstage of the storage surface, we apply a very sharply rising positive pulse, shown as the dashed line of Figure 7. Because of the capacitance between this backing plate and the front of the storage surface, where the electrons fall, this drives the front of the storage surface positive. Then the pulse applied to the conducting backing falls slowly to zero, as shown. However, the action of the electrons falling on the surface tends to make it assume the potential Vs, and so if the pulse falls off slowly enough the portion of the surface on which electrons fall is left at the potential Vs, as shown by the solid line of Figure 7. Application of the pulse will leave the portion of the storage surface behind the open window at the potential Vs regardless of whether its initial potential is Vs or 0, and the pulse will not affect portions of the surface behind closed windows, because no electrons reach them.
____________________
Figure 7. To make an element of the storage surface assume a potential VS, its window is opened, it is flooded with electrons, and a sharp pulse is applied to the conducting backing. This drives the surface positive through capacitive coupling. The pulse is allowed to fall gradually to 0 – dashed curve. The surface at first falls with the pulse, but the action of the electron stream tends to hold it at a potential VS. A sharp negative pulse will leave the surface at 0 potential.
____________________
This tells us how we can bring any selected area of the storage surface to the potential Vs which, we can say, corresponds to writing 1 in a particular cell of this memory tube. By flooding a given area or cell with electrons and applying a sharply falling, negative pulse, which rises again gradually toward 0 – the dashed pulse of Figure 7 upside down – we can bring any selected area of the storage surface to 0 potential, and thus write 0 in any selected cell of the memory.
Thus, each little area of the storage surface behind each window of the selectron grid is a cell of our memory. By opening a particular window – through making one lead of each of the four sets of eight selectron grid leads positive – and pulsing the conducting backing positive or negative, we can make the little area of the storage surface behind that window assume a potential Vs or a potential 0, and so can, in effect, write 1 or 0 in that particular memory cell. By opening all windows periodically and flooding all areas with electrons, we can periodically bring all little areas back to their proper potentials, either VS or 0, despite leakage of electrons to or away from the little areas. We can, that is, put thousands of pieces of information into the selectron and keep them there. What about reading? How can we get this information out?
Imagine that the entire inner storage surface is covered with a phosphor or fluorescent material like that used on cathode-ray tube screens or inside of fluorescent lights. Now, suppose we open one window of the selectron, shooting electrons at a particular area of the surface. If that area has a potential 0, the electrons will be repelled from it. But, if that area has a potential Vs, corresponding to 1, the electrons will strike the fluorescent surface vigorously, emitting a glow of blue light. Suppose we let this light fall on a photo-multiplier, of the type Dr. Rajchman worked on earlier in his career. Then, when we open a given window of the selectron, if the potential of the surface behind the window is 0, we get nothing out of the multiplier. But, if the potential is Vs, there is a flash of light, and a pulse of current from the multiplier. And so, we can not only write a 0 or a 1 in each little memory cell of the selectron, we can not only keep this information there indefinitely, but we can also read it off at will.
Dr. Rajchman has devised other ways for reading the stored information in the selectron, but the use of a phosphor-coated storage surface together with a photo-multiplier has been one of the preferred method. I have spoken of the phosphor as one giving blue light. This is because the photo-multiplier is more sensitive to blue light than to other colors. And so, I predicted that the memory cells of the thinking machines will be not only multitudinous and small, but also blue.
Of course the selectron provides only a part of the thinking machine – that is, the memory. Associated with it there must be circuits in tubes to seek out stored in tubes to seek out stored information, to make use of it to obtain new formation, to write in that new information, and to make use of the new information in turn. All is a field apart. Still, there is one wrinkle which is so intimately connected with the use of the selectron that it deserves mention here. I have referred to the 0 or 1 a cell of the selectron which can tore a binary digit or, alternately, as a letter of the electronic alphabet which the machine understands. Now, usually we don’t want to store isolated digits or letters: we want to store complete numbers or words – combinations of 1 and 0, as, 10011. This is 19 in binary notation, and might in some instance stand for the nineteenth word in a dictionary. When we look up a number or a word, we want it all at once, not piecemeal.
When we want to write many multi-digit numbers in a book, as, in a table of logarithms, for instance, we usually assign a vertical column for each digit to be stored, and write each digit of a given number in a different column, along the same row. Thus, entries in a log table appear as in Figure 8. Suppose that in using the selectron we assign a different tube to each binary digit of the numbers to be stored. If we wish to store twenty-digit numbers, we will need twenty tubes. Each tube will, in effect, be a given column of our storage space. The different cells in a tube will represent different rows. Thus, Cell 1 of Tube 1 will be Row 1 Column 1, Cell 1 of Tube 2 will be Row 1 Column 2, while Cell 10 of Tube 1 will be Row 10 Column 1, et cetera.
____________________
Figure 8. In storing multi-digit numbers in a table, we write the different digits of a given number in different columns, so that all of a given number will lie along a horizontal row, as in the log table above. In storing binary digits of multi-digit binary numbers using electrons, a separate selectron is provided to represent each column. The rows are represented by the different windows. Thus, the first window of the first selectron is Row 1 Column 1, while the first window of the second selectron is Row 1 Column 2, et cetera.
____________________
We want to look up all the digits in a given row at once. This means that we want to open corresponding windows in all the tubes at once, and so we can connect the corresponding selectron grid leads of all twenty tubes together. Thus, if want to store a number in Row 1, we apply voltages to the selectron grid leads which will open Window 1 in all tubes. We are then ready to read the number in Row 1 or to write a new number in. The twenty photo-multipliers which read the twenty selectrons are not connected in parallel, but are connected separately to carry off the twenty digits of the number in Row 1 to their proper destinations. Perhaps these twenty leads from the twenty photo-multipliers may go to the twenty backing plates of another twenty selectrons to which it is desired to transfer the number. We see, thus, how a whole table of numbers can be stored in twenty selectrons. The windows 1, 2, 3 et cetera, can represent, for instance, the angle of which we want the sine. The first selectron can store the first digits of all the sines, the second selectron can store all the second digits, et cetera. The twenty digits of the sine of any angle – any window number – can be read off simultaneously from the photo-multipliers of the twenty selectrons.
The selectron isn’t perfect by any means. Perhaps it’s not even the final answer. At the moment, in its early form, it may be almost expensive as relays, but that’s partly because it’s new. It’s certainly great deal more compact than relays, a very great deal faster, and probably more reliable as well. It represents a first huge stride in the electronics of the thinking machine. Just how far it takes us is up to a lot of mathematicians, a lot of circuit gadgeteers, and, especially, to Dr. Jan A. Rajchman and RCA, to whom we must look for smaller, cheaper and better selectrons.
This has long been so in the realm of science fiction, a striking example of which – perhaps arising from equal measures and intuition and imagination – appearing in Astounding Science Fiction in mid-1949. That year, Eric Frank Russell’s three-part serial “Dreadful Sanctuary” was serialized in the June, July, and August issues of the magazine.
__________
June, 1948: Cover by William F. Timmins.
(Note Timmins’ name on the “puzzle piece” in the lower left corner!)
__________
July, 1948: Cover by Chesley Bonestell
__________
August, 1948: Cover by: Alexander Cañedo
__________
With interior illustrations by Timmins, the story, set in 1972, is centered upon the efforts of one John J. Armstrong – an iconoclastic combination of entrepreneur, inventor, and unintended detective – to accomplish the first successful manned lunar landing as his entirely private venture, in the face the inexplicable mid-flight destruction of each of his organization’s spacecraft. Armstrong doesn’t fit the cultural stereotype of inventor or scientist. As characterized by Russell, “Armstrong was a big, tweedy man, burly, broad-shouldered and a heavy punisher of thick-soled shoes. His thinking had a deliberate, ponderous quality. He got places with the same unracy, deceptive speed as a railroad locomotive, but was less noisy.”
While Russell’s story commences in the June issue as a solid – and solidly intriguing – mystery, effectively conveying a sense of wonder; with characters who portend to be more than two-dimensional; the events, plot, and underlying tone gradually change. With the installments in Astounding’s July and and August issues, what had been a tale with an eerie undertone of Fortean inexplicability, technical conjecture (such as the “ipsophone”, a video-telephone imbued with aspects of artificial intelligence – cool! – we’re talking 1948!), and a well-crafted mood of impending threat … gradually and steadily falls flat.
A pity, because to the extent that the story succeeds – and in parts it does succeed, and creatively at that – it does so far more as a hard-boiled (and very ham-fisted) detective tale than science-fiction.
Regardless of the story’s literary quality (I don’t think it’s ever been anthologized) the physical and psychological presence of the aptly named Armstrong (“arm”?! “strong”?! get it??!) remain consistent throughout. Iconoclastic and independent, he’s extremely intelligent, and if need be, a man capable of brute intimidation, self-defense, and violence. He’s also canny, cunning, and psychologically astute. It’s these latter qualities that lead to Armstrong’s discovery – after meeting a police captain – of a most intriguing device, at his residence in the suburbs of New York City.
Correctly suspicious of surveillance by adversaries, on reaching his residence, “…Armstrong cautiously locked himself in, gave the place the once-over.
“Knowing the microphone was there, it didn’t take him long to find it though its discovery proved far more difficult than he’d expected.
“Its hiding place was ingenious enough – a one hundred watt bulb had been extracted from his reading lamp, another and more peculiar bulb fitted in its place.
“It was not until he removed the lamp’s parchment shade that the substitution became apparent.
“Twisting the bulb out of its socket, he examined it keenly.
“It had a dual coiled-coil filament which lit up in normal manner, but its glass envelope was only half the usual size and its plastic base twice the accepted length.
“He smashed the bulb in the fireplace, cracked open the plastic base with the heel of his shoe.
“Splitting wide, the base revealed a closely packed mass of components so extremely tiny that their construction and assembling must have been done under magnification – a highly-skilled watchmaker’s job! The main wires feeding the camouflaging filament ran past either side of this midget apparatus, making no direct connection therewith, but a shiny, spider-thread inductance not as long as a pin was coiled around one wire and derived power from it.
__________
(July, 1948, page 101)
__________
“Since there was no external wiring connecting this strange junk with a distant earpiece, and since its Lilliputian output could hardly be impressed upon and extracted from the power mains, there was nothing for it than to presume that it was some sort of screwy converter which turned audio-frequencies into radio or other unimaginable frequencies picked up by listening apparatus fairly close to hand.
“Without subjecting it to laboratory tests, its extreme range was sheer guesswork, but Armstrong was willing to concede it two hundred yards.
“So microscopic was the lay-out that he could examine it only with difficulty, but he could discern enough to decide that this was no tiny but simple transmitter recognizable in terms of Earthly practice.
“The little there was of it appeared outlandish, for its thermionic control was a splinter of flame-specked crystal, resembling pin-fire opal, around which the midget components were clustered.” (July, 1948, pp.116-117)
I’ll not explain the origin of this device (it’d spoil the story should you read it!), but suffice to say that in the world of the “Dreadful Sanctuary”, things and people are not as they seem, in terms of their origin, nature, and purpose.
In our world, however, it seems that Eric Frank Russell created a literary illustration – at least in terms of its diminutive size and the delicacy of its fabrication – of what would in only a few years be known as the integrated circuit.
