Though by definition and nature a science fiction<\/em> publication, Astounding <\/em>(akin to its post-WW II counterparts and rivals Galaxy Science Fiction<\/em>, and, The Magazine of Fantasy and Science Fiction (“F&SF”)<\/em>) also published non-fiction material.\u00a0 Such non-fiction material included leading editorials, book reviews, and letters, as well as articles – typically, one per issue – about some aspect of the sciences.\u00a0 As in 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<\/em> 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.\u00a0 (Biology it seems, not so much!)\u00a0 Viewed as a whole, these subject areas\u00a0 – 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.<\/span><\/p>\n Here are the (non-fiction) science articles that were published in Astounding Science Fiction <\/em>in 1949:<\/span><\/p>\n January: “Modern Calculators<\/strong>” (Digital and analog calculation<\/em>), by E.L. Locke; pp. 87-106<\/span><\/p>\n February: “The Little Blue Cells<\/strong>” (The “Selectron” data storage tube<\/em>), by J.J. Coupling; pp. 85-99<\/span><\/p>\n March: “The Case of the Missing Octane<\/strong>” (Chemistry of petroleum and gasoline<\/em>), by Arthur Dugan; pp. 102-113 (Great caricatures by Edward Cartier!) April: “9 F 19<\/strong>” (Hydrocarbons<\/em>), by Arthur C. Parlett; pp. 46-162<\/span><\/p>\n May: “Electrical Mathematicians<\/strong>” (Machine (electronic) calculation<\/em>), by Lorne MacLaughlan; pp. 93-108<\/span><\/p>\n ______________________________<\/span><\/p>\n June: “The Aphrodite Project<\/strong>” (Determining the mass of the planet Venus<\/em>), by Philip Latham; pp. 73-84. (Intriguing cover art by Chesley Bonestell.) July: “Talking on Pulses<\/strong>” (Electronic transmission of human speech and other forms of communication<\/em>), by C. Rudmore; pp. 105-116.<\/span><\/p>\n August: “Coded Speech<\/strong>” (Electronic speech; noise reduction<\/em>), by C. Rudmore; pp. 134-145<\/span><\/p>\n September: “Cybernetics<\/strong>” (Review of Norbert Wiener’s book by the same title<\/em>), by E.L. Locke; pp. 78-87<\/span><\/p>\n October – First article: “Chance Remarks<\/strong>” (Communication research<\/em>), by J.J. Coupling; pp. 104-111<\/span><\/p>\n October – Second article: “The Great Floods<\/strong>” (Review of great floods in human history<\/em>), by L. Sprague de Camp; pp. 112-120<\/span><\/p>\n November: “The Time of Your Life<\/strong>” (Time; Determining the length of the earth’s day<\/em>), by R.S. Richardson; pp. 110-121<\/span><\/p>\n December – First article: “Bacterial Time Bomb<\/strong>“, by Arthur Dugan; pp. 93-95<\/span><\/p>\n December – Second article:\u00a0 “Science and Pravda<\/strong>“, by Willy Ley; pp. 96-111<\/span><\/p>\n Regardless of the topic, a notable aspect of the non-fiction science content of Astounding<\/em> (likewise for Galaxy<\/em> and F&SF<\/em>) is that mathematics – in terms of equations and formulae, let alone Cartesian graphs – was kept to a minimum, if not eschewed altogether.\u00a0 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 material.\u00a0 <\/span><\/p>\n One such example – from February of 1949 – is presented below, in the form of J.J. Coupling’s article “The Little Blue Cells”.\u00a0 <\/span><\/p>\n ______________________________<\/span><\/p>\n This issue features great cover art by Hubert Rogers for Jack Williamson’s (writing under the pen-name “Will Stewart”) serial “Seetee Shock”.\u00a0 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”.\u00a0 Like much of Rogers’ best work, symbolism is as important as representation.\u00a0 (You can enjoy more of Rogers’ work at my brother blog, WordsEnvisioned<\/a>.) \u00a0\u00a0\u00a0 Coupling’s article is notable because it addresses a subject frequently addressed by Astounding<\/em>, with continuing and likely indefinite relevance: recording, storing, preserving, and accessing information – computer memory. \u00a0\u00a0\u00a0\u00a0\u00a0 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<\/a>.\u00a0 As implied and admitted by Coupling’s article, even at the time of the device’s invention there was ambivalence about its long-term economic and technical viability, despite its functionality and innovative design.\u00a0 <\/span><\/p>\n \u00a0\u00a0\u00a0\u00a0 An image of a Selectron Tube, from Giorgio Basile’s Lamps & Tubes<\/a>, is shown below.\u00a0 (Scroll down to end of post for a photograph showing a Selectron Tube in the hands of its inventor, illustrating its relative size.) \u00a0\u00a0\u00a0\u00a0\u00a0 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.\u00a0 To no avail.\u00a0 Both tube designs were superseded by magnetic core memory in the early 1950s.\u00a0 <\/span><\/p>\n \u00a0\u00a0\u00a0\u00a0 As for J.J. Coupling?\u00a0 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, <\/em>and other publications.\u00a0 His lengthy oeuvre is listed at The Internet Speculative Fiction Database<\/a>. \u00a0\u00a0\u00a0\u00a0 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.\u00a0 His article, with accompanying illustrations, is presented below. ______________________________<\/span><\/p>\n THE LITTLE BLUE CELLS The <\/em>most acute problem in the design of a robot, a thinking machine, or any of the self-serving devices of science-fiction is memory.\u00a0 We can make the robot\u2019s body, its sensory equipment, its muscles and limbs.\u00a0 But thinking requires association of remembered data; memory is the essential key.\u00a0 So we present the Little Blue Cells!<\/em><\/p>\n 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.\u00a0 The small, self-portable models I admire from a distance, but I feel no temptation to speculate about their inner secrets.\u00a0 The workings of the big thinking machines have intrigued me, however.\u00a0 It used to be that I didn\u2019t know whether to believe in them or not.\u00a0 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.\u00a0 With the other readers of Astounding I\u2019ve had a sort of inducted tour through the brain cases of these monsters in \u201cModern Computing Devices\u201d by E.L. Locke.\u00a0 I\u2019m pretty much convinced.\u00a0 It\u2019s beginning to look as if we\u2019ll know the first robot well long before he\u2019s born.<\/p>\n Perhaps some readers of science fiction can look back to the old, unenlightened days and remember a prophetic story called, I believe, \u201cThe Thinking Machine.\u201d \u00a0The inventor of that epoch had first to devise an \u201celectronic language\u201d before he could build his electrical cogitator.\u00a0 The modern thinking machine of the digital computer type comes equipped with a special electronic alphabet and vocabulary if not with a complete language.\u00a0 The alphabet has the characters off<\/em> and on<\/em>, or 0 and 1, the digits of the binary system of enumeration, and words must certainly be of the form 1001-110\u2014and so on.\u00a0 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.\u00a0 The machine\u2019s most abstruse thought, or its fondest recollection – if such machines eventually come to have emotions – will be stored away as off\u2019s and on\u2019s in the multitudinous blue cells of the device\u2019s memory.<\/p>\n I\u2019m sure that I\u2019m right in describing the memory cells of the machine as multitudinous and little – that is, if it\u2019s a machine of any capabilities at all.\u00a0 To describe them as blue is perhaps guessing against considerable odds, but there are reasons even for this seemingly unlikely prognostication.<\/p>\n The multitudinous part is, I think, obvious.\u00a0 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.\u00a0 The human brain, for instance, has around ten billion nerve cells.\u00a0 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.\u00a0 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.\u00a0 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.\u00a0 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.\u00a0 But, if one hundred thousand and ten million both small numbers as far as the human brain is concerned, they\u2019re big numbers when it comes to building a machine, as we can readily see.\u00a0 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.<\/p>\n For instance, some present-day computers use relays as memory cells.\u00a0 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.\u00a0 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.\u00a0 Further, one would probably mount about a thousand relays on one relay rack, and so there would be a thousand relay racks.\u00a0 These could perhaps be packed into a space of about six thousand square feet – around eighty by eighty feet.\u00a0 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.\u00a0 This would increase the cost and the space occupied a good deal.\u00a0 The thing isn\u2019t 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.<\/p>\n Fortunately, at just the time it as needed, something better than the relay has come along.\u00a0 That something, the possessor of the little blue cells, is the selectron<\/em>.\u00a0 It is a vacuum tube which can serve in the place of several thousand relays.\u00a0 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.\u00a0 The selectron was invented by an engineer, Dr. Jan A. Rajchman – pronounced Rikeman<\/em> – for the purpose of making an improved computer and so its appearance at the right time is, after all, no accident.\u00a0 Instead, it is a tribute to Dr. Rajchman\u2019s great inventive ability.\u00a0 Lots of people who worked on computers knew what the problem was, but only he thought of the selectron.<\/p>\n You might wonder how to go about inventing just what is needed, and if Dr. Rajchman\u2019s career can cast any light on this, it\u2019s certainly worth looking into.\u00a0 Did he, for instance, think about computers from his earliest technical infancy? \u00a0The answer is that he certainly didn\u2019t.\u00a0 I have a copy of his doctoral thesis, \u201cLe Courant R\u00e9siduel dans Les Multiplicateurs D\u2019Electrons<\/em> Electrostatiques<\/em>,\u201d 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 \u201cUniverses to Order,\u201d in Astounding for February, 1944.\u00a0 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. \u00a0It 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.<\/p>\n What was needed in connects with computers was, as we have said, a memory cell, or, rather, lot, of them.\u00a0 What do these cells have to do? \u00a0First 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.\u00a0 Then, one must be able to put into the cell the equivalent of a 0 or a 1.\u00a0 One must have this stay there indefinitely, until it is deliberately changed.\u00a0 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<\/em>.\u00a0 The selectron has these features.<\/p>\n You might be interested in some of the earlier suggestions for using an electron tube as a memory in a computing machine.\u00a0 The electron beam of a cathode ray tube sounds like just the thing for locating a piece of information, for instance.\u00a0 One has merely to deflect it the right amount horizontally and vertically to reach a given spot on the screen of the tube.\u00a0 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.\u00a0 This would mean producing the exact voltages on the deflecting plates when the formation was stored, and that is by no means easy.\u00a0 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.\u00a0 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.\u00a0 The selectron has a radically different means for getting electrons to a selected spot – the selectron grid.<\/p>\n The features of the selectron which Dr. Rajchman holds in his hand – page 163 – are illustrated simply in Figure 1.\u00a0 There is a central cathode and around it a concentric accelerating grid.\u00a0 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.\u00a0 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.\u00a0 Figure 2 shows a portion of the selectron grid formed by the rings and bars.\u00a0 The rings and bars together form a number of little rectangular openings or windows.<\/p>\n The thoughtful reader will have noticed, by the way, that there are only sixty-three spaces between sixty-four bars.\u00a0 This, however, omits the space out to infinity from Bar 1 and back from infinity to Bar 64.\u00a0 We can in effect shorten this space by adding an extra bar beyond the sixty-fourth and connecting it to Bar 1.<\/p>\n 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. \u00a0Thus, in the end we have four sets of eight leads each, two sets the bars and two for the rings.\u00a0 We make positive one wire in each set at a time.\u00a0 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.\u00a0 The action is entirely positive.\u00a0 A given window is physically located in a given place.\u00a0 Small fluctuations in the voltages applied to the bars and rings will not interfere with the desired operation.\u00a0 This is a lot different from trying to locate a given spot by waving an electron beam around.<\/p>\n The selectron grid and its action are- of course, only a part of the mysteries of the selectron.\u00a0 They provide a means for directing a stream of electrons through one of several thousand little apertures at will. \u00a0But, how can this stream of electrons be used in storing a signal and then in reading it off again? \u00a0Part of the answer is not new.\u00a0 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.\u00a0 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.\u00a0 It is easy enough to do either of these things, as we shall see in a moment.<\/p>\n There are two very serious difficulties with, such a scheme, however.\u00a0 First, how shall we keep the positive or negative charge on the insulating surface indefinitely? \u00a0It will inevitably tend to leak off.\u00a0 Second, how can we determine whether the surface is charged positively or negatively without disturbing the charge? \u00a0The logical exploring tool is an electron beam, but won\u2019t the beam drain the charge off in the charge off in the very act of exploration? \u00a0Both of these difficulties are overcome in the selectron.\u00a0 To understand how, we must know a little about secondary emission.<\/p>\n Beyond the accelerating and selectron grids of the selectron, as shown in Figure 1, there is a sheet of mica indicated as \u201cstorage surface.\u201d \u00a0This has a conducting backing.\u00a0 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.\u00a0 The essential ingredients of the situation are illustrated in the simplified drawing of Figure 5.\u00a0 Here the accelerating grid and the selectron grid are lumped together and shown as positive with respect to the cathode.\u00a0 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.\u00a0 What happens? \u00a0That depends on the potential of the storage surface with respect to the cathode.<\/p>\n As the potential of the storage surface rises further above V0<\/sub>, current for a time becomes more and more positive.\u00a0 Then, abruptly the neighborhood of the potential VS<\/sub> of the selectron grid itself, the current becomes negative again and stays negative.\u00a0 Why is this?\u00a0 The the primary electrons still strike the storage surface energetically and drive out more than one electron each.\u00a0 The fact is that these secondary electrons leave the surface with very little speed.\u00a0 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.\u00a0 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.\u00a0 Thus, the current to the storage surface is again negative.<\/p>\n Our mechanism for holding the storage surface positive or negative is immediately apparent from Figure 6.\u00a0 If the surface is more positive than Vs, the current to it is negative and its potential will tend to fall.\u00a0 If the surface has a potential between V0<\/sub> and Vs, the current to it is positive and its potential will tend to rise.\u00a0 Hence, if the storage surface initially has any potential higher than V0<\/sub>, current will flow to it in such a way as to tend to make its potential VB, the potential of the selectron grid.\u00a0 If, on the other hand, the potential is between O and V0<\/sub>, the current to the surface will be negative and the potential of the surface will tend to fall to O.\u00a0 If the potential of the surface is negative with respect to the cathode – less than O – there is no current to it from the electron stream and hence no tendency for the potential to rise and fall.\u00a0 Actually, some leakage would probably result in 3 very slight tendency for the potential to rise.<\/p>\n 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<\/sub> O.\u00a0 If it has initially any other potential, it tends to come back to one of these.\u00a0 Which potential it assumes is determined by whether the initial potential is greater or less than V0<\/sub>.\u00a0 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 \u2013 meaning, say, 1 – or O – meaning, O – and if this potential changes a little through electrical leakage, perhaps adjacent portions at a different potential, we can recover or regenerate<\/em> the original potential merely by opening the window of the selectron grid and flooding the area with electrons.\u00a0 In fact, we can periodically regenerate the potentials behind all windows by opening all windows at once and flooding the whole surface with electrons.\u00a0 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.<\/p>\n How do we get the information on the portions of the storage surface beind the various windows? \u00a0That is, how do we initially bring some portions of the surface to the potential Vs and others to the potential V0<\/sub>?\u00a0 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.\u00a0 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.\u00a0 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.\u00a0 Then the pulse applied to the conducting backing falls slowly to zero, as shown.\u00a0 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.\u00a0 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 O, and the pulse will not affect portions of the surface behind closed windows, because no electrons reach them.<\/p>\n Thus, each little area of the storage surface behind each window of the selectron grid is a cell of our memory.\u00a0 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 O, and so can, in effect, write 1 or 0 in that particular memory cell.\u00a0 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<\/sub> or O, despite leakage of electrons to or away from the little areas.\u00a0 We can, that is, put thousands of pieces of information into the selectron and keep them there.\u00a0 What about reading? \u00a0How can we get this information out?<\/p>\n 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.\u00a0 Now, suppose we open one window of the selectron, shooting electrons at a particular area of the surface.\u00a0 If that area has a potential O, the electrons will be repelled from it.\u00a0 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.\u00a0 Suppose we let this light fall on a photo-multiplier, of the type Dr. Rajchman worked on earlier in his career.\u00a0 Then, when we open a given window of the selectron, if the potential of the surface behind the window is O, we get nothing out of the multiplier.\u00a0 But, if the potential is Vs, there is a flash of light, and a pulse of current from the multiplier.\u00a0 And so, we can not only write a O 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.<\/p>\n 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.\u00a0 I have spoken of the phosphor as one giving blue light.\u00a0 This is because the photo-multiplier is more sensitive to blue light than to other colors.\u00a0 And so, I predicted that the memory cells of the thinking machines will be not only multitudinous and small, but also blue.<\/p>\n Of course the selectron provides only a part of the thinking machine – that is, the memory.\u00a0 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.\u00a0 All is a field apart.\u00a0 Still, there is one wrinkle which is so intimately connected with the use of the selectron that it deserves mention here.\u00a0 I have referred to the O 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.\u00a0 Now, usually we don\u2019t want to store isolated digits or letters: we want to store complete numbers or words – combinations of 1 and O, as, 10011.\u00a0 This is 19 in binary notation, and might in some instance stand for the nineteenth word in a dictionary.\u00a0 When we look up a number or a word, we want it all at once, not piecemeal.<\/p>\n 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.\u00a0 Thus, entries in a log table appear as in Figure 8.\u00a0 Suppose that in using the selectron we assign a different tube to each binary digit of the numbers to be stored.\u00a0 If we wish to store twenty-digit numbers, we will need twenty tubes.\u00a0 Each tube will, in effect, be a given column of our storage space.\u00a0 The different cells in a tube will represent different rows.\u00a0 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.<\/p>\n The selectron isn\u2019t perfect by any means.\u00a0 Perhaps it\u2019s not even the final answer.\u00a0 At the moment, in its early form, it may be almost expensive as relays, but that\u2019s partly because it\u2019s new.\u00a0 It\u2019s certainly great deal more compact than relays, a very great deal faster, and probably more reliable as well.\u00a0 It represents a first huge stride in the electronics of the thinking machine.\u00a0 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.<\/p>\n – J.J. Coupling, 1949 –<\/span> References<\/span><\/em><\/span><\/p>\n Dr. Jan A. Rajchman<\/span><\/em><\/p>\n Jan A. Rajchman<\/a> (at Wikipedia)<\/span><\/p>\n Jan. A. Rajchman<\/a> (at I.E.E.E. History)<\/span><\/p>\n J.J. Coupling (Dr. John R. Pierce)<\/span><\/em><\/p>\n J.J. Coupling<\/a> (at Wikipedia)<\/span><\/p>\n J.J. Coupling <\/a>(at Speculative Fiction Database)<\/span><\/p>\n Machine Hearing and the Legacy of John R. Pierce<\/a> (at Cal Tech) (at CalTech.edu)<\/span><\/p>\n Creative Thinking, by John R. Pierce<\/a> (at Tom Schneider’s “Molecular Information Theory and The Theory of Molecular Machines”)<\/span><\/p>\n Selectron Tube<\/span><\/em><\/p>\n Pierce, John R. (as J.J. Coupling), “The Little Blue Cells”, Astounding Science Fiction<\/em>, 1949, Vol. 42, No. 6, February, 1949, pp. 85-99<\/span><\/p>\n Lamps & Tubes \/ Lampen & R\u00f6hren<\/a> (Giorgio Basile’s website)<\/span><\/p>\n Selectron Tube<\/a> (at Wikipedia)<\/span><\/p>\n RCA Selectron<\/a> (at Charles Osborne’s “RCA Selectron.com” – superb and comprehensive website)<\/span><\/p>\n \u041f\u043e\u0447\u0435\u043c\u0443 \u0444\u043e\u043d \u041d\u0435\u0439\u043c\u0430\u043d \u0432\u0435\u0440\u0438\u043b \u0432 SELECTRON<\/a> (“Pochemu fon Neyman Veril v Selectron”) (Why Von Neumann believed in the Selectron) (In Cyrillic)<\/span><\/p>\n Astounding Science Fiction<\/em><\/span><\/p>\n Analog Science Fiction and Fact<\/a> (at Wikipedia)<\/span><\/p>\n","protected":false},"excerpt":{"rendered":" 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.\u00a0 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 … Continue reading The Age of Science: Computer Memory, in Astounding Science Fiction<\/em> – February, 1949<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4,358,7,2],"tags":[353,272,403,401,402,361,400,406,405,404,370,359,378,371,407],"class_list":["post-2626","post","type-post","status-publish","format-standard","hentry","category-art","category-science-fiction","category-sociology","category-technology","tag-astounding-science-fiction","tag-calculating-machines","tag-chesley-bonestell","tag-computer-memory","tag-data-storage","tag-dr-jan-a-rajchman","tag-dr-john-r-pierce","tag-httplampes-et-tubes-infoindex-d-htm","tag-httpwww-isfdb-orgcgi-binindex-cgi","tag-httpwww-rcaselectron-com","tag-hubert-rogers","tag-j-j-coupling","tag-jack-williamson","tag-seetee-shock","tag-selectron-tube"],"_links":{"self":[{"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/posts\/2626","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/comments?post=2626"}],"version-history":[{"count":55,"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/posts\/2626\/revisions"}],"predecessor-version":[{"id":2968,"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/posts\/2626\/revisions\/2968"}],"wp:attachment":[{"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/media?parent=2626"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/categories?post=2626"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/thepastpresented.com\/index.php\/wp-json\/wp\/v2\/tags?post=2626"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
\n<\/span><\/p>\n
\n<\/span><\/p>\n
\n<\/span><\/p>\n
\n<\/span><\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/09\/Astounding-Science-Fiction-1949-06-223x300.jpg\")
<\/a>______________________________<\/span><\/p>\n
\n<\/span><\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-Hubert-Rogers-MGM-753x1024.jpg\")
<\/a>______________________________<\/span><\/p>\n
\n<\/span><\/p>\n
\n<\/span><\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/09\/RCA-Selectron-Tube-Lampes-et-Tubes-242x300.jpg\")
<\/a><\/p>\n
\n<\/span><\/p>\n
\n<\/span><\/p>\n
\n<\/strong>By J.J. COUPLING<\/span><\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-87-Figure-1-300x238.jpg\")
<\/a>![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-88-Figure-2-175x300.jpg\")
<\/a>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.\u00a0 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.\u00a0 What happens? \u00a0Let us consider first the bars of the selectron grid.\u00a0 Figure 3 tells the story.\u00a0 If two neighboring bars are negative, the approaching electrons are simply repelled and turned back.\u00a0 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.\u00a0 Only if the bars on both sides<\/em> of the space which the electron enters are positive does the electron get through.\u00a0 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.\u00a0 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.\u00a0 If both <\/em>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.\u00a0 Thus, we have a means for letting electrons through one window at a time.<\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-89-Figure-3-300x297.jpg\")
<\/a><\/strong>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.\u00a0 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? \u00a0The tube would certainly work if it had one hundred twenty-eight leads to the selectron grid, but Dr. Rajchman\u2019s ingenuity has cut this down instead to thirty-two, a saving by a factor of four.\u00a0 How is this done? \u00a0The table of Figure 4 tells the story.\u00a0 Here we have in the top row the numbers of the bars, in order, sixty-four in all.\u00a0 These bars are connected to two sets of eight leads.\u00a0 The second and third rows show to which lead of a given set a bar is connected.\u00a0 Thus, Bar 1 is connected to Lead 1 of Set I.\u00a0 Bar 2 is connected to Lead 1 of Set II, while Bar 64 is connected to Lead 8 of Set II.\u00a0 To save space, some of the bars have been omitted from the table.<\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-91-Figure-4-300x109.jpg\")
<\/a><\/strong>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<\/em><\/strong>, 29, 45 and 61 will be positive.\u00a0 Then, if we make Lead 2 of Set II positive and all the other leads of Set II negative Bars 4, 8, 12<\/strong><\/em> and 16 will be positive.\u00a0 All the bars which do not appear in either of the above listings will be negative.\u00a0 Now, the only adjacent bars listed are 12<\/em><\/strong> and 13<\/em><\/strong>, which have been written in italics.\u00a0 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.\u00a0 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.<\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-92-Figure-5-300x288.jpg\")
<\/a><\/strong>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.\u00a0 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.\u00a0 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.\u00a0 The curve shows how current to the surface varies as the potential of the surface is varied.<\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-94-Figure-6-300x271.jpg\")
<\/a><\/strong>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.\u00a0 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.\u00a0 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.\u00a0 These are whisked away to the more positive selectron grid. \u00a0These negative electrons leaving the surface are equivalent to a positive current to the surface.\u00a0 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.\u00a0 Finally, at some potential labeled V0<\/sub> 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.\u00a0 If the potential of the storage surface is higher than V0<\/sub>, 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.\u00a0 This tends to make the storage surface more positive.<\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-97-Figure-7-300x267.jpg\")
<\/a><\/strong>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.\u00a0 By flooding a given area or cell with electrons and applying a sharply falling, negative pulse, which rises again gradually toward O – the dashed pulse of Figure 7 upside down – we can bring any selected area of the storage surface to O potential, and thus write O in any selected cell of the memory.<\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-98-Figure-8-167x300.jpg\")
<\/a><\/strong>We want to look up all the digits in a given row at once.\u00a0 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.\u00a0 \u00a0Thus, 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<\/em>.\u00a0 We are then ready to read the number in Row 1 or to write a new number in.\u00a0 \u00a0The 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.\u00a0 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. \u00a0We see, thus, how a whole table of numbers can be stored in twenty selectrons.\u00a0 The windows 1, 2, 3 et cetera, can represent, for instance, the angle of which we want the sine.\u00a0 The first selectron can store the first digits of all the sines, the second selectron can store all the second digits, et cetera.\u00a0 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.<\/p>\n
\n<\/em><\/span><\/p>\n![\"\"](\"http:\/\/thepastpresented.com\/wp-content\/uploads\/2017\/08\/Astounding-Science-Fiction-1949-02-p-163-Figure-9.gif\")
<\/a>______________________________<\/span><\/p>\n
Notice: ob_end_flush(): failed to send buffer of zlib output compression (0) in /home4/emc2ftl3/public_html/wp-includes/functions.php on line 5493