Showing papers in "Ibm Journal of Research and Development in 1963"
TL;DR: A new mathematical model is proposed to describe the distribution of the occurrence of errors in data transmission on telephone lines that can be well approximated by a law of Pareto of exponent less than one and that the relative number of errors and the equivocation tend to zero as the length of the message is increased.
Abstract: This paper proposes a new mathematical model to describe the distribution of the occurrence of errors in data transmission on telephone lines. We suggest: a) that the statistics of telephone errors can be described in terms of an error probability depending solely on the time elapsed since the last occurrence of an error; b) that the distribution of inter-error intervals can be well approximated by a law of Pareto of exponent less than one; the relative number of errors and the equivocation tend, therefore, to zero as the length of the message is increased. The validity of those concepts is demonstrated with the aid of experimental data obtained from the German telephone network. Further consequences, refinements, and uses of the model are described in the body of the paper.
144 citations
TL;DR: A character recognition system has been developed for the recognition of handwritten numerals using a logically controlled cathode ray tube scanner to generate basic measurements that characterize significant features of the numeral shapes.
Abstract: A character recognition system has been developed for the recognition of handwritten numerals. This system uses a logically controlled cathode ray tube scanner to generate basic measurements that characterize significant features of the numeral shapes. A contour-follower procedure is used to control the scanner. In addition, special scanner subroutines initiated by feedback from the recognition logic are utilized. Character shape data are generated in a sequential form, which can be analyzed for recognition with an easily realizable logic.
An experimental model has been built that recognized 99.3% of numerals written by 45 subjects after 30 minutes of training. The error rate ftohre se characters was 0.11%. The rejected character ratew as 0.59%.
91 citations
TL;DR: A computer program has been written to design character recognition logic based on the processing of data samples to search for logic circuits having certain constraints on hardware design and evaluate these logics in terms of their discriminating ability over samples of the character set they are expected to recognize.
Abstract: A computer program has been written to design character recognition logic based on the processing of data samples. This program consists of two subroutines: (1) to search for logic circuits having certain constraints on hardware design, and (2) to evaluate these logics in terms of their discriminating ability over samples of the character set they are expected to recognize. An executive routine is used to apply these subroutines to select a complete logic with a given performance and complexity. This logic consists of 39 to 96 AND gates connected to a shift register and a table look-up or resistance network comparison system.
The methods were applied to the design of recognitionl ogics for the 52 upper and lower case characters of IBM Electric Modern Pica type font and lower case Cyrillic characters scanned from Russian text. In both cases when the logics were tested on data different from that used to design the logics, the substitution rate was about one error per thousand. A single logic was designed to read two different Cyrillic fonts. For this design, an error rate of one error per hundred characters was observed.
Several experiments are reported ona number of logics designed for typewritten data, and single- and two-font Cyrillic data. The performances of different recognitionsy stems are compared as a function of the complexity of the recognition logics.
90 citations
TL;DR: A new type of diffraction loss for a thin light-emitting layer surrounded by light-absorbing material is calculated.
Abstract: Mathematical expressions are derived for the minimum current density necessary to cause stimulated emission in injection lasers. A new type of diffraction loss for a thin light-emitting layer surrounded by light-absorbing material is calculated.
67 citations
30 citations
TL;DR: An improved method for fabricating multiple-image photographic masks of high quality is described for the fabrication of silicon transistors and diodes in photoengraving-type fabrication process.
Abstract: This Communication describes an improved method for fabricating multiple-image photographic masks of high quality. Photographic masks are commonly used in the fabrication of silicon transistors and diodes. Such masks are needed in sets of five, ten, or even twenty, each being used for one of the steps in the photoengraving-type fabrication process. Arrays of many devices are made at one time on a silicon wafer. The photographic masks are the corresponding arrays of microscopic images. Figure 1 shows a portion of two representative transistor masks. The center-to-center distance between the devices may typically be 0.020″, 0.025″ or 0.030″, with a total of perhaps 1000 devices in the array.
26 citations
25 citations
TL;DR: This novel approach to the file addressing problem is illustrated with a specific design to show feasibility and its effectiveness is illustrated by comparing test results obtained from a simulated calculation, which used typical data, against values calculated from an ideal model.
Abstract: Techniques used in the theory of error-correcting codes are applied to solve the problem of addressing a large file. This novel approach to the file addressing problem is illustrated with a specific design to show feasibility. Its effectiveness is further illustrated by comparing test results obtained from a simulated calculation, which used typical data, against values calculated from an ideal model.
24 citations
21 citations
TL;DR: Continuous stimulated emission was achieved with the diode immersed in liquid helium II with a direct current flowing through the unit, operation was maintained as long as the liquid helium lasted (more than one-half hour) for more than ten runs.
Abstract: Figure 1 Recorder traces of the light output spectrum of a cw laser at 1.95° K. The light is observed out the long side of the unit in the plane of the junction. There is 11.25 A between markers with an extra marker every ten. The zero for the 38 rna is at the bottom of the chart and each successive curve is displaced upward by 2 divisions from the one below except for the 85 rna curve, which is displaced from the 60 rna curve by 5 divisions. The broad background is the spontaneous emission. The sharp spike is the stimulated emission. The 7.5 A spectral slit width determines the width of the stimulated emission. The area of the junction is 5AX 10-4 em>. 74 Stimulated emission of radiation from forward-biased GaAs diodes was recently observed in several laboratories.tz The observations reported were made with short pulses « 20 JLsec) and a low duty cycle at 77°K. In this Letter continuous stimulated emission is reported. The structure used is a rectangular parallelepiped with dimensions 100 JL X 120 JL X 450 JL similar to the ones in which directionality effects have been observed.t-s The p-n junction is parallel to the long dimensions of the crystal and is about in the middle of the unit. Continuous stimulated emission was achieved with the diode immersed in liquid helium II. The bath temperature varied between 1.95° and 2.0 throughout the runs, higher input powers producing small but observable temperature rises (=O.03°K). With a direct current flowing through the unit, operation was maintained as long as the liquid helium lasted (more than one-half hour) for more than ten runs. The stimulated emission was not observed for bath temperatures above 2.10 Figure 1 is a photograph of the recorder traces of the spectrum observed from the broad side of the unit at several currents. The broad background, which has a full width at half maximum intensity of 75 A, is the spontaneous emission. At 38 ma, which is below the threshold, only this spontaneous emission is observed. At 40 ma a prominent narrow stimulated emission line appears at 8383 A. (For Fig. I the slits were opened to display both the background and the narrow emission on the same scale.) With increasing current, the stimulated emission at first increases, more rapidly than the background, up to about 70 ma, at which point it starts to decrease. By 85 ma it has disappeared and only the spontaneous emission remains. This quenching is due to heating of the diode as evidenced by the shift of the 85 ma curve to longer wavelengths. Comparisons of the respective areas under the traces for the narrow peak and the background indicate that Vl .... z => >'\" -c '\" .... '\" '\" « z
20 citations
TL;DR: Results in coding theory can be used in solving this file addressing problem and it is shown, in particular, how the theory of Bose-Chaudhuri codes can be applied.
Abstract: In this paper a file address problem is proved to be equivalent to a problem in coding theory. Results in coding theory can thus be used in solving this file addressing problem. It is shown, in particular, how the theory of Bose-Chaudhuri codes can be applied. A simple transformation from the input information to its address is given. This method can be easily implemented using either a computer or shift registers.
TL;DR: It has been proven that the principle of superposition applies to a magnetic read-back waveform, and each pulse can be treated as an isolated transient, and a linear passive network can be used to reduce its width.
Abstract: It has been proven that the principle of superposition applies to a magnetic read-back waveform. Consequently, each pulse can be treated as an isolated transient, and a linear passive network can be used to reduce its width. In many cases this effective increase in read-back resolution would permit an increase in the operating density. The isolated read-back pulse is first approximated by a Gaussian curve. A second approximation is effected in the frequency domain, and a table of transfer functions is obtained. A network is designed using one of the transfer functions, and the solution is given an algebraic form. One particular case is illustrated numerically and the laboratory results are shown.
TL;DR: The hydrodynamic air lubrication of magnetic tape over magnetic heads, to prevent contact and wear, closely approximates the foil bearing and there is a great deal of interest in this configuration.
Abstract: Classic lubrication theory has generally been directed toward fluid films bounded by solid objects that were assumed to be infinitely rigid. This assumption, however, is contrary to the physical nature of things. Blok and van Rossum
1
introduced the concept that the rigid bearing could be replaced with a perfectly flexible band. This configuration has been termed the foil bearing. It represents the other extreme situation. Between these limits then are the cases of practical engineering interest. The hydrodynamic air lubrication of magnetic tape over magnetic heads, to prevent contact and wear, closely approximates the foil bearing. From the viewpoint of engineering design there is a great deal of interest in this configuration. The parameters of the foil bearing discussed in this Communication are shown in Fig. 1.
TL;DR: The large-signal transistor model evolved for the purpose of carrying out such an analysis; methods of measuring device parameters with relevant theory; computational techniques most adaptable; and correlation between predicted and observed transients are described.
Abstract: Characterization of devices for high speed ACP (Advanced Circuit Program) circuits demands an accurate study of transients and switching delays. This paper describes (a) the large-signal transistor model evolved for the purpose of carrying out such an analysis; (b) methods of measuring device parameters with relevant theory; (c) computational techniques most adaptable; and (d) correlation between predicted and observed transients.
Many new ideas in the development of the device model, measurements, and computational procedure are reported and could be used for any general circuit analysis.
TL;DR: Design techniques which result in a compatible set of logic circuits and power amplifiers are described and Germanium tunnel diodes are used in power driving and transmission line driving circuits.
Abstract: Directly coupled logic circuits utilizing silicon transistors have been developed in the Advanced Circuit Program. Silicon transistors possess sufficient field gradient in the collector-base "diode" to permit high-speed operation under a forward-biased condition. Current-switching techniques provide the well-controlled voltage levels which permit operation in this region without sacrificing high speed. Operation in this region permits circuit simplification and low power dissipation.
A desirable state-of-the-art device is defined. Design techniques which result in a compatible set of logic circuits and power amplifiers are described. Germanium tunnel diodes are used in power driving and transmission line driving circuits.
In order to measure the delay times, a model with 424 circuits was subsequently built. ACP circuits (the name adopted for this circuit family) were found to perform two levels of logic in 3.2 nsec.
TL;DR: Experimental results do not support a previous theory that the reversal of birefringence is reversed because of various particle sizes, but it is believed that in dilute solutions this behavior is caused by some inherent property of the bentonitem icelle.
Abstract: Colloidal suspensions of bentonite were observed to undergo a negative electro-optical birefringence at low-voltage fields even in dilute suspensions, a phenomenon not reported earlier. Experimental results do not support a previous theory that thes ign of birefringence is reversed because of various particle sizes. Although interaction betweenp articles was found to augment this low-voltage anomalous behavior, we believe that in dilute solutions this behavior is caused by some inherent property of the bentonitem icelle. This contention is supported by variation in the relaxation times in negative and positive birefringence regions, thus suggesting orientation on two different axes in the two respective regions.
TL;DR: A general theory of cyclic codes correcting a set of given types of errors is presented, which offers several new classes of codes that are competitive with existing ones and may be better for certain message lengths.
Abstract: A general theory of cyclic codes correcting a set of given types of errors is presented. Codes published by Abramson, Fire, Melas and others are accounted for in this theory, which also offers several new classes of codes. These new codes are competitive with existing ones. In burst-error correction, for certain message lengths they may be better since they may need fewer parity check bits.
TL;DR: A method is developed for determining, to the first order, the deflection from a straight path of a perfectly flexible tape moving near a rigid cylinder.
Abstract: A method is developed for determining, to the first order, the deflection from a straight path of a perfectly flexible tape moving near a rigid cylinder. The case of a parabolic cylinder is considered.
TL;DR: An analysis is given of the magnetization of a cylindrical thin film exhibiting a uniaxial anisotropy in the circumferential direction and the technique of superposition may be applied to other film geometries.
Abstract: An analysis is given of the magnetization of a cylindrical thin film exhibiting a uniaxial anisotropy in the circumferential direction. The magnetization and demagnetizing fields are derived for the cylinder where the magnetization is not uniform. The derivation is accomplished by dividing up the cylinder into a large number of uniformly magnetized, coaxial cylindrical regions in superposition and by integrating their individual contributions. Some applications of the derivation are shown for several specified field geometries. The technique of superposition may be applied to other film geometries.
TL;DR: High-resolution measurements of the line width as well as the dependence of the GaAs stimulated emission characteristics on hydrostatic pressure are reported.
Abstract: Stimulated emission from forward biased p-n GaAs junctions has been reported by several laboratories. 1,2 Most of the reported spectroscopic measurements have been made under medium resolution. This Communication reports high-resolution measurements of the line width as well as the dependence of the GaAs stimulated emission characteristics on hydrostatic pressure.
TL;DR: The saturation front moving along a string of films is shown to be the shock front and the shock speed, which is determined in terms of the parameters of the circuit, is then the operational speed limit of a thin film memory.
Abstract: This paper considers the propagation of waves in a transmission line loaded with thin permalloy films. Since the films are driven to saturation, the transmission line equations which are derived to describe the wave propagation are nonlinear. The nonlinearity requires the use of shock wave analyses, and a derivation of the appropriate shock relations is included. The problems of a line loaded with a single film and a line loaded with a periodic array of films are both treated. The saturation front moving along a string of films is shown to be the shock front. The shock speed, which is determined in terms of the parameters of the circuit, is then the operational speed limit of a thin film memory.
TL;DR: This paper describes the factors leading to the present design approach, and examples of engineering applications are given, including a cross-sectional model of a “scratch pad” memory for use with Advanced Circuit Program logic circuitry.
Abstract: Beginning with a brief history of tunnel diode memories, this paper describes the factors leading to the present design approach. Array design criteria are discussed at length. Examples of engineering applications are given, including a cross-sectional model of a “scratch pad” memory for use with Advanced Circuit Program logic circuitry.
TL;DR: A computer-oriented logic for obtaining the determination of the linear sequence of amino acids in proteins by applying successively stronger decision rules is described.
Abstract: A fundamental problem for biochemistry is the determination of the linear sequence of amino acids in proteins. This paper describes a computer-oriented logic for obtaining such determination. The logic applies successively stronger decision rules to extract the required information on the protein sequence.
TL;DR: The photoelectric emission from a nickel ribbon has been observed as a function of temperature from 25° to 760°C, over a wavelength range from 2250 to 2530 A, qualitatively confirming and extending Cardwell's earlier work.
Abstract: The photoelectric emission from a nickel ribbon has been observed as a function of temperature from 25° to 760°C, over a wavelength range from 2250 to 2530 A, qualitatively confirming and extending Cardwell's earlier work. The yield increases with temperature at all wavelengths, with an upward bulge near the Curie point. Fowler-Dubridge analyses of the emission from the front face of the ribbon, which is found to contain mainly (111) facets after extensive outgassing, yield values of the work function ranging from about 5.07 ev at 25°C to about 5.20 ev at 760°C. Behavior below the Curie point may be consistent with the magnetization-squared dependence recently suggested by Wonssowski, et al.
TL;DR: Examination of the diffraction pattern showed Au to be the only metal of the three to form an intermetallic compound with lead, indicating that compound formation aids in nucleation.
Abstract: Lead films evaporated onto thin nucleating layers of Cu, Ag, and Au were studied by the techniques of electron diffraction and electron microscopy. Electron micrographs indicated that films nucleated with Au become continuous much sooner than films nucleated with Ag or Cu. Examination of the diffraction pattern showed Au to be the only metal of the three to form an intermetallic compound with lead, indicating that compound formation aids in nucleation.