Byte, octubre del 85

Portada de la revista Byte de octubre de 1985. El tema de portada es Simulating Socienty. Lo ilustra una hoja de papel de impresora que envuelve unas caras humanas.

Vamos allá con nuestra relectura de lo último en informática…de hace cuarenta años, a través de los archivos de la revista Byte en archive.org. Hoy toca octubre de 1985.

Para comenzar, no os quejéis de que no estáis presenciando los grandes avances de la historia. Os presento… ¡el disquete de alta densidad! (Creo que la mayoría de los que me leéis ya sois talluditos y apreciaréis que saltar de 720 kilobytes a 1.44 megas, sin ser revolucionario, sí fue todo un salto.)

Sony, Toshiba Prepare High-Density 3 ½ inch Disks Sony announced in Tokyo that it has developed a 2-megabyte 3½ inch floppy disk, storing 1.6 megabytes (formatted) by doubling the number of sectors per track. The 2-megabyte medium uses a 1 micron magnetic layer (half the thickness of current 1 -megabyte disks) and requires a higher coercivity (700 rather than 600-620 oersteds). While the 2-megabyte versions use the same magnetic technology as earlier 3 ½-inch disks and drives, the magnetic heads of the drives require higher tolerances. An additional disk cartridge hole allows drives to distinguish between 1- and 2-megabyte disks. Although it has already licensed 38 companies to produce 2-megabyte disks, Sony says it is waiting for formal standards to be set before marketing the disks and drives, which should be available to OEMs next year, probably at prices about 20 percent higher than 1-megabyte versions. An even denser 3 ½-inch drive from Toshiba uses perpendicular recording technology to squeeze 4 megabytes of data onto a single-sided disk coated with barium ferrite. Toshiba plans to release evaluation units early next year, with full production slated for 1987 While the 2-megabyte versions use the same magnetic technology as earlier 3 '/2-inch disks and drives, the magnetic heads of the drives require higher tolerances. An additional disk cartridge hole allows drives to distinguish between 1- and 2-megabyte disks. Although it has already licensed 38 companies to produce 2-megabyte disks, Sony says it is waiting for formal standards to be set before marketing the disks and drives, which should be available to OEMs next year, probably at prices about 20 percent higher than I -megabyte versions. An even denser 3 '/2-inch drive from Toshiba uses perpendicular recording technology to squeeze 4 megabytes of data onto a single-sided disk coated with barium ferrite. Toshiba plans to release evaluation units early next year, with full production slated for 1987.

Que levante la mano quien supiese / recordase que antes de Access, la base de datos de Microsoft (que no llegaría hasta 1992), hubo un Microsoft Access para conectarse a servicios de información a través del módem (yo no tenía ni idea / no lo recordaba en absoluto). La hegemonía del Access base de datos es tal que apenas he sido capaz de encontrar más información al respecto.

Anuncio de Microsoft Access. Lo ilustra un ordenador sobre el que hay el auricular de un teléfono de sobremesa, roto por la mitad. El titular es Don't get mad, get Access

En nuestra habitual sección «crees que esto se acaba de inventar, pero no» tenemos a la sección de libros, que se hace eco de Computer culture : the scientific, intellectual, and social impact of the computer, disponible, como no, en archive.org, que recogía las ponencias de la conferencia del mismo nombre, porque no es solo en Despacho 42 que nos preocupamos de estos temas y que, naturalmente, ya se preocupaba del impacto de la IA…

Artificial Intelligence Approximately one-fourth of Computer Culture (four papers and one panel discussion) deals specifically with artificial intelligence. The panel discussion on the impact of Al research is the most thought-provoking contribution in the book. As you might expect, this discussion is not so concise as an article dealing with the same topic, but the interaction among the panel members is intriguing. The panel consists of two philosophers (Hubert Dreyfus and John Searle) and three computer scientists (John McCarthy, Marvin Minsky, and Seymour Papert). Much of the discussion is spent identifying important questions about Al. Each panelist has a distinct viewpoint, resulting in a diversity of questions. Among these, however, two issues are of overriding concern: Can machines think? If they can, is machine thinking the same as human thinking? The panelists seem to agree that computers can be used to study thinking, if for no other reason than to provide a contrast with human thought processes. On the other hand, the suggestion that appropriately programmed computers could duplicate human thought processes is much more controversial. Aside from the philosophical issues, Papert makes a very important point when he argues that it is dangerous to reassure people that machines will never be able to challenge the intellectual capabilities of human beings. If people are lulled into a sense of security about machine capabilities, they will be ill prepared to deal with situations in which machines become better than people at doing specific jobs, he says. Whether or not the machines are described as thinking in these situations, the social and psychological issues raised by machine capabilities demand attention.
(Enlazo a la página de portada de la sección de libros, en vez de la específica del fragmento que tenéis aquí. En cualquier caso, vale la pena leer la crítica completa… e incluso el libro, si tenéis la oporunidad)

Más cosas que no se inventaron ayer. Uno ve poco fútbol del de darle patadas a un balón, pero bastante fútbol americano, un deporte en que las retransmisiones no serían lo mismo sin la obligatoria skycam, ua cámara que sobrevuela el terreno de juego colagada de cuatro cables. Y sí, cumple cuarenta años:

Skycam: An Aerial Robotic Camera System A microcomputer provides the control to add three-dimensional mobility to TV and motion picture cameras On a morning in March 1983, a group of technicians gathered at Haverford High School in a suburb of Philadelphia. Each brought an electrical, mechanical, or software component for a revolutionary new camera system named Skycam (see photo 1). Skycam is a suspended, mobile, remote-controlled system designed to bring three-dimensional mobility to motion picture and television camera operation. (See the text box on page 128.) I used an Osborne 1 to develop Skycam's control program in my basement, and it took me eight months of evenings and weekends. As of 3 a.m. that morning, however, the main control loop refused to run. But 19 hours later, Skycam lurched around the field for about 15 minutes before quitting for good. Sitting up in the darkness of the press booth, hunched over the tiny 5-inch screen, 1 could see that the Osborne 1 was not fast enough to fly the Skycam smoothly. In San Diego 18 months later, another group of technicians opened 20 matched shipping cases and began to get the Skycam ready for an NFL preseason game between the San Diego Chargers and the San Francisco FortyNiners. The Skycam was now being run by an MC68000 microprocessor based Sage computer, and a host of other improvements had been made on the original. [Editor's note: The Sage Computer is now known as the Stride: however, the machine used by the author was purchased before the company's name change. For the purpose of the article, the machine will be referred to as the Sage.] For the next three hours, Skycam moved high over the field fascinating the fans in the stadium while giving the nationwide prime-time TV audience their first look at a new dimension in sports coverage. Skycam represents an innovative use of microcomputers. The portable processing power needed to make Skycam fly was unavailable even five years ago. That power is the "invention" upon which the Skycam patents are based. It involves the support and free movement of an object in a large volume of space. The development team used the following experiment to test the movement and operation of the Skycam. At a football field with one lighting tower at each of four corners, the team members bolted a pulley to the top of each pole, facing inward. Then they used four motorized winches, each with 500 feet of thin steel cable on a revolving drum and put one at the base of each tower. Next, they ran a cable from each motor to the top of its tower and threaded the cable through the pulley. They pulled all four cables from the tops of the towers out to the middle of the field and attached the cables to a metal ring 2 feet in diameter weighing 10 pounds (see figure 1). A motor operator was stationed at each winch with a control box that enabled the operator to slowly reel in or let out the cable. Each motor operator reeled the cable until the ring was suspended a few feet from the ground, and then they were ready to demonstrate Skycam dynamics. All four motor operators reeled in the cable. The ring moved upward quickly. If all four motors reel in at the same rate (and the layout of lighting towers is reasonably symmetrical) the ring will move straight up. In the experiment, the two motors on the left reeled in and the two on the right reeled out. The ring moved to the left and maintained its altitude. An instruction was given to the two motor operators on the left to reel out and the two on the right to reel in just a little bit. The ring moved right and descended as it moved back toward the center. The theoretical basis of this demonstration is quite simple. For each point in the volume of space bounded by the field, the four towers and the plane of the pulleys, there is a unique set of four numbers that represents the distances between that point and each of the four pulley positions. Following the layout above for an arbitrary point on the field, you can...

Pero este mes me quedo con el tema de portada: el uso de simulaciones informáticas para modelar la sociedad:

Simulating Society THE NEED FOR GREATER RIGOR in the social sciences has long been acknowledged. This month's theme examines computer-based simulation as a means to achieving that end. Simulation may be able to assist in evaluating hypotheses, not in the sense that an experiment in the physical sciences can test a hypothesis, but in the sense of making plain the ramifications of a hypothesis. The value of specifying a hypothesis with sufficient clarity to be amenable to programming and of examining the consequences of that hypothesis should not be underestimated. Indeed, one of the interesting aspects of the work presented here is that these researchers appear to be developing a tool for the social sciences that is not simply a poor stepchild of physical science methodologies. Our first article, "Why Models Go Wrong" by Tom Houston, is a wonderfully readable account of the ways that you can misuse statistics. Next, Wallace Larimore and Raman Mehra's "The Problem of Overfitting Data" discusses a difficult but important topic. Overfitting happens when your curve traces the noise as well as the information in your data. The result is that the predictive value of the curve actually deteriorates. In "Testing Large-Scale Simulations," Otis Bryan and Michael Natrella show how validation (determining whether the specification for the simulation corresponds with reality) and verification (determining whether the simulation program corresponds with the specification) were achieved on a large-scale combat simulation they developed for the Air Force. The ways of economic modeling are illustrated by Ross Miller and Alexander Kelso, who show how they analyzed the effects of proposed taxes for funding the EPA Superfund in "Analyzing Government Policies." Michael Ward discusses his ongoing research in simulating the U.S.-Soviet arms race in "Simulating the Arms Race." Several authors discuss new and surprising applications of simulation. In "EPIAID," Dr. Andrew Dean describes the development of computer-based aids for Centers for Disease Control field epidemiologists. Royer Cook explains how he fine-tuned a model in "Predicting Arson," and Bruce Dillenbeck, who uses an arson-prediction program in his work as a community activist, discusses modeling in "Fighting Fire with Technology" Articles in other sections of the magazine that relate to this theme include Zaven Karian's review of GPSS/PC and Arthur Hansen's Programming Insight "Simulating the Normal Distribution." When I began researching this theme, I took an excellent intensive course in simulation from Edward Russell of CACI. Dr. Russell's is the unseen hand guiding the development of this theme. Of course, any blame for bias in the choice of theme topics belongs to me, but much of the credit for the quality that is here must reside with him.

No os perdáis los artículos sobre los problemas, comenzando por los dos que abren la sección, sobre los riesgos del mal modelado (un tema que, desafortunadamente, tiene hoy todavía más importancia que hace cuarenta años), y siguiendo con el de modelado económico con Lotus 1-2-3, o el de epidemiología.

Ah, y aprovechando que la cosa iba de modelado… ¿sabíais que SPSS/PC+, no solo ya existía en 1985, sino que fue lanzado en 1968? Si a alguien se le ocurre un software que lleve más tiempo en el mercado, que avise.

Anuncio del programa SPSS/PC+. El eslogan es Make Stat Magic. Lo ilustra la foto de un sombrero de copa, como los de los magos, del que sale un disquete de 5¼ etiquetado SPSS/PC+

Y no vamos a dejar de hablar del Amiga, claro. Esta vez, es Bruce Webster, otro de los columnistas estrella de la revista, el que nos explica lo mucho que ha alucinado con la potencia, el precio y la elegancia del sistema:

According to Webster Commodore's Coup Product of the Month: Amiga Last month, I made a few comments about the future of the home computer market, based on rumors I had heard about the Amiga from Commodore. In essence, I said that if what I had heard was true the Amiga might be the heir to the Apple II in the home/educational/small business marketplace. Since writing that. 1 have seen the Amiga. I have watched demonstrations of its abilities; I have played with it myself; and I have gone through the technical manuals. My reaction: I want to lock myself in a room with one (or maybe two) and spend the next year or so discovering just what this machine is capable of. To put it another way: I was astonished. Hearing a description of a machine is one thing, seeing it in action is something else especially where the Amiga is concerned I can tell you that the low-resolution mode is 320 by 200 pixels, with 32 colors available for each pixel (out of a selection of 4096). But that does not prepare you for just how stunning the colors are especially when they are properly designed and combined. It also doesn't tell you that you can redefine that set of 32 colors as the raster-scanning beam moves down the screen, letting you easily have several hundred colors on the screen simultaneously. It also doesn't tell you how blindingly fast the graphics hardware is. If you've seen some of Commodore's television commercials demonstrating the Amiga's capabilities, or if you've looked at the machine yourself, you have some idea as to what the machine can do. If you haven't, I'm not sure I can adequately describe it. Having seen the graphics on the Amiga, I have to smile when I hear people lump it together with the Atari 520ST. The high resolution mode on the ST is 640 by 400 pixels with 2 colors (out of 512); on the Amiga, it is 640 by 400 pixels with 16 colors (out of the 4096). and you can redefine those 16 colors as the raster-scanning beam goes down the screen. Also, the graphics hardware supporting all those colors is much faster. Little wonder, then, that a friend of mine, a game developer with several programs on the market, came back from the Amiga developers' seminar with plans to return the Atari ST development system at his house and to turn his attentions to the Amiga instead. As I guessed last month, the real strength of the Amiga is its totally open architecture. An 86-pin bus comes out of one side of the machine, giving any add-on hardware complete control of the machine What's more 512 K bytes of the 68000's 16-megabyte address space have been set aside for expansion hardware, 4K bytes each for 128 devices. A carefully designed protocol tells hardware manufacturers what data they should store in ROM (read-only memory) so that the Amiga can automatically configure itself when booted. This is a far cry from the closed-box mentality of the Macintosh, which has forced many hardware vendors through weird contortions just to get their devices to talk consistently to the Mac without crashing. The memory map is well thought out. The Amiga comes with 256K bytes of RAM (random-access read/write memory); an up...

Snif.

Si os lo leéis entero, por favor no os asustéis cuando lleguéis al momento en que comenta que la RAM está a 350 dólares (algo más de mil, actualizando la inflación) por 256 kilobytes. Vamos, que lo por lo que costaban 256 kilobytes hoy te puedes comprar unos 320.. gigabytes. Un millón a uno. (Y supongo que no os sorprenderá mucho comprobar que los márgenes de beneficio de Apple al vender RAM para sus sistemas no son una cosa del siglo XXI.)

Y lo dejamos aquí por este mes. Nos vemos el mes que viene, con el número de noviembre.

Deja un comentario