Byte, septiembre del 85. Diez años de Byte

Portada de Byte de septiembre de 1985. Un tema de portada es el 'homebrewing', el otro, el supersistema de Ciarcia, con un ordenador compatible con Z80 a 6 megahercios y con 256 kilobytes de RAM

Pues vamos allá con el número de septiembre del 85 de Byte, el del décimo aniversario de la revista… En portada, un ordenador, pero uno construido por uno de los autores estrella de la revista, Steve Ciarcia, que se sacaba de la manga un ordenador de 8 bits para la era de los 16, con todo lujo de esquemas para que te lo montaras tú mismo:

BUILD THE SB180 SINGLE-BOARD COMPUTER PART 1: THE HARDWARE by Steve Ciarcia This computer reasserts 8-bit computing in a 16-bit world Newer, faster, better. These words are screamed at you in ads and reviews of virtually every new computer that comes to market. Unfortunately, many of the proponents of this rhetoric are going on hearsay evidence. While advertising hype has its place in our culture, a more thorough investigation may lead you to alternative conclusions. Generally speaking, quotes of increased performance are basically comparisons of CPU (central processing unit) instruction times rarely involving the operating system. The 68000 is indeed a more capable processor than the 6502, but that doesn't necessarily mean that commercial application programs always run faster because the CPU has more capability. People owning 128K-byte Macintoshes have discovered this. The bus size of the processor is only one factor in the performance of a computer system. Operating-system design and programming styles contribute much more to the overall throughput of a computer. It is not enough to simply compare 8 to 16 bits or 16 to 32 bits. For example, the Sieve of Eratosthenes prime-number benchmark runs faster in BASIC on the 8-bit 8052-based controller board presented in last month's Circuit Cellar than it does on a 16-bit IBM PC.

Y por si esto no fuera suficiente para la sección «cosas que no veríamos en una revista generalista de informática hoy»…

AN ANALYSIS OF SORTS by Jonathan Amsterdam How to choose one sorting algorithm over another A friend told me recently that 90 percent of all the computer programs in the world sort. I can believe it. Our society's passion for organization has elevated the simple task of putting things in order to a position of major importance. And who better to carry out the job than those informational beasts of burden— computers? Because of their significance, sorting algorithms have been thoroughly studied. Some are slow and some are fast. Some sort a few items and some sort millions of items. Here I want to discuss sorting in the context of three different algorithms: Selection Sort, for small lists. Quicksort, for larger lists, and Mergesort, for lists of a size so monstrous they can't fit into memory all at once. But first we will need to develop some simple tools to help us with our analysis of these algorithms. Analysis Our goal is to understand the efficiency of some sorting algorithms. But we are immediately faced with a problem: How can we study an algorithm in the abstract without considering the language it's written in or the machine it's running on? For example, any algorithm written in a high-level language will run faster when written in assembly language. And any program running on a microcomputer would run faster on a mainframe. We want to abstract away from these facts, to talk about an algorithm's running time independent of machine or language.

Efectivamente: una discusión lo suficientemente sesuda como para una asignatura de Algoritmos de primero de carrera sobre los diferentes algoritmos de ordenación (temazo de lectura siempre muy recomendable), con sus grafiquitas sobre complejidades de tipo lineal, cuadrática y «n log n», algoritmos no tan básicos…

Figure 1: the rates of grouth of n, n log n and n squared Listing 1: The algorithm for Selection Sort. Selection Sort. Input: an array, A, and its size, n. Output: the same array A, in sorted order, begin for i : = 1 to n do begin m : = i; for j : = i + 1 to n do compare A[j] to A[m], making j the new m if it is less; swap A[i] and A[mj; end end.

…el merge sort (o ordenamiento por mezcla),

Figure 2. The mergesort treeListing 2. The algorithm for Mergesort. Mergesort. Input: a list, L. Output: a sorted list, S. begin If L is one item long, then S = L Otherwise, split L into two lists L1 and L2, each about half as big. Mergesort L1 into S1. Mergesort L2 into S2. merge S1 and S2 into S. end.

…o el mismísimo Quicksort:

Listing 3: The algorithm for Quicksort. Quicksort. Input: an array A, with items from 1 to n. Output; the same array, sorted, begin choose a pivot; partition the list so that all items < = pivot are < i; Quicksort A from 1 to i - 1 ; Quicksort A from i to n; end.

Pero la gracia de este número era la celebración del primer decenio de vida de la revista, y el consiguiente echar la vista atrás, entrevistas incluidas con el ya citado Ciarcia o Jerry Pournelle (de quin hablamos en el número de julio).

NO CELEBRATION of BYTE's 10th anniversary would be complete without the acknowledgment of some of the events and contributions that helped to shape the magazine. In the too-few pages that follow, we tried to capture some of the flavor of the past 10 years. Special thanks to all contributors and to the BYTE staff, especially Gregg Williams, who chaired the project, Richard Shuford, Rich Malloy, Mark Welch, and Stan Wszola. A Microcomputing Timeline Notable Quotes Evolution of the Microprocessor Interview: Carl Helmers Interview: Steve Ciarcia Ciarcia's Prodigious Output Interview: Robert Tinney Tinney Favorites Interview: Jerry Pournelle

Vale la pena ir siguiendo los enlaces a la revista que dejo en cada imagen, aunque solo sea para disfrutar de las maravillas del diseño industrial de la segunda mitad de la década de los setenta y la primera mitad de la de los ochenta…

Fotos de los ordenadores Sphere 1, Kim-1 y el Altair 8800, todos ellos de de 1975

…como el Sphere 1, el Kim-1 o el mitiquísimo Altair 8800 por ejemplo.

Recuperando temas que a veces no recordamos que vienen de muy lejos, los teclados:

Que el Keyport 717 tiene bien poco que envidiarle al más loco de los teclados actuales.

En la Kernel del mes nos encontramos con el lanzamiento de Excel en una conferencia conjunta de Microsoft y Apple porque, como igual no sabías, Excel era originalmente una aplicación para el Mac.

Kernel The ongoing construction work at Chaos Manor made it desirable for Jerry to escape yet again. He attended a joint press conference held by Microsoft and Apple in New York. The product introduced at the conference, Excel, is a spreadsheet for the Macintosh. Comments made at the press conference caused Jerry to put down some thoughts on software integration and whether or not we need it. He also looked at several new products, including a new version of BASIC from the inventors of the language. This being our anniversary issue, Dick Pountain brings us a condensed history of personal computing in Great Britain. He also introduces us to a rugged new lap-held portable, the Husky Hunter. From Japan, Bill Raike sends us an abbreviated history of that country's microcomputers and also discusses an innovative new product from Brother Industries— the SV-2000 Software Vending System. In this month's According to Webster, Bruce describes his experiences at the West Coast Computer Faire. He discovered that it isn't as much fun as it used to be, but he found some interesting products on display. He also discusses Apple's plans for the Macintosh, predicts success for the Amiga, and looks forward to testing a host of new products. Bob Kurosaka discusses the world of transcendental numbers in Mathematical Recreations. Some of them are familiar to us, such as e, the base of natural logarithms, and ir. He looks at some hiding places for these two numbers and some ways to approximate their values.

Y no podíamos saltarnos, claro está, sobre el textito que le dedica Pournelle al Amiga, por el que vota como sucesor del Apple II a finales de los ochenta. Ojalá, Jerry. Ojalá.

Amiga Among its other faults, Apple has been shamelessly neglecting the Apple II family, and specifically the Apple IIe. When the IIc came out a year ago, Apple cut the price of the IIe and slowed production, figuring the machine would die of its own accord. Instead, the sales jumped dramatically, easily outselling the IIe. People would see the IIc ads, come into the computer store, and walk out with a IIe. Why? Because the IIe had slots, while the IIc (like the Mac) was a closed machine. The IIe is a chameleon: With the right set of boards, you can make it look like and do just about anything. Case in point: The nicest development system I've ever used, including mainframes and minis, was an Apple IIe with 128K bytes of RAM, an AcceleratorIIe card (3.5-megahertz 65C02), and two Axlon 320K-byte RAM disks (configured as four 160K-byte floppy disks). Apple's response to the increased IIe sales was to cut back on production and raise its price (while discounting the IIc). Even so, it wasn't until late 1984 that the IIc finally started outselling the IIe. What does this have to do with the Amiga? Well, several machines are competing in the low-end market: the Atari 520ST the Apple IIe, the Mac (to a lesser extent), and the Amiga. Guess how many of these are easily expandable? Just one: the Amiga. Guess which machine will probably end up being the Apple II of the late eighties? I don't think the IIe will, nor the Mac, and the ST is a tightly closed, nonexpandable box. My vote is for the Amiga. From what I can see, the Amiga's graphics, sound, 68000 processor, memory map (allowing up to 8 megabytes of RAM), and expansion bus give it the potential of a long and successful life. There's always the chance that Apple will, indeed, come out with a souped-up Apple II next year, but even with the Western Design Center chips (65816. etc.) and the nifty 3 /2-inch Duodisk (1.6 megabytes of storage), it will probably be too little, too late.
(Dejo la imagen enlazada a la versión grande de la imagen, y no a la fuente en el Archive (aunque siempre tenéis la opción de leer el texto alternativo de la imagen).)

Y cierro con dos piezas más. La primera, lo normal en las revistas actuales (no): el típico artículo de dedicado a los números π y e…

pi, e, and All That Sneaking up on transcendental numbers by Robert T. Kurosaka God made integers, all else is the work of man. — Kronecker This famous quote of Leopold Kronecker serves as the starting point for this month's column. The integers (the whole numbers) can be used to construct other numbers. We can construct rational numbers by dividing one integer by another. When we do so, we get either a terminating decimal (1/4=0.25) or a nonterminating, repeating decimal (7/18 = 0.388888 ...). Repeating, or cyclic, decimals are a fascinating study I may explore in a future column. Irrational Numbers We can also construct numbers that are both nonterminating and nonrepeating. It is a rather amazing notion that a string of digits may go on forever without having to establish a pattern. It's such an odd notion that the ancient Greeks originally did not believe it possible— or even imaginable. When it was established that the square root of 2 was such a number, the Greeks called this kind of number irrational. The root meaning of irrational is "without ratio,'' or unable to be expressed as a fraction. The Greeks found such numbers irrational not only in the sense of "non-ratio-able" but also in the sense of "nonsensical." The differences between rational and irrational numbers are substantial. It can be shown that no more rational numbers exist than do whole numbers, but irrational numbers outnumber rational numbers. This fact, which is often presented as a paradox, is not especially surprising when you look at how we have constructed rational numbers. They are built up out of whole numbers and can be expressed as integer fractions. As I said above, irrational numbers cannot be so expressed. TWO TYPES There are two different kinds of irrational numbers. The first, like the square root of...

…y el segundo dedicado al Versabraille II, un ordenador diseñado para funcionar usando braille, porque la preocupación por la accesibilidad tampoco es nueva:

VersaBraille II Telesensory Systems has introduced the VersaBraille II system, a portable, disk-based electronic information processor for the blind. This braille computer lets you electronically store, process, and retrieve information. A special telephone modem can link VersaBraille II to other computers. VersaBraille II consists of a standard 3 '/2-inch microfloppy-disk system and a braille display that substitutes for a video monitor. Its memory holds up to 30.000 characters; disk support boosts the unit's capacity to 77,000 characters. This is adequate for many word-processing procedures, such as formatting, high-speed searching, and inserting, deleting, and relocating text. The system can simultaneously output braille and print information. VersaBraille II is fully programmable. Menus guide the user to each of the system's programs. The manufacturer provides special software that converts VersaBraille II into a four-function calculator with algebraic logic, floating decimal point, square root, and percent. Plans for other software packages include a 50,000word spelling checker, a two-way braille translator. and a language interpreter. The price of a VersaBraille II system is S6995 plus shipping and handling.

He encontrado poca información sobre el Versabraille II, pero si alguien quiere investigar sobre su antecesor, el Versabraille original, aquí un documento por el que comenzar.

Apa. Volvemos el mes que viene con el número de octubre. Por cierto, que si alguien quiere hacer los deberes por su cuenta, además del archivo de la revista en el Archive, también tenéis esta chulada de navegador que me pasó hermanito hace unas semanas.

¡Hasta la próxima!

2 opiniones en “Byte, septiembre del 85. Diez años de Byte”

  1. Pues me parece extremadamente llamativo el uso del término «single-board computer». Yo diría que descubrí el término en 2012 con el lanzamiento de la Raspberry Pi.

Deja un comentario