A 50-year old engineer's journey through nostalgic times My first experience with floppies and floppy diskettes was 8-inch BSMDs (Burroughs Standard Mini Diskettes) and ISMDs (Industry Standard Mini Diskettes) on the Burroughs CP-9500 super mini at TCS Maker Towers, Cuffe Parade in Oct. 1982. BSMDs were heavily engineered. They had a hard sleeve and the drive in which they were used was complex. The drive would slide open a window in the hard sleeve, spin the diskette and engage the heads. Needless to say, reliability problems, were the order of the day.
But lets back track another 8 years to when I was 16, an Inter Sc. Student at SIES, Sion. The Intel 8080, the first real microprocessor that ushered in the microcomputer revolution was announced in April 1974. The following summer in 1975, I joined engineering college.
In Diwali 1975, my father gifted me a hand-held 4-function calculator-- a Casio Pocket-Mini. Its CPU and display driver were integrated in one chip(μPD974C). It had an eight-digit vaccuum fluorescent display. I had been trying to fathom programming, reading a Fortran-IV programming textbook.
The Casio calculator and Fortran stimulated interesting conversation with a senior colleague of my dad's. Got the coveted opportunity as a first-year engineering student, to go to Bhabha Atomic Research Center (BARC) during the summer vacations to get hands-on training in computer programming. That year, Paul Allen and Bill Gates wrote their BASIC interpreter for the MITS Altair, the first hobby microcomputer and founded what would become Microsoft. They worked out of a mall in Albuquerque, New Mexico.
Back then, BARC had a Soviet BESM-6 (БЭСМ-6) mainframe. БЭСМ stands for “Быстродействующая Электронно-Счетная Машина” “Bystrodeystvuyushchaya Yelektronno-Schetnaya Mashina”, meaning, “High-speed Electronic Calculating Machine”. The BESM-6 was a 9-MHz, 1 MIPS, 48-bit machine with 60,000 transistors and 1,70,000 diodes.
At college, we were studying numerical methods. My first computer program was to implement the Newton-Raphson root-finding algorithm to find the roots of a polynomial.
So at BARC, in the summer of 1976, I used the powder-blue IBM Type 029 and Type 129 key punches to produce decks of Hollerith cards with Fortran-IV source code. A deck was submitted for batch-processing and in the morning a 132-column drum-printed output would be waiting for me in my pigeon-hole in the varnished plywood rack at BARC North site.
Back in the late seventies, computer architecture was implemented with TTL chips and MSI logic. As an engineering student at BARC during the Diwali-1976 holidays, I learned to program in assembly language on an ECIL 16-bit TDC-316 mini.
The TDC-316 was developed in India, contemporary to the PDP-11 at Digital Equipment Corp. (DEC),Maynard, Massachusetts.
In 1976 DEC announced their first 32-bit supermini: the VAX series. Suddenly the current 16-bit machines were outdated. Data-General immediately launched their own 32-bit effort to beat DEC to market. They called it the “Fountainhead Project”.
However, two years later in 1978, the VAX 11/780 was released. Fountainhead proj. mgmt. had failed to beat DEC to market. DG then killed Fountainhead and launched their “Eagle Project” a crash 32-bit effort based on the Eclipse.
Tracy Kidder's 1981 book, “The Soul Of A New Machine” described these travails. The book won the Pulitzer. The DG Eclipse MV/8000 was finally delivered in 1980, the year I graduated.
Going back, the September 1977 issue of the Scientific American, was an eye-opener. I convinced my college librarian to let me borrow the library reference copy long enough to photocopy it.
In 1977, photocopying was a laborious expensive process. The plain-paper xerographic electrophotocopier was the size of an autorickshaw.
You mounted the page to be photocopied on an easel. Then the photographer would charge a selenium-coated photo-receptor plate with static electricity. He would then place the plate into his bellows-type view camera and photograph (shoot) the page. The plate with an electrostatic image would then be placed in an aluminum box and dusted with fine black toner powder. The toner would stick forming an image on the plate, which would then be placed in a “fixer” along with the paper. A heat-fixing process would cause the toner to adhere to the paper.
Using this wonderful machine, I copied the whole Sept.’77 issue of the Scientific American. That photocopy marked a watershed in my engineering ambitions. It inspired me to build a career on the microcomputer frontier.
Computer design has made great strides since the Intel 8080. An ever-increasing amount of functionality is integrated in-silicon. CPU-design battles are now fought with large gate count, IP-based, bus-intensive system-on-chips (SoC), designed using SystemVerilog.
The essential spirit of the high-tech industry, the feverish pace, the mystique, the go-for-broke approach to business continues, as the industry pursues mind-bending technological innovations with new blood pouring out of the engineering colleges.
Ram |