CompuServe Messages

#9600 Modem

    29-Nov-91 16:57:25
Sb: #41843-#9600 Modem
Fm: Shraddhan 70374,754
To: Don Curtis/SYSOP 76703,4321
Aaah! The good old days! The first computer I ever used was the university mainframe – an Elliott 803B. It was the fully expanded version – 8192 words of memory at 39 bits (plus parity) each. The Algol compiler and run-time code took up about 75% of this, and longer programs had to be compiled out onto paper tape. I had the luxury of having two high speed paper tape readers, capable of almost 330 charcters per second (5 bits per character), but running at such a high speed, the readers were prone to making errors, and I remember I had to slug the code somewhat to drop the speed to a still fast 300 chars/sec. The output devices were a pair of high-speed paper tape punches, which zipped long at 100 characters per second. The output tape then had to be wound up and carried next door to a spare Creed teleprinters (there were three in all, for maximum throughput), where the output was printed at a tolerably fast 10 chars/sec. The computer was controlled via a keyboard, about the size of a desk, at which instructions could be entered in machine code. Of course, as with all the best computers, there was also a loudspeaker. The speaker was pulsed every time a jump or floating point math instruction was executed (yes, it even had a math co-processor!), and with a bit of cunning, you could write a program to play a tune. The processor was perhaps a little slow by modern standards – 576 microseconds for most instructions (excluding floating point math), but only 288 microseconds for a jump instruction. Yes, microseconds, not the new-fangled nanoseconds, but bearing in mind that it was a serial processor (i.e. it carried out its instructions on a bit-by-bit basis, not byte-by-byte), this was quite understandable. Ah, the joys of twenty years ago! Shraddhan – via Whap!