While this question ask been well answered, it does bring to mind a good old trivia question.... "What significant advantage did octal have over hex notation (especially in the late '60s timeframe)?"

derMouse: You win the grand prize! The actual devices that prompted it however were printers! A printing calculator could be adapted to produce hardcopy. These were in (relatively) common use and available. "Most" systems used binary for lamps/leds so nixie tubes for non decimal values were not predominant. As you know the use of text in computing [for entry/display] came later. David. >>> -----Original Message----- >>> From: cctalk-bounces(a)classiccmp.org >>> [mailto:cctalk-bounces@classiccmp.org] On Behalf Of der Mouse >>> Sent: Friday, June 25, 2004 8:17 PM >>> To: General Discussion: On-Topic and Off-Topic Posts >>> Subject: Re: Really stupid PDP assembler question >>> >>> > While this question ask been well answered, it does bring >>> to mind a >>> > good old trivia question.... >>> >>> > "What significant advantage did octal have over hex notation >>> > (especially in the late '60s timeframe)?" >>> >>> Well, I wasn't around then (at least not with respect to computers). >>> But I'd hazard a few guesses. >>> >>> - Word lengths that were multiples of 3 were commoner than >>> word lengths >>> that were multiples of 4 (of course, some, eg 36 bits, >>> were both). >>> >>> >>> - Using letters as "digits" ran into human mindset trouble; using >>> decimal representation for hex digits runs into bigger trouble, >>> using multiple characters per functional digit. >>> >>> - Humans have trouble with a 16-digit system. (I know I >>> do; I always >>> have to stop and pay attention to avoid getting B and D >>> confused.) >>> >>> How close did I come? :-) >>> >>> /~\ The ASCII der Mouse >>> \ / Ribbon Campaign >>> X Against HTML mouse(a)rodents.montreal.qc.ca >>> / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

Tom Jennings wrote: Don't know if that was the real reason, more important probably was the fact the 5 bits could code up to 32 symbols, sufficient for the alphabet. IBM did not "buy" the hollerith card, IBM was the evolution of a company, the Tabulating Machine Company, created by Herman Hollerith in 1896. In 1911 this merged with two other companies to form CTR, the Computing Tabulating and Recording Company. Thomas Watson took the reins of CTR as general manager in 1914 and became its president in 1915. He renamed the company International Business Machines, IBM, in 1924. An the rest is history ;-) -- HansP

On Sat, 2004-06-26 at 14:44, Tony Duell wrote: I'm sure you are right.

On Sat, 2004-06-26 at 13:10, der Mouse wrote: You're extrapolating modern ideas backwards. It doesn't work. I would like to see some evidence for "existing printer mechanisms" being the basis for entire notational systems. 2, 4,, 8... is "logical" but that's immaterial -- bi-quinary was used a lot, 1 2 4 8 representations of BCD were not the only ones used, and people insisted on decimal internal representations when it's "obvious" that binary is better. Well yes, 6-bit characters, octal is convenient, was my (unstated) intent... I defer to Hans on IBMs history, and he may be right on the 5-bit thing too (I deleted before I got to reply), but Emike Baudot's code was manually entered, and variations of the code sure look like they ended up in ITA2 aka Baudot teleprinter code.

On Sat, 2004-06-26 at 14:41, Tony Duell wrote: THere is ABSOLUTELY NO advantage to one scheme over another, globally. Locally, there may be, eg. within a given machine architecture. Remember "split octal" in the 8080 world? A horrible joke, merely because the Intel designers put stuff on byte/nybble boundaries. It doesn't have to be that way, it can be anything. Now I'm a bit unclear -- before dot matrix took over the world, print mechanism schemes were almost countless -- belts, bands, drums, baskets, cylinders, wheels, daisies, cups, raster, XY, whatever -- and notation for whatever wacko scheme chosen would be utterly arbitrary. (I miss all the assorted mechanism, it was fun to see and listen to. I don't miss unreliability, incompatibility, mass, cost, ... :-)

I remember those, just, but only on fully mechanical calculators. But I also haev a Victor Comptometer calculator with solenoids mounted over the keys that was used as a printer on a data logging system, so maybe such machines were hacked onto computers too. I've seen early 70's calculators with tiny drum printers. A drum the full width of the paper -- perhaps 15 columns -- with the digits, -ve sign, decimal point, round it and a row of hammers that print the approriate characters. I've also seen them with a typewheel per column which were positioned and then all columns printed simultaneously. And then I've seen one with a carriage containing a single typewheel and a hammer mechanism. One of the strangest printers I've seen came off the data logger for a gas chromatagraph (a Kent Chromalog II if anyone wants to try to find details). There was a typewheel per column with a solenoid to step it round one posisition, and a switch that gave a carry function. It was essentially a long ripple counter. There was a hammer assembly the full width of the paper. You fed in the pulses (the instrument was essentially an integrator), and at hte approrpriate time you fired the hammer. I can't rememebr how the reset-to-zero worked. I don't think there was a separate solenooid for that. Maybe you just fed in a clock signal to all solenoids together and turned each one off when the carry contact flipped or something. -tony