24 Nov
2009
24 Nov
'09
3:42 a.m.
Chuck Guzis wrote:
On 20 Nov 2009 at 9:54, CSquared wrote:
I suppose it is all about what one becomes familiar with. I always
wanted but could never seem to afford one of the Altairs, IMSAI's,
Cromemco's etc. and had to be content with the stuff at work. Besides
the Microdata, that consisted of 6800's, then 6809's (what a great chip
IMHO) with no switches or blinkey lights at all - just a debugger
running from a terminal. Then there were Datacraft/Harris 6024's and
SDS 910's which had both switches and blinkey lights but were 24 bit
word length machines with everything in octal and in groupings of 3 bits
per group. I think it would have been tough thinking in hex for those.
Characters for the SDS were only 6 bits so one could pack 4 per word.
Since we only had 4K or 8K words in the 910, packing characters that
way was fairly important. Hmm, lets see, there was also a bit of Z80
and lots of 8086 but none of those devices had a really fun control
panel either. I do seem to recall from the Z80 days noticing that some
of the instruction coding seemed more natural in octal than hex which
fits with your thoughts about the 8080.
Later,
Charlie C.
My old Microdata 820 has a 4-4-4-4 front panel
pattern which is rather
easier to deal with I think. Of course you realize this discussion
has me wanting to drag it out of the garage and get it going again. :)
The switches on the MITS Altair are grouped in threes; the IMSAI
allowed one to (by virtue of paddle colors) group in 3 or 4 (or 2).
I simply ignored the grouping on the MITS and thought in hex anyway.
It would have been much more difficult had the display on the system
been octal via 7-segment displays.
(I know that 8080 instructions, especially register moves, lend
themselves to octal representation, but that was less important to
me).
--Chuck