Thomas "Todd" Fischer posted the following on comp.os.cpm. I thought it was worth cross-posting.
This discussion seems to still have legs, so here's a rather profound statement from Joe Killian in the first-ever contact between him and Ed Roberts last Monday regarding the attainment of "critical mass" of acceptance for the bus standard developed by Roberts and Bill Yates:
"Your card size, bus connector and signal definitions were copied, by IMSAI, Processor Tech, and countless others, either for complete systems or add-in boards. This was the sincerest form of flattery, done because I and others saw supporting your design as the best business path to pursue. I do think that this support in the form of add-in cards and alternate platforms using the same bus was fortuitous for all of us, in that it snowballed into the fledgling industry's standard. I've always sort of felt that my choice of using your bus, and thus IMSAI quickly being out there as a second MITS compatible machine, tipped the balance on the part of all the others wanting to enter the microcomputer market.
"Without such a standard, I would guess that neither MITS nor the industry would have taken off nearly so quickly. Nothing else from your machine was copied in the design of the IMSAI (save the use of the 8080 chip, of course). By contrast, I had the dubious privilege along the way of examining a competing chassis that copied my IMSAI chassis right down to holes I put in for options that we never used."
Ed Roberts responded back to Joe about the observation that on the original Altair processor board, there are no signal line vias between the
8080 processor chip and the connector. Joe asked if this was done to ease the layout task. Ed responded:
"Bill Yates and I personally did the PC layouts. And yes the layouts were done bas[ed] on what was the easiest and quickest not on what was the most logical.
We were grossly under staffed, overwork[ed] and essentiall[y] bankrupt at the time we did the Altair.
Maybe this will give you a view of my perspective of early events.
So now you can reflect back when dealing with issues of bus interface next time you're working on your own "time machine", regardless of it's name!
-Thomas "Todd" Fischer