I learned programming on an Indian Altair/IMSAI 8080 clone that used the S100 bus in 1978 - the ECIL Micro-78. One memorable weekend was spent trying to debug an obscure bug which turned out to be single memory location failure in a 16 KByte DRAM board that used Intel 4k DRAMs for the first time. DRAMs had a well-earned reputation for unreliability at that time, but had ~4x the density of SRAM from the same era, so they were more economical. In 1982, I started programming in C for the first time on a Cromemco Z80 box at college, and that one was as bulletproof as it got for a dual 8" floppy box of that era.

The importance of a plug-in bus like the S100 was a crucial insight that enabled the PC revolution, and the first wave of the microcomputer revolution happened mainly due to bus compatibility. Eventually, Apple and IBM also used plug-in buses which were essentially simplified or smaller versions of the S100, using signals from the 6502 and 8088 instead.

The 16-bit extension of the IBM PC plug-in bus became known as ISA, and ensured its enduring domination even to this day (it's still around as the X-bus inside modern highly-integrated Southbridge ASICs, though it's no longer brought out as an external plug-in bus, except in some legacy industrial PCs).

> They had been recommended to me as 'tanks' and were much more reliable than the dynamic memory cards of the time.

Interesting to hear that.... I've recently been reading through old issues of DTACK Grounded, and Hal Hardenberg said similar things, and at length, about the reliability improvements you get with static memory over dynamic. (A few years in, he was also selling dynamic memory products, presumably for scale, and was no longer able to ground DTACK in his designs.)

I missed Hal's rants, you can read DTACK Grounded online here: http://www.easy68k.com/paulrsm/dg/