From: Richard Erlacher <edick_at_idcomm.com>
Date: Sun Oct 28 13:16:34 2001

No, the evolution of the S-100 market was really quite different from that of
the IBM-PC market. The longer the S-100 was allowed to progress on its own,
board vendors, with the notable exception of Godbout/CompuPro, who really never
even guaranteed his boards to work with others of his own manufacture, tended to
make their products more and more interoperable, as the market demanded it. It
was a terrible overhead for them to buy back boards that wouldn't work with a
customer's other equipment, so they made every effort to get around that.

The proposal of a "standard" with the "optional" features they managed to put
into it, put a quick stop to that. Most of the board makers had worked out most
of the interoperability issues by that time, and the development of a standard
offered them an opportunity to reverse the process, forcing their customers,
foolish enough to believe their marketing line suggesting some sort of
superiority or other, to buy new boards, all from the same manufacturer. This
meant that whereas one previously used a CCS CPU and FDC, SD Systems' RAM, and
TEI's I/O board, now one had to have boards with the same maker's trademark. In
the case of CompuPro, the trademark wasn't enough. One had to have boards with
the same series number, and, in some cases, the same manufacturing date, or they
couldn't be relied upon to work together. One had to cope with the fact that
some boards required a bus signal be grounded while others required it be tied
high, while others yet required it be left floating or have a specific signal on

The technical discussions in the standards committee meetings quickly
deteriorated into discussions over which features should be mandatory and which
should be optional. When the standard was adopted, it really wasn't a standard
at all, since so many features, labeled optional, when implemented in accordance
with one alternative, could cause another maker's product, equally compliant
with the standard, to malfunction to such extent that it or the entire system
didn't function.

Once the standard was adopted, and by that I mean within weeks, the S-100 bus
was a dead duck. From that point forward, the only way to ensure a set of
boards complied with the same standard was to buy the set at the same time, from
the same retailer, with a warranty to assure the buyer that the boards would
work together. Of course one quickly learned he was high and dry as far as
other products were concerned.

The PC market started with a firm standard, however goofy (e.g. positive-going
interrupt and reset signals on the bus, etc.) that, when faithfully adhered to,
worked pretty much all the time so long as one didn't tamper with it. Once IBM,
with its massive influence on the market (at least back then) set a standard,
3rd party manufacturers knew they had a large enough market for products that
worked on that standard, and, frankly, they could live without IBM's
competitors. Now that IBM is really of no further interest in the market, like
every other large corporation that's attempted to assume the role of Mr. BIG on
the market, nobody sets the standards, hence there's no penalty for failint to
adhere to the standard/convention/whatever, that drives the design of new

A typical example is the 64-bit PCI. I've seen several desirable cards offered
for the 64-bit PCI, yet I've not seen even ONE advertised motherboard that
actually supports 64-bit PCI. From what I've read, 32-bit PCI boards work in a
64-bit slot, yet nobody seems to be offering that feature in their
advertisement. I'd guess that's because it's cheaper/easier to diverge slightly
from the published standard in making these products, hence they don't make
claims about compatibility.

It's all about money, after all.


----- Original Message -----
From: "M H Stein" <mhstein_at_usa.net>
To: "'ClassicComputers'" <classiccmp_at_classiccmp.org>
Sent: Sunday, October 28, 2001 9:13 AM
Subject: Re: VLB SCSI?

> Well, I'm not sure that the S-100 "standard" was all that rigidly adhered to
either, not to mention that everything else was proprietary at that time (not to
mention IBM's infamous Slot 8, PS/1's and MCA, non-standard memory modules,
> New PSU fans are still abundantly available up here in Toronto; if ya really
need one, or a complete PSU (especially an odd-sized one) or anything else for
an old PC, drop me a line.
> mike
> -----------------Original Message---------------
> Date: Sat, 27 Oct 2001 17:19:50 -0600
> From: "Richard Erlacher" <edick_at_idcomm.com>
> Subject: Re: VLB SCSI?
> <snip>
> because IBM no longer provided a firm ISA-equipped standard that everyone else
> clone, there was nothing but chaos, and that's the way it's been ever since.
> The whole concept of "standard" was corrupted in the course of this
> step, and it was a step into the quagmire we're all swimming in today. <snip>
> <snip>
Received on Sun Oct 28 2001 - 13:16:34 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:34:21 BST