Actually I do mean PS/2, not microchannel. I once wrote a
keyboard driver for DOS, and the interface itself was a piece of
cake. The point of the PS/2 interface wasn’t revolution but
compatibility. But there’s more to an interface than an API or a
register access discipline: here you are hearing from someone
who has being quite active on the hardware side of the fence,
so, I look at the packaging and wiring too, not just at what the
programmer sees. To me who was actively desigining PCs back in
1987, PS/2 was the whole thing: microchannel, the new keyboard,
the new machine packaging, and so on. Had IBM not being so
jealous about protecting the microchannel as their intellectual
property, we might have seen the microchannel bus standard
depart from its PS/2 roots; but it didn’t.
And the PS/2 mouse, well, it’s still in use today.
About PnP, I don’t think it’s that much of a good thing, at
least not the way it’s set up today. First, it inverts the
natural order of things: software should be written for
hardware, not the other way around. Second, hardware standards
should be industrywide and not wielded by software companies,
so, any kind of PnP standard should come built in inside specs
such as the PCI bus and so on. PnP, actually, should be a
hardware thing and not a software thing, and it should not be
driven by a software company but by a consortium of hardware
providers. Now, power is wildly overrated, and here again, it
should be a hardware standard.
Power management belongs in the Bios, and it should be totally
and completely OS-independent: the OS has no business overriding
the Bios. In fact, I strongly believe that low level engines
such as dispatching processes and such should be included in the
Bios, not in the OS. About 10 years ago, power was handled by
power management chips that worked in tight cooperation with the
Bios; and if the idea of a protected mode Bios had established
itself (like for example IBM’s ABIOS), much of that
functionality would be implemented at hardware and firmware
level and it would be common to any OS.
This is, in fact, the key to what I’ve said: I believe that a
good part of today’s OS functionality in Windows (and Linux!)
belongs in the hardware and in the bios, not in the OS. As I see
it, today’s OS’s are big, fat, inflated matrons, trying to
emulate amoebas and gobble up everything around them. Also, OS’s
should come in removable memories and no OS component should
require to be installed in a hard drive: hard drives should be
for user data alone. Oh, don’t get me started on this, today’s
OS’s are vintage 1970 at best, and there’s no sign of any
relevant innovation anywhere in the horizon.
As for the original PS/2, if I remember correctly it had not
cables and no screws to fiddle with: everything snapped in and
out. I also remember much of it being plastic and not metal. It
took about one or two minutes, if I remember, to take the whole
machine apart, and another minute or so to put it back together,
because everything snapped in, again, no cables and no screws,
no jumpers, no twisting hands and arms to reach a hard-to-get-to
place in the motherboard, nothing but snap-in.
Alberto.
----- Original Message -----
From: “Tim Roberts”
To: “Windows System Software Devs Interest List”
Sent: Tuesday, November 22, 2005 1:44 PM
Subject: Re: [ntdev] PS/2 mouse state machine
> Alberto Moreira wrote:
>
>> /Actually, given the vintage of the PS/2 interface - 1987 if
>> I’m not mistaken, before the PCI bus, hardly after Multibus
>> hit the world, at a time where ISA and EISA were the only PC
>> alternatives, it was quite an improvement./
>
>
> The IBM PS/2 and the Micro Channel Architecture bus were
> introduced in 1987.
>
> At the time, ISA was the ONLY alternative. EISA was created
> in response to MCA, and VLB in response to that. PCI work did
> not begin until 1990.
>
> However, the PS/2 keyboard and mouse interfaces are unrelated
> to MCA, except that they first appeared on the same systems.
> The PS/2 keyboard interface is identical to the AT keyboard
> interface (except that the connector has one additional pin),
> and the PS/2 mouse interface was little more than a wrapper
> around a serial port.
>
>> / Some of you may have worked with the old IBM PC keyboard,
>> so, you will note the difference. Moreover, the PS/2 was an
>> impressive line of machines, and it was the first time in
>> personal computing where the hardware was described by
>> configuration files (I forget the name of those little
>> files!) and access to it was done through a set of
>> configuration i/o port addresses specifically allocated for
>> the purpose./
>
>
> EISA called them, oddly enough, “configuration files”, with a
> .CFG extension.
>
>> /The PS/2 interface was an example of something that I
>> believe we lost today: simplicity. No PnP. No power
>> management. No bells and whistles./
>
>
> You mean “Micro Channel” interface, not “PS/2” interface,
> right? The PS/2 keyboard and mouse interfaces were anything
> but revolutionary. They were just a repackaged version of
> what already existed.
>
> I’m not actually sure what you’re trying to say here. PnP is
> a software thing, and I doubt there is a single person who
> would argue that hardware power management is a bad thing.
> It’s tricky to manage in software, but it would have been
> tricky with Micro Channel as well.
>
>> /This brings me good memories. I was working for Wang
>> Microsystems back in 1987 when the PS/2 came to live, and we
>> queued up early morning at the local computer store waiting
>> for it to open so that we could get our hands on the very
>> first machines. We bought a Model 60 and (if my memory
>> doesn’t fail me) a model 80, and took them to the lab. The
>> packaging was absolutely innovative, so much so that even
>> today no existing machine can match it - and we made a little
>> contest among the engineers to see who could unassemble and
>> reassemble the PC in least time./
>
>
> I think your memory is enhancing things a bit. Micro Channel
> boards were easy to remove, because they had a big plastic
> handle at one end, but there are lots of machines today with
> cases that are much easier to disassemble than the PS/2.
>
>> / /
>> /The whole microchannel concept was very nice and rather
>> innovative, too bad it died such an early death./
>
>
> It died because of predatory licensing demands by IBM, not
> because of any serious design flaws.
>
> –
> Tim Roberts, xxxxx@probo.com
> Providenza & Boekelheide, Inc.
>
>
>
> —
> Questions? First check the Kernel Driver FAQ at
> http://www.osronline.com/article.cfm?id=256
>
> You are currently subscribed to ntdev as: xxxxx@ieee.org
> To unsubscribe send a blank email to
> xxxxx@lists.osr.com
