Ah, yes, the IBM PC, that used a busted CPU that used busted memory addressing, whose keyboard was deliberately busted so as not to compete with their own division that sold turnkey wordprocessing systems -- creating repetitive-strain injuries in generations of office workers.

The 'legacy of a standard BIOS' is a legacy of a standardized crippled architecture.

Would Linux have ever gotten anywhere if IBM hadn't been the company it was and is?

Remember the IBM PCjr? Now that was deliberately crippled.

They sold those things with a monitor for about $1000. It wasn't really a bad deal at the time, although you'd need to buy software like Lotus 1-2-3 on a cartridge not unlike a C64. The non-geeks might have been happy with it.

It was a terrible deal for computer stores because it brought in a completely different set of customers who'd expect the same quality of service.

> Would Linux have ever gotten anywhere if IBM hadn't been the company it was and is?

Well, the reason that IBM jumped on the Linux bandwagon was because Linux was already going somewhere. If having a good Unix clone for free was the only consideration, I'd think that one of the BSD family would have been picked on [narrowly] technical and licensing grounds.

It's been so long since I've used an IBM PC keyboard that I don't even remember the keys on it (I've got a stack of about 40 IBM AT keyboards at home, and, maybe, one PC keyboard, and that only because I've got a couple of old 5150 motherboards floating around.

Alan: forget the architecture and think of the externalities -- namely, standardization. (Or would you rather live in a world where CP/M-style incompatability reigned unchallenged?)

I'll grant you that an IBM PC design that came out a year later and used the 68000, VMEbus, and Xenix would have been one whole lot nicer. But it would also have cost too much to compete with the offerings from companies like ACT (remember them?) or Apricot (in the UK). A good idea? You be the judge ...

At 04:07 PM 12/5/2004, you wrote:
>Alan: forget the architecture and think of the externalities --
>namely, standardization. (Or would you rather live in a world
>where CP/M-style incompatability reigned unchallenged?)

How is the PC as we got it much different from CP/M-style
incompatibility? :-)

>I'll grant you that an IBM PC design that came out a year later
>and used the 68000, VMEbus, and Xenix would have been one whole
>lot nicer. But it would also have cost too much to compete with
>the offerings from companies like ACT (remember them?) or
>Apricot (in the UK). A good idea? You be the judge ...

Various manufacturers did release many machines in the 80's that
were effectively what you describe, a 68K and some form of Unix,
and you're right, they were all way too expensive.

My office mates were reminiscing about this recently. They're a bunch of IBM mainframe geeks and were lamenting IBM's decision to go with the 8086 running a CP/M clone, rather than the S/370-on-a-chip running MVS. Would this have been OS/2 ten years early? It probably would have had the same cost problems as the real-world Unix-based architectures did, but it's fun to imagine worlds where powerful software tech won in the market. (It could still happen.)

David: in case you've forgotten the nightmare that was CP/M ... a different BIOS for every manufacturer's machine was only the start. There was then the problem that every manufacturer picked a different floppy disk formatting scheme, the better to lock out their competitors. Not to mention using different TTY control codes. This wasn't too bad for the software developers if they restricted themselves to writing code that used the BDOS and no undocumented or proprietary BIOS calls and which had some sort of bastard cousin of Termcap to figure out how to address the screen, but for high-performance stuff it was a nightmare of market fragmentation. It meant the business computer market (at least, for S-100 based CP/M boxen) was as fragmented as the UNIX market was, a decade later.

The one unequivocally good thing that the IBM PC brought was a measure of standards-based sanity, because even crap standards are better than deliberate market-sequestrating incompatability.

At 02:07 PM 12/7/2004, you wrote:
>David: in case you've forgotten the nightmare that was CP/M ...

Alas, no, I can't seem to purge the last memories of it from my mind!
>a different BIOS for every manufacturer's machine was only the
>start. There was then the problem that every manufacturer picked
>a different floppy disk formatting scheme, the better to
>lock out their competitors.

Yeah, that one had me pretty cheesed...and it wasn't just the CP/M folks, either. All the freakin' other systems all had totally
whacked formats, too. Apple ][ floppies different that C-64 floppies different that Atari 400/800 floppies...but yeah the CP/M universe was that madness writ large.

> Not to mention using different TTY
>control codes.

I blame the terminal manufacturers for starting that fine tradition!
> This wasn't too bad for the software
>developers if they restricted themselves to writing code that
>used the BDOS and no undocumented or proprietary BIOS calls and
>which had some sort of bastard cousin of Termcap to figure out
>how to address the screen, but for high-performance stuff it was
>a nightmare of market fragmentation.

Early MS-DOS wasn't a lot better in this regard, but not QUITE as nightmarish...christ I'm having flashbacks to writing assembler
graphics routines back in the day now.

>It meant the business
>computer market (at least, for S-100 based CP/M boxen) was as
>fragmented as the UNIX market was, a decade later.
>
>The one unequivocally good thing that the IBM PC brought
>was a measure of standards-based sanity, because even crap
>standards are better than deliberate market-sequestrating
>incompatability.

Well I suppose jacking around with batch files and config.sys
endlessly IS at least do-able for mortals, compared to "you need
to fix this BIOS call in assembler" :-)

-David

Let's not forget that IBM didn't really want a standard either. They were just too big that whatever they did would become it.

If IBM had based their PC on CP/M then we'd have seen the same effort by others to reverse engineer and make their systems compatible with that. It's probably a good thing that Microsoft was able to sell to other hardware companies, but I think we'd have had a standard regardless.

And I still do batch files.

There's a big memory hole that's developing here. Am I the only one who remembers when IBM was the leading manufacturer of electric typewriters? Wasn't that office equipment, too?

At 11:33 AM 12/9/2004, you wrote:
>There's a big memory hole that's developing here. Am I the only
>one who remembers when IBM was the leading manufacturer of
>electric typewriters? Wasn't that office equipment, too?

Oh don't I know it, I learned to touch type on Selectrics while
taking a typing class after being frustrated from 5 years of hunt and peck typing. IBM Selectrics with no letters on the key
caps were used for tests. So you really learned to type.

Like someone mentioned in this thread a few messsages ago, the
original PC keyboard sucked partially to not cannibalize their
dedicated wordprocessor sales; they also didn't want to kill their (now dead? :-) typewriter division.

I guess they couldn't see the typewriting on the wall.

Edited by author 09-25-2006 01:53 AM
I am looking for an old keyboard with keys which were only about 1/8" tall, and bigger around than the modern keyboards. I saw and used this keyboard when I was in college in the computer lab in the late 80's. I have carpal now, and am looking to find one of these computer keyboards to make writing easier.

Does anyone here know of such a keyboard?