and created a lot of new bugs

For a slightly different perspective, check out Silicon Cowboys: The Story of Compaq Computer. It has some great interviews and TV clips from back in the day. I worked at Compaq from 1988 to 1992 - one of the best companies to work for.

Silicon Cowboys Trailer: https://www.youtube.com/watch?v=7wjJYqUkHd8

Dave...

Interesting. I wonder if there's a comparable book. I see
Open: How Compaq Ended IBM's PC Domination and Helped Invent Modern Computing by co-founder Rod Canion. This book was possibly previously called The Compaq Revolution: An American Success Story

And I see there's an audio interview and a condensed transcript of an interview with him here.

It's a good book if you are interested in that time in history. I got mine autographed by Rod at the premier of Silicon Cowboys. Rod's a great guy. I left Compaq when they fired him - I knew it was going to be run by finance guys instead of engineers.

This link doesn't work. It is unusual because it is archived...

_________________
my blog about processors

How unfortunate! Try this link instead.

Thank you.

I dare to add a comment about Dr. David Patterson's "Fifty (or Sixty) Years of Processor Development… for This?" mentioned above:

Indeed, the 8086 was slow, but its competitors were also slow. The 68000 was much faster on 32-bit and large array operations. However, for processing bytes, the 68000 might even be slightly slower. I mean the cases where the 68000 and 8086 use the same clock speed. The 68000 was available at higher clock frequencies and this gave it more performance benefits. However, these benefits were offset by its higher price tag and very low FP performance. The 8087@5 MHz can easily outperform the 68000 at higher frequencies. The NEC V20 and V30 were accelerated versions of the 8088 and 8086 respectively, and they made the 68000 performance advantage visible on more rare occasions. The 80286 just closed the "68k performance advantage" topic.
IMHO it is still known not much about the Z8000 performance. We have a cryptic revelation from Masatoshi Shima that they intentionally were producing the Z8000 at lower frequencies because they didn't want the Z8000 to be faster than the 8086. This indirectly suggests that Z8000@4MHz has the performance close to the 8086@5Mhz. Maybe it is true but without thorough benchmarking it is just a suggestion.

I also dare to claim that the IBM PC was actually the best PC in 1981. Its CPU is at least 2-3 times faster than the 6502@1MHz. Maybe the TI-99/4A has slightly faster processor but it had only 256 byte of fast RAM and slow serial ROM interface. The Atari 800 performance might be almost two times greater than the Apple II performance and it was very close to the IBM PC performance...
The CGA graphics is much better than the Apple II, Commodore VIC-20, Tandy Coco, TI-99/4A graphics. The Commodore PET didn't have graphics at all. Only the Atari 800 had competitive graphics but Atari management wanted to sell more Atari 2600s than advanced new models. There were contacts between Atari and IBM managements then - one can think that this could affect Atari marketing...
Maybe only the IBM PC's audio system was completely inferior compared to the Atari 800, Tandy Coco or TI-99/4A.
But as it was mentioned afore the IBM PC had the best keyboard for that time and it was also an important factor.

_________________
my blog about processors

Warning: 2nd hand info .. old guy story ..

I was down at the December 1982 DECUS symposium and DexPo (sp?). I was there mainly for RSTS/E sessions, but at the show I saw DR had a booth set-up showing off Concurrent CP/M for the DEC Rainbow. We had a couple of them at the time running the most embarrassing MS-Dos imaginable, so being someone that cut my teeth on CP/M I was interested to see it. While I was in the booth Kildall showed up. Apparently he was giving a talk at the symposium. I never talked directly to him nor did I get to his presentation as my dance card was full. However, when I went out to diner that night Kildall was in the restaurant with a bunch of DEC guys, one of which I knew well. Next evening I met that guy at a pub and he said the question of why MS got the deal and not DR had come up. Acceding to him Kildall was in good faith negotiation with IBM at the time but Kildall was unaware that timelines at IBM had changed. He was on a business trip to see another client, which he undertook by flying his own plane to the meeting. While he was underway the whole dynamic of the deal changed as IBM had shifted gears and wanted to wrap things up then and there. Kildall did not get news of this until after the meeting with the other client and only as he was on his way out of the building. Apparently his office had called the other client to tell Kildall. Kildall called his IBM contact from the other client's office and was under the impression that he had a couple of days for him and his legal team to look at the particulars. He found out when he got back to his offices that higher-ups at IBM would not wait and had decided to pull the trigger with MS. Apparently they were still going over the "breach of understanding" or something like that at the time I heard this. I don't know how accurate all this is, but it's what I heard from a guy I trusted that claims he heard it directly from Kildall. FWIW.

Edit: I will say that we did end up buying the Concurrent CP/M, and quite honestly it was 3 orders of magnitude better and more stable than the MS-Dos that we had been using which was just a joke. I think the world would have been a much better place if Kildall had not taken that trip. The whole concept of being accepting of buggy software at the MS level would never have materialized. CP/M was rock solid, fast, multitasking and had a far superior set of features and utilities.

_________________
Bill

It is interesting that the article claims that "The 8086 was such a powerful chip that an IBM PC so equipped might convince some customers to choose it in lieu of IBM’s own larger systems; IBM wanted to take business from other PC manufacturers, not from their own other divisions".



Thank you very much for this interesting information. However it seems DR was in no hurry with the CP/M-86, which was only released in November 1981. So it only happened a few months after the launch of the IBM PC.
I can also blame DR that they stopped supporting their most successful OS, CP/M in 1983. More than 20 millions CP/M computers were sold after 1983 and people would definitely have used better CP/M versions but they couldn't.

_________________
my blog about processors

Re: Warning: 2nd hand info .. old guy story ..

That's an interesting story, but I can certainly say it wasn't the story that echoed in the hallways and buildings at the IBM Boca Raton facility. I can say that as I was there... my office was in Bldg 235 and I was the level 2 support specialist for PC hardware and operating systems, which obviously, was DOS at the time.

In a nutshell, the story was one of Dr. Kildall going out for a "fun ride", apparently knowing that the IBM execs were scheduled for meetings that day. The rest of this particular story can be found online in several places. Of course, there are generally two sides to every story, but the significance of this one exceeds that number based on the impact that it did, or would, have had on the industry.

I've also read online that there were concerns about signing the NDA with IBM and his wife (not) signing it (in Dr. Kildall's absence) when the execs arrived. This story I wouldn't believe.... after 38 years with IBM, I never visited a client or business partner in person (usually requiring air travel) unless an NDA was already signed and in place (when an NDA is required, due to the nature of the business). That's general practice in the industry, for all of my time with IBM and two telco startups afterwards.

I kinda doubt the "real story" will ever become generally known, but everyone is entitled to their own view and their own sources.

_________________
Regards, KM
https://github.com/floobydust