Comments by "Lawrence D’Oliveiro" (@lawrencedoliveiro9104) on "The Computer Chronicles - Gary Kildall Special (1995)" video.

  1. If Kildall had lived just a few years longer, he would likely have invested in Linux.
    5
  2. 6:25 I feel there was a missed opportunity there. As a minicomputer guy at the time (DEC PDP-11, then VAX), the one revolutionary new thing that the micros brought to the table was being able to access the video memory directly from the CPU. This mean instant full-screen updates, allowing a level of interactivity beyond anything a bigger and more powerful mini could manage by accessing text terminals through anemic serial lines. But there was no hardware-independent way of accessing the video RAM. And when the display hardware became capable of graphics, this hardware dependence grew even worse. Imagine if CP/M had added another layer above the BDOS—call it the “BVOS”, the “Basic Video Output System”—to abstract away the details of the display hardware, allowing full-screen and even graphical apps to be portable across different machines, they could have stolen a big march on the Apple Macintosh.
    4
  3. 3:20 Near as I can tell, Kildall’s PL/M compiler was always a cross-compiler: it never ran native on any 8-bit microprocessor. To use it, you had to have access to a big (and expensive) DEC PDP-10 machine. I think this ran the TOPS-10 OS, which was a pretty typical DEC OS. A filespec consisted of a device name, a directory part, a name part, an extension and a version number. Clearly this was the inspiration for CP/M filespec syntax: the device name (which could be multiple characters) was simplified down to a single character, the directory part and version number were omitted, the name limited to 8 characters, and the extension to 3. And aspects of that syntax were carried over to MS-DOS, and still persist in Windows to this day. Particularly the single-character drive names, which seem pretty ridiculous on modern hardware.
    1
  4. CP/M only ran on Z80 and 8080/8085 chips. There were other popular microprocessors out there (like the 6800 and 6502), which it never supported. At this point, Unix was already available on machines with entirely different hardware architectures and instruction sets. So you could say that Unix portability was aiming for a much more ambitious target.
    1
  5. I think he would have seen the benefit in supporting Linux, too. Even then it would have been clear that this was the only possibility for a real threat to Microsoft.
    1
  6. 23:05 There is room for multiple companies to compete where you have common standards that avoid vendor lock-in. Bill Gates built his business on vendor lock-in.
    1
  7.  @OpenGL4ever  386, not 286. The 286 was a stupid chip that only IBM, with its bloated OS/2 offering, cared about.
    1
  8.  @OpenGL4ever  That success of Windows 3.0 actually started with its predecessor, Windows 2.1 a.k.a “Windows 386”. The name should give you a clue.
    1
  9.  @OpenGL4ever  Just do a search for “Windows 386”. I was there.
    1
  10. Capitalism is efficient when there is a level playing field without vendor lock-in. Unfortunately, the history of the computer history is full of instances of vendor lock-in, which has reduced the efficiency of market competition. On the bright side, most of these cases do go away eventually, so they mostly slow down innovation rather then prevent it altogether.
    1
  11. 7:00 It did draw some foundation from the minicomputer industry, though, particularly DEC machines. Natural, since CP/M was originally cross-developed on DEC machines.
    1