Re: OK, let's sort this out...
Warning: Bloody long post. With only one tiny, irrelevant provider reference:
Originally posted by Humble Narrator
In order of creation
<snip>
PC-DOS: IBM's very own version of DOS that was NOT licensed from MS, but was not "100%" compliant with MS-DOS, thereby starting MS' tactic of becoming the de facto standard and then immediately adding their own proprietary elements making it virtually impossible to replicate in near-real time. This tactic has served them very well up to this day.
DR-DOS: Kildall's entry (late) to the party that was always a superior product, but again, with MS always shifting the target and marketing the hell of it, they never caught. Oh, and maybe the bullying of hardware vendors might have had some impact as well. ;-)
FYI, Gates suggested that IBM contact Digital Research, but they never did. Instead, they came back to him for a solution, which he bought from SCP for $50k! [/B]
[Background info: I worked for DRI at the time they got eaten by Novell (who later, as was common for that company, regurgitated them)]. That was much after all of the following happened.
PC-DOS
was based on MS-DOS, with absolutely trivial differences. However, at the time that IBM was selling PC-DOS, it was also selling PSYS and CPM/86.
DR-DOS was very loosely based on the original CPM/86 code, but little of CPM/86 still existed in the end product (which you can verify... it's now open source), and absolutely _NONE_ of Kildall's code still existed (Languages, platforms, etc). Just to chain up the very beginning of all of this, Gary Kildall was at Monterey's (Uh-oh: Ambiguity... it was either DLI or NPG... I'll bet hard on NPG, though) Naval Post-grad school, and he was contracted by Intel (who?) to write a P/LM compiler for something called an 8080 (which no one ever really used for long, switching to Zilog's Z80 instead... PCHL). To test his compiler, he did what any neurotic geek would do: He wrote an operating system (well, the joke was that he wrote half of an OS, and left the other half as an exercise for the student). The File system was kind of fun for the times (1975... I was just a boy)
CP/M was quite successful, as it basically had no competitors (no one really counts NDOS [Northstar DOS], PSYS [IBM liked it, but no one else did], or CROMIX (A UNIX clone running in that small a box was .... amusing [CROMIX was sold by CROMEMCO - or Crowther Memorial Commons, a Stanford Dining hall]).
Bill Gates thought "Hey, wouldn't it be cool to have a BASIC interpreter for this OS?" He wrote it (in a hotel room in Phoenix(?) with Paul Allen. ObProviderRef: Who else was there?). It was a truly great hack, even if one never used it. He then started a small company to sell it. Then he wrote a very bad linker and assembler (MASM-80.) Then he wrote a bunch of other tools. Not great. Not bad.
Years pass. IBM decides to do a PC. Makes lots of mistakes (8088? How brain dead could you get? And what gives with that half of a bus? On the other hand, you could already get tools and platforms for an S-100) Goes to Bill to buy development tools and (most importantly) BASIC. Puts newly purchased BASIC in ROM of PC (therein previewing what would be the differences between PC-DOS (until 2.0) and MS-DOS... MS-DOS didn't use, rely on, or acknowledge the BASIC in ROM.)
What follows is the product of conjecture and fact... Kind of like an "Historical Novel". IBM then says to our friend Bill: "Now, what do you think about an OS?"
Bill says, you mean, you've not talked to DRI yet? (Note that it really was DRI at this point [ Digital Research, Inc ], the company having changed from InterGalactic Digital Research, Inc with the non-release of CP/M 0.86.) So, [Fact] IBM makes an appointment to talk with Gary Kildall very soon. [/Fact] Rumor has it: I.) He (Gary) was too stoned to make the appointment; II.) He decided he wanted to go fly his new toy (and I wish to hell I could remember what it was... Beechcraft Baron?) III. Demons attacked Pacific Grove's Butterflies, and he had to fight off the demons and return the beautiful princess to her castle. Whatever.
IBM calls up Bill, and says "Hmmmph. We still need an OS. We'll see you tomorrow" Now, Microsoft is still a tiny company. Can't tell you how many employee's they had, but it wasn't a campus, it wasn't a road, and manuals were done by daisywheels (that is not sexual innuendo) and bound in loose-leaf binders (ugly brown ones). I again am not sure, but I believe that most of the resources of the company were working on the CP/M card for Apple's. Microsoft was absolutely dwarfed by companies like MicroPro (WordStar), Ashton-Tate (Vulcan first, then dBase), and Software Arts (Visicalc).
Bill and MS, being small, move quickly. He calls up SCP, hacks together a business deal to buy rights to qdos, (and later ignores some of the provisions of that deal, for which they did indeed got caught), and completely changes the direction of the company. Overnight. SCP is a
retail operation which had hacked together an OS (sort of) for the 8086 S-100 boards, called qdos [Quick and Dirty Operating System]. It was
not CP/M compatible: You did INT 21 instead of CALL 05 for syscalls... to say the least (the CP/M emulation jump was not in early versions of qdos, but I don't know if SCP or MicroSoft put it in). Also, there were no user areas on a disk (user areas were kind of like directories, except that there were exactly 16 of them, and they were called 0: ... 15: Oh, also, they were essentially pointless, but M/PM (Multi-user CP/M) needed them for protection, so CP/M got them later at 2.1). qdos and ms-dos had no PIP, no ERA, etc. Finally, and this is the big one, the File system for qdos and ms-dos was FAT based.
When IBM comes back to Microsoft, MS sells a snow job (" Why, yes we DO have an OS, although still under development, and this is what it looks like").
Someplace around here, I may still have a file comparing all of the versions of MS-DOS, but, other than 2.0 (which added the unix-y stuff like file handles and hierarchical directories), no-one really cares.
Later, IBM wanted to go with OS/2 (which Microsoft used for development of Windows from Win/386 onward), and Microsoft liked Windows, which they'd started developing in 83 or so. Neither Window's nor OS/2 had hit the press as anything other than vapor, but MS and IBM decided to hate each other. At this point, MS was the underdog, BigBlue the scary monster, and tech people rooted for MS, wsb.
DRI tried to save itself years later by enhancing CPM/86 (which had, again, been sold with the original IBM 5150. It cost $500. psys cost $300. PC-DOS cost $100, and PFS Perfect Writer came with it) by adding in a MS-DOS compatibility layer. Quite a nice one. They also worked for a while on a windows clone (stay away from Int 2A (or 2F?). Microsoft did some nasty stuff to kill the commercial prospects of all of those.
Stuff I don't know: Sorcim wrote SuperCalc. Then Computer Associates sold SuperCalc. Was that the beginning of CA?
With all this utterly useless knowledge swirling around my head, is it any wonder I have to see providers? Who else would listen to a really old guy?
[Edited by beep9 on 07-24-2001 at 10:41 PM]