🚀 go-pugleaf

>| It's funny that many DOS users don't know that their beloved OS is based
>| heavily on CPM. In fact Ver 1.0 of MS-DOS looked like a recompiled CPM
>| with a few modifications.
>
>Believe it or not, MS-DOS actually WAS a recompiled version of CP/M.. !! 
>Bill Gates stole the CP/M code (bought it cheap, re-compiled, and sold to 
>IBM with copyright, etc. to himself... )  So CP/M lives even in '95, eh? :)


Here is a quote from Dettmann and Johnson, "DOS Programmer's 
Reference"  (QUE, 1992), p11:

"DOS was first marketed by Seattle Computer Products as 86-DOS for 
that company's line of computers.  The original DOS was written by Tim 
Patterson starting in April 1980, and first shipments were made in 
August of that year.  At that time, Digital Research's CP/M was the 
most widely used microcomputer operating system.  86-DOS was 
specifically designed to make porting applications from CP/M easy: it 
kept the same structure for file control blocks [not widely used  
since MS-DOS ver. 1] and functions so that a mechanical translator 
could convert a program directly to 86-DOS... Seattle Computer 
Products established a base of several dozen customers, including at 
least one other hardware manufacturer.  Then Microsoft approached SCP 
about writing a customized version for an anonymous customer... By 
January 1981, Patterson knew who the customer was... in April of that 
year, Patterson left Seattle Computer Products and joined Microsoft, 
where he spent the next several months tailoring the system to IBM's 
needs... In July 1981, Microsoft bought from Seattle Computer Products 
all rights to 86-DOS." 

"After the PC's original release, DOS was still not prominently 
displayed in some stores.  IBM had selected CP/M-86 and Softech's
p-system as alternative PC operating systems.  But vendors were slow 
to deliver both products..."

I have another MS-DOS reference that claims the original, internal 
name for 86-DOS was QDOS, which ostensibly stood for "Quick and Dirty 
Operating System." 

Remember that back in those days, most programs were written in 
assembly language.  The 8086/8088 maintained a great deal of downward 
compatibility with the 8080 (look at vestiges like SAHF, which proved 
useful with the 8087), so "mechanical translation" was possible as 
long as the CP/M OS calls mapped into functional equivalents in
86-DOS.  The 86-DOS functions were not *stolen* from CP/M; they were 
just designed to use similar interrupts, and take the same input and 
return the same results.  This correspondence was easy, since the CP/M 
interface was widely published for programmers writing CP/M 
applications.