17 Oct
2011
17 Oct
'11
1:59 p.m.
On 17 Oct 2011 at 13:03, Eric Smith wrote:
Incorrect. Those were open source until the great
unbundling of June
23, 1969. In fact, it went beyond just "open source"; they were in
the public domain, along with the operating systems they ran on (e.g.,
OS/360 21.8, DOS/360 and TOS/360 26.2, MVS 3.8, TSS/370 3.0).
There was a somewhat different attitude back before the days of
"commodity CPUs". If you, as a customer, had an interest in
modifying or adapting any system code, you could, with a little
wrangling perhaps, get a copy of the source code. Site analysts
routinely suggested corrective fixes to code.
The system code was a necessary evil to sell the hardware.
And the picture was often mutual--customers contributed whole
products that became part of the standard line. In at least one case
that I'm familiar with, the whole operating system for a machine was
written by a customer.
I know that IBM was irked by systems such as the RCA Spectrola or one
of the Two-Pi boxes that could run IBM object code, but I don't know
if IBM ever filed suit.
It was a very good idea to have the capability to run competitor's
customer source code without too much tweaking, as well as being able
to read their tapes.
When AT&T essentially offered Unix for free, it was no big deal. The
big deal was that it was written in C and if you could get a C
compiler going for your CPU (or at least a cross-compler), you
*might* stand a chance of getting a functional OS up quickly.
But heck, this isn't the first case of an OS being written in an HLL.
That probably goes to the system software for the B5000 series
(extended Algol).
During the 70s, CDC even made a move away from assembly to a dialect
of Pascal called SYMPL and other vendors did likewise.
--Chuck