14 May
2013
14 May
'13
7:44 a.m.
On 5/14/13 5:40 AM, David Riley wrote:
On May 13, 2013, at 9:06 PM, Dave McGuire <mcguire
at neurotica.com> wrote:
There was the issue of not wanting review from the internal 'competition', but
mainly they
had a religion that you ONLY were to know about the published interfaces since
they felt that most of the problems with the Mac OS was how exposed the implementation
was (low memory globals, etc.).
I know the person who wrote the TCP stack. Smart guy, but he REALLY preferred
Mesa to C++. He implemented TCP/IP for the Xerox Star.
Also, when I said Apple was happy when 'they' went off to Taligent, I meant the
Pink group, and not Phil and Erich. Phil left for General Magic, then was one of the
founders of WebTV, Erich went to Amazon a bit later.
On 05/13/2013 08:57 PM, Al Kossow wrote:
Jeeze. Were they THAT embarrassed about it? ;)
Alas, they probably thought it was THAT great (and, I guess, easy to
copy if you caught a glimpse?).
I
remember a friend of mine at IBM took the Taligent source code (which
was all C++) and ran it through a program he wrote to graph the object
relationships.
At least IBM got to see it. Internally NO ONE outside of the Pink group
was allowed to look at it. We had two people who were cleared to look at
it (Phil Goldman, and Erich Ringwald who both were original members of the
Pink microkernel team that came to RISC Products) and they had to go over
there in a locked room with no network and no removable storage on the
machines. Now, imagine trying to write drivers when you can't look at the
kernel code, and there is no published driver API.