Andrei's Google Talk

[Walter Bright]

Russel Winder wrote:
> I am sure Walter reads other code, and especially learns about languages
> other than C, C++ and D -- not to do so would be to abdicate one's
> professional responsibility to stay up to date in one's field.  The
> problem is to set up a clear barrier to the accusation of copying.  So
> to take a well known example:  FOSS project contributors should never
> read Microsoft copyright code, and Microsoft employees should never read
> GPL and LGPL code -- substitute any profit-making company that relies on
> licence revenues for Microsoft, it is just that using Microsoft here is
> almost certainly not libellous.   Rigidly following behaviour such as
> this is not a total defence in a copyright action, but it sets a pattern
> of corporate and individual behaviour that makes a copyright
> infringement less likely and more difficult to prove sufficiently to win
> an action.

Back in the 80's, making a clone PC required copying the BIOS, or at least 
people assumed it would. This put a huge barrier in place, as IBM had a 
reputation of aggressively suing anyone who might infringe.

This nut was eventually cracked by Phoenix, who set up 3 groups:

Group 1. read the IBM PC BIOS, and wrote a specification for it.

Group 2. implemented the specification, and WAS NOT ALLOWED TO EVER LOOK AT THE 
IBM PC BIOS CODE.

Group 3. a group of lawyers who read (1) and (2) and verified there was no copying


No communication between G1 and G2 was allowed that did not go through G3. They 
were kept physically separated.

Extensive documentation was made of every step in this process. It left 
absolutely no crack for IBM to drive a lawsuit through, and it worked. The 
market for clones was blown wide open, and Phoenix made a fortune supplying the 
BIOSes for them.