core-logic clock / was Re: TTL homebrew CPUs

[Jim Battle]

William Donzelli wrote:
>>  Dr. An
>> Wang (the inventor of magnetic core memory) did a lot of experimentation
>> with core logic, and had patented it.  But, IBM offered him big bucks to
>> assume the patent, and he took the money.  Later, Dr. Wang realized that
>> he got ripped off by IBM,
> 
> That is now how IBM felt. They paid thru the nose for the license. The
> also spent a large amount of money trying to get around the license
> before they caved.

Chapter 7 of "IBM's Early Computers" by Bashe, Johnson, Palmer, and Pugh 
is all about IBM's experience in developing their core memory expertise 
and getting it into production, along with some of the patent intrigue, 
core memory history, and avenues that were explored and abandoned.  This 
is all written from IBM's perspective.

After reading this chapter, it disabused me of the meme that Wang 
invented core memory.  Others had proposed using magnetic rings before 
Wang.  IBM even had done some experimenting with them before hearing of 
Wang's results at Harvard, which were far more advanced than what IBM 
had done to that point.

Munro Haynes read Wang's paper and started his own experiments at the 
University of Illinois.  It was Hayes who figured out coincident current 
switching, which makes addressing the cores a whole lot more pratical. 
Haynes later joined IBM.

MIT too worked on practical core memory, and there was some back and 
forth with IBM about it.

I won't replay the whole chapter, but the main points are two:

	(1) like many inventions, there wasn't a single father,
	    rather a concurrent development by many parties that
	    were sometimes aware of the work of the others,
	    sometimes not

	(2) IBM didn't pay through the nose for Wang's patent;
	    they got Wang's patent for "several hundred thousand
	    dollars" according to the book.

	    IBM payed through the nose (in their estimation)
	    for rights to core memory technology patented
	    by Jay Forrester at MIT (through MIT's patent
	    management company).  MIT was insisting on royalties
	    of two cents per bit (core).  Their tussle went on
	    for years, and in the end IBM paid out a one time
	    $13M fee for the rights to the Forrester patent.
	    By that time (1964), IBM was using more than one
	    billion cores per year.

The chapter has a lot of other really interesting detail and is worth 
the read.