How point-contact transistor work?

In December 1947 Bardeen, Brattain and Shockey introduced a device which is considered to be the first transistor. This transistor was point-contact device. So, it used metal-semiconductor junctions.
According to book History of semiconductor engineering by Bo Lojek (2010), a principle of such transistor has never been fully understood:

…but as matter of fact, a good solid physical theory of point-contact transistor does not exist and even today we do not know, for example, why point-contact transistors occasionally exhibit negative resistance behavior.

As to produce point-contact transistor (or diode) is rather art (because each piece of semiconductor can have different impurities, damaged crystal grid etc. and it also depends where a metal contact is made), it seems obvious that each transistor could show different behavior (mentioned negative resistance, unstable gain etc.). These were reasons why Shockley worked out theory of P-N junctions and based on it introduced junction transistor which behavior is predictable.

I understand that point-contact transistor is now very obsolete device, however, I wonder if there is (or was) any research trying to understand point-contact transistors functions. Any information, paper etc. will be highly appreciated.

Cross-posted on SE.Electrical Engineering