In general, I do not think SF made a positive contribution to computerchess.
The ultra fast testing procedures, the level of pruning were contributions of V. Rajlich, and himself was inspired by Fabien Letouzey.
The tuning procedures were a contribution of several authors, among these Peter Österlund, Thomas Petzke, etc. I am sure there are many others.
The use of neural nets are an idea of Matthew Lai, implemented thereafter by Youri Matiounine and Jonathan Rosenthal.
Basically, although I am sure programmers know it much better, I do not see what could be revolutionary in SF.