Social Processes and Proofs of Theorem and Programs (Lipton, Perlis, DeMillo)

I have to admit that I included this article into my top ten list of articles read at MIT this semester.

“If mathematical process were really one of strict, logical progression, we would still be counting on our fingers.”

The article opened me a small window into the world of mathematicians. Our distorted understanding of the their world lead us to erroneous perception of how software programming “world” operate.  In both of these worlds, social processes, which are not necessarily obvious to an outsider, are major driver in test and acceptance of new ideas, theorems, products and systems.

The authors argue that "A program is a human artifact, a real-life program is a complex human artifact; and any human artifact of sufficient size and complexity is imperfect.” Thus people are not able to “create perfect mechanisms and   verification process is nothing but a model of believability”.  As the result, software programs have to be reliable but not necessarily perfect.

A conclusion that can be drawn from the article is that in software engineering social processes have to more visible and acceptable form of test and acceptance. By social processes I mean per-reviews, beta-user reviews and end-user reviews. These processes may not completely verify the system, but encounter the system with the real-world environment and force the designers of the system to adopt it.