I have pointed out that interesting development efforts begin with incomplete knowledge. Going in, you do not have compete information of the requirements, the system design, and how the the software and system will act in the wild. The latter point is especially true in the world of large-scale integration and the use of the software and systems. Incomplete information leads to uncertainty. In fact, non-linear dynamics theory, we know that perfect predictions are impossible even if you have really good information. Otherwise we could have much more certainty in weather prediction.
I am auditing an online course on probability, Statistic 110 by Joe Blitzstein at Harvard, available at ITunes U. He made some really goods, He says in lecture 1, “Math is the logic of certainty, statistics is the logic of uncertainty.” While I agree with the point, at first it struck me as odd as probability is a kind of mathematics.I think he means we can be certain that probability is the way to reason about uncertainty
He also points out that probability is unintuitive for most of us. He cites an instance where Isaac Newton gave the wrong answer to a probability question. For our field to mature, we collectively need to have better intuitions about uncertainty. In fact, fields we should understand like big data and cognitive computing are applications of applied probability.
The challenge for us is that most of us were not trained well-enough in probability to have developed good intuitions on how to apply probability. I partially addressed this in a paper I wrote, http://www.ibm.com/developerworks/rational/library/filling-in-the-blanks-1/. Of course, this course is far superior than just reading the paper and I highly recommend it to you and your friends.
Leave A Comment