(Russian article) Raises a valid question: how do we statistically interpret PERT deviation? Can we use it to estimate confidence level? Author thinks it is too much inaccurate, from pure statistics point of view. Proposes alternative approach (based on 2 point estimations, min & max, and Normal distribution.
(Russian article) Interesting idea for sustaining long projects: while working on bugs/features in a long project, you get to know its problems, deep architectural problems. So, spend some time and work on the "ideal" architecture, the one you'd build from the very beginning if you knew everything you know now. Knowing this "ideal" architecture will help you drive the project in the right direction.
Detailed and well-thought-thru article on brainstorming. But, as I see it, it is only author's opinion on the effective brainstorms. I have seen/participated/led brainstorms which violated some of the rules he proposed, and they still provided good value outcome.