8 hours of TCL (and not so much of Tk)
Our guys had invited some Altair guys to teach us some specific topics in Hypermesh (henceforth abbreviated as HM). Today’s topic was automation and customisation. Since I do a lot of the former, it was a great way to introduce myself to the methods of HM in this regard. The API of HM is exposed in TCL. I knew that already. What I didn’t know was TCL.
The trainer from Altair had only 30 odd slides to show us every aspect of TCL programming. His lectures were more inclined towards scraping the TCL surface for most users. I learned something new on almost every slide. This was not because he was explaining something radical but because my exposure to the LISP family and C family of languages kind of forced me to decipher (or attempt to understand) the mechanism of the syntax.
It turns out that there are only 12 cardinal rules to program in TCL. And here they are . I wish he had started with that.
The strange part is that Altair has decided to move away from TCL in favour of Python. Python is a hip language at the moment. I wonder why they decided to use TCL in the first place. Or what would happen to the legacy solutions they have implemented in the industry?
Another hour of Machine Learning
I had registered myself on the Stanford University’s Machine Learning course  on Coursera. It’s conducted by none other than the founder of Coursera himself . Because I was in a good flow today, I decided to complete the Week 1 materials.
I like something about the way courses are designed by the US universities. They don’t jam a course with umpteen pre-requisites. For example, knowledge of calculus would help one grasp the mathematical aspect of the course better but at no point does the instructor say that it would be a barrier in understanding the course.
Sudipto  (my childhood friend, singer of Dark Project) works on OpenStack and he says that there are major contributors to this open source project, who have background in liberal and performing arts. Go figure!