Well, now I've gotten a week or so into Stanford's Machine Learning course. The second week of the course is all about learning the Octave programming language. In my previous post I described why the professor, Andrew Ng, says he chose the Octave language to teach the course.
It turns out that Octave is essentially an open-source derivation of Matlab, which is nice for students since Octave is free, and commercial versions of Matlab run several hundred dollars, and even the student version of Matlab is $100. As I work through the Octave examples, I am struck by the similarities to J and APL. This means that the Matlab language must also be similar to J/APL, though I have never used it. Generally, the syntax of Octave is a bit more convoluted than J and the underlying concepts of Octave/Matlab aren't quite as elegant as J, but much of the functionality of Octave is similar to the basic ideas in J/APL. This is particularly true with the idea of extending scalar functions to vectors and matrices. I will be working through the assignments over the next few days, so I will get more insight into the similarities. I'll give another report in a few days. Skip -- Skip Cave Cave Consulting LLC ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
