I found this talk extremely interesting. Of note, he describes: How they're making syntax close to that of mathematical notation. How a huge number of program

I found this talk extremely interesting. Of note, he describes: How they're making syntax close to that of mathematical notation. How a huge number of programming statements can be modeled by a generator, a comprehension, and a reducer. How they're using traits, a variant of multiple inheritance, to encode and test for algebraic properties. How, knowing these algebraic properties, and knowing specifics of the hardware system, generators, comprehensions and reducers can derive a parallelization strategy. I'm excited enough by this language I've actually started to use it, for numerical work. It's not mature -- there's no compiler, just an interpreter -- but the input method is easier than latex, so at worst I'll have well typeset pseudocode. :-)