Monday, February 11, 2008


Busy times, but interesting ones:

  • Reading "Synchronous Kahn Networks": relating efficient synchronous computation and listlessness optimizations by presenting streams as lazy lists following a clock calculus. Addition of dependent types to the clock calculus is embarrassingly impeding my progress, but a good motivator to get a better feel for it. End goal: "N-Synchronous Kahn Networks" follow-up paper.
  • Reading "The Categorical Abstract Machine": slide 4 describes the goal here - compiling the original basis of ML in a rigorous way. End goal: points! Or knowledge.
  • Wrote the final technical part of my paper, so filler, an example, and heavy editing are all that's left. Oh, and submitting it somewhere.
  • The more I try to answer what a browser and web program are, the less I can. Tricky to make a definition that will hold in 5-15 years. Wikipedia's definition of a browser is sufficiently vague and conservative to squeak by, but I do not buy its one for web applications.
  • Coming enlightenment by Bill Thies (streaming computation? developing region and assistive tech? race detection? programmable microfluids?), as well as Burak Emir (pattern based extraction support in Scala, Done Right)
Also, something about camping in Death Valley, least RE set encodings for lambda calculus, and running something ginormous on the similarly ginormous NERSC cluster through LBL. For the latter - what will it be? A type miner, a concolic tester, a browser bug finder, ... oh boy oh boy. Maybe I should give Dawn a poke, I bet she could burn a few cycles in the name of Good :)

Ah, and congrats to Chang-Seo for rocking his prelims. He is that much closer to being a free man.

No comments: