26.10.05
SPLS, 25 Oct
The Scottish Programming Language Seminar goes from strength to strength. We consistently have thirty to forty researchers coming from across Scotland, and the range and quality of talks is excellent. All of this with very lightweight organization and coordination.
About thirty souls turned up at Stathclyde, coming from Glasgow, Edinburgh, and St Andrews, and visitors from as far away as Australia. There were five talks, ranging from semantic theory to hardware design. Alex Simpson described an elegant (and beautiful and surprising) way to extend Reynold's semantic parametricity to Moggi's computational lambda calculus. Wim Vanderbauwhede presented a Scheme-like language for programming "Systems on a chip" consisting of hardware modules interconnected by a chip-wide network. Greg Michaelson introduced Hume, an attempt to put fp to work on real applications producing embedded systems with guarantees on the time and space used; they are applying the system to real vision problems in automated vehicles. Joe Wells provided a clear introduction to a calculus for linking, with strong technical results (can be analyzed in close to quadratic time, where the first analysis proposed in this area was NP complete). I liked that he focussed on the simplest possible calculus that could express the problem. Edwin Brady explored how a dependently typed language similar to Epigram can express resource bounds; Epigram is beginning to look familiar to me and I quite like it.
Many thanks to Strathclyde and David Lievens for organizing the meeting. The next will be at St Andrews in January 2006.
About thirty souls turned up at Stathclyde, coming from Glasgow, Edinburgh, and St Andrews, and visitors from as far away as Australia. There were five talks, ranging from semantic theory to hardware design. Alex Simpson described an elegant (and beautiful and surprising) way to extend Reynold's semantic parametricity to Moggi's computational lambda calculus. Wim Vanderbauwhede presented a Scheme-like language for programming "Systems on a chip" consisting of hardware modules interconnected by a chip-wide network. Greg Michaelson introduced Hume, an attempt to put fp to work on real applications producing embedded systems with guarantees on the time and space used; they are applying the system to real vision problems in automated vehicles. Joe Wells provided a clear introduction to a calculus for linking, with strong technical results (can be analyzed in close to quadratic time, where the first analysis proposed in this area was NP complete). I liked that he focussed on the simplest possible calculus that could express the problem. Edwin Brady explored how a dependently typed language similar to Epigram can express resource bounds; Epigram is beginning to look familiar to me and I quite like it.
Many thanks to Strathclyde and David Lievens for organizing the meeting. The next will be at St Andrews in January 2006.
