User loginNavigation |
A functional-programming view of timeSee fig. 3.38 in SICP, where the authors show a model of a joint bank account shared by two people, Paul and Peter. The bank account can be modeled as a process that operates on a stream of transactions and generates a stream of responses. But the difficulty is merging the individual streams of transactions by Paul and Peter in a purely functional style. They argue that one can't get away from explicit synchronization. They end this this chapter by pointing out the essential difference between the two views of modeling of the world: as a set of interacting stateful objects vs a single timeless stateless entity. They say "A grand unification has yet to emerge". IIRC this part has not changed between the first and second edition of SICP. [This is probably confusing so you might wish to read this last section of chapter 3 in the online copy of SICP, or view the very excellent video lecture 6B from MIT opencourseware @ about 52 minutes. The Q&A at the end is also relevant] I am wondering if things have changed since then. I doubt FPLs can allow modeling such a merge but how far have we come? Can we express synchronization in some way in FPLs? Is there a grand unification of something like CSP and FPL? Are SICP authors looking at this the wrong way? By Bakul Shah at 2010-10-08 18:37 | LtU Forum | previous forum topic | next forum topic | other blogs | 9922 reads
|
Browse archives
Active forum topics |
Recent comments
25 weeks 4 days ago
25 weeks 4 days ago
25 weeks 4 days ago
47 weeks 5 days ago
52 weeks 16 hours ago
1 year 1 week ago
1 year 1 week ago
1 year 4 weeks ago
1 year 8 weeks ago
1 year 8 weeks ago