Is Ungar focusing on general-purpose computing or just high-performance computing?
Unless he's strictly talking about HPC, he could be way off the mark. For the past 5-10 years there's been a general assumption that massive parallelism will be necessary as CPU speeds max out. But then there's talk that something like graphene could take us into the terahertz range. And isn't 4 ghz enough? My first computer was 4 mhz. Bloatware still pushes hardware's limits, but in my own code, performance just isn't a big obstacle anymore. Even if there does turn out to be a simple and general way to do parallel programming, there'll always be tradeoffs weighing against it - energy usage and design complexity, to name two obvious ones. BTW, good point about the problem of hyper-specialization in CS academia. -Tom On Thu, Mar 29, 2012 at 10:12 PM, David Barbour <dmbarb...@gmail.com> wrote: > > > On Thu, Mar 29, 2012 at 12:29 PM, Max Orhai <max.or...@gmail.com> wrote: > >> Probability is highly applicable to (bounded) nondeterminism, but I get >> the impression that most CS theorists don't tend to learn much about it, >> and I know for sure that it gets extremely short shrift in the applied CS >> curriculum at my school. >> > > How would you suggest applying probability to non-determinism? What > benefit would it provide? > > I've seen some cases where probability is applied to non-determinism, e.g. > concurrent-constraint probability models. But the benefits it offers seem > marginal. > > > >> >> Dave Ungar loves being deliberately provocative, but I really don't >> understand why he's so attached to the (obviously unscalable) shared memory >> imperative programming model... except, perhaps, he thinks that's the only >> model the great unwashed masses of industry coders can handle. If so, I >> sure hope he's wrong. >> > > I also hope he's wrong. After all, the unwashed masses barely handle even > that. Even if we exclude concurrency, most bugs and complexity involve > state. > > >> >> But, lets face it, after decades of real-world deployment, Erlang is >> still considered an exotic language, and hardly anybody outside the ivory >> towers has even heard of Kahn nets, FRP, CALM, etc. These don't get taught >> in the undergrad CS curriculum either. >> > > Most professors don't know them either. Doctorates in CS are generally for > a specialized contribution, not broad knowledge. It's unfortunate. > > >> >> Programmers, like everybody else, only get to choose their problems >> inasmuch as they are aware of the choices. >> > > And the only way to make them aware is to provide a killer app to extend... > > Regards, > > Dave > > -- > bringing s-words to a pen fight > > _______________________________________________ > fonc mailing list > fonc@vpri.org > http://vpri.org/mailman/listinfo/fonc > >
_______________________________________________ fonc mailing list fonc@vpri.org http://vpri.org/mailman/listinfo/fonc
