in reply to Re^2: The State of Parallel Computing in perl 2007?
in thread The State of Parallel Computing in perl 2007?

An attractive talk was just posted to Lambda the Ultimate about Erlang and concurrency: LCA2007: Concurrency and Erlang.

⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊

  • Comment on Re^3: The State of Parallel Computing in perl 2007?

Replies are listed 'Best First'.
Re^4: The State of Parallel Computing in perl 2007?
by BrowserUk (Patriarch) on Jan 22, 2007 at 16:55 UTC

    It's a big file for relatively few slides, but they're certainly tantilising aren't they? I don't suppose that you came across any more details of the talks content did you? I followed all the obvious looking links, but didn't find anything that put any flesh on the bones of the slides. I'd love to hear the talk, but that ain't gonna happen any time soon.

    There is an aweful lot to like about Erlang. One thing that I think contributes to it's effectiveness and useability is that it was developed and evolved by non-academics to solve real-world problems with real world, commercial, time and money constraints. As such, it avoids many of the theoretical and 'purity' imperatives that litter academically devloped languages.

    This is especially true with the absence of gargon and notation in the documentation. It's very refreshing to see the purpose and feature sets of an FP language described in terms of the real world problems that they evolved to solved; rather than having the real world treated as a dirty black box that the pure world of FP can encapsulate and control.

    As an aside. It's nice to see that others think like me about that "Computers are state machines; threads are for people who can't do state machines" quote ;) I've tried before to think of an appropriate analogy for that non sequitur and I finally came up with one:

    Cooking is Chemistry; and people without Chemistry degrees shouldn't be allowed to cook.

    Anyone who's seen Heston Blumenthal's TV series will understand that whilst the first clause of that statement is very true, if the second were, mass starvation would not be confined to war and drought ridden areas of the Third World.


    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.

      It took awhile for the slides to download. Mostly I posted that link because it had just appeared as the top item at Lambda the Ultimate and I was also thinking about how much concurrency sucks in Perl. I've been a bytstander when you and others have discussed how to use Perl threads and it's all looked like a big pile of pain. I have a hard time coming up with cases where I'd want to have concurrency in Perl with threads+shared variables rather than with forks+anything external. Blech. At least Parrot and Pugs seem to be addressing these directly. Parrot with it's concurrency PDD and then both with their interest in STM.

      Of course, I'm really novice when it comes to thinking about architecting concurrency. Just this morning by the bus-stop it occurred to me why single-assignment variables (and therefore functional style) go hand-in-hand with concurrency oriented language. It's about getting only "forward" motion which I assume has something to do with having manageable data flow dependencies.

      ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊