in reply to Re^2: Time to write a "serious" http server in Perl?
in thread Time to write a "serious" http server in Perl?

Considering that most (Unix) OSses are written in C, it would be irrelevant for a language to scale better than C ever could. Because if such a statement is true, it wouldn't help whatever application was written in Erlang, it couldn't display its superiour scalability - the limitations of the OS kick in.
  • Comment on Re^3: Time to write a "serious" http server in Perl?

Replies are listed 'Best First'.
Re^4: Time to write a "serious" http server in Perl?
by chromatic (Archbishop) on Aug 13, 2008 at 19:35 UTC
    Considering that most (Unix) OSses are written in C, it would be irrelevant for a language to scale better than C ever could.

    Nonsense. A Unix-like OS is an opinionated, general purpose operating system intended to run a wide variety of applications equally well. A well-designed application tuned for a specific purpose written in an appropriate language may outperform a generic application written in C. Go look at the language used for the best FFT libraries on your system; I bet there's Fortran involved. Then look at some of the compile-time optimizations you get in languages where explicit or even implicit purity by default allows for automatic memoization and expression elimination. Throw in automatic parallelization, and you can beat C in terms of performance, scalability, and ease of development.

Re^4: Time to write a "serious" http server in Perl?
by dragonchild (Archbishop) on Aug 13, 2008 at 23:16 UTC
    If you're discussing a single machine, sure. However, Erlang's major claim to fame is that it scales (nearly) effortlessly across a cluster of machines and does so for (nearly) anything written in it. And it does so with fault tolerance, load balancing, and failover built into the VM. Given that I haven't built a production system that had only one machine for any given part of the application stack, this sounds rather appealing, don't you think?

    My criteria for good software:
    1. Does it work?
    2. Can someone else come in, make a change, and be reasonably certain no bugs were introduced?