Loading ...
Sorry, an error occurred while loading the content.
 

Advocacy and Talking Points

Expand Messages
  • Fred Moyer
    Hi, Yesterday I was asked about the viability of mod_perl ( and more generally Perl web based solutions ) in scaling web based applications. I have my own
    Message 1 of 3 , Sep 28, 2005
      Hi,

      Yesterday I was asked about the viability of mod_perl ( and more
      generally Perl web based solutions ) in scaling web based applications.
      I have my own personal experiences with scaling mod_perl, and am well
      versed with the best practices, case studies, and methodologies out
      there in scaling mod_perl web apps.

      So I was able to cite a few large sites from memory which used mod_perl,
      and also in hopes of bolstering their confidence, I said something to
      the effect that I would wager my career that mod_perl could scale with
      the best of the other technologies out there, and I meant it. I cited a
      couple of links [1] post conversation, and blogged about this today on
      use.perl [2].

      So in wanting to improve upon my advocacy pitch the next time this
      question is asked of me, I would like to ask the list for advice on
      talking points. And also experiences with a situation like this when
      you are asked to compare mod_perl's scalability to some other technology
      in an apples vs oranges comparison. I've looked through the list
      archives and didn't find much in this area, if anyone has a link to
      relevant content please send it my way.

      Many Thanks,

      Fred

      [1]
      http://perl.apache.org/outstanding/sites.html
      http://www.masonhq.com/?MasonPoweredSites

      [2]
      http://use.perl.org/~Phred/journal/26908
    • Bill Whillers
      I m an advocate. But this argument can be broken into at least a few areas or more (beyond the walls of mod_perl) since scaling applies to human as well as
      Message 2 of 3 , Sep 28, 2005
        I'm an advocate.

        But this argument can be broken into at least a few areas or more (beyond the
        "walls" of mod_perl) since scaling applies to human as well as hardware and
        capital infrastructures.

        This may be old news and common, but I hear less of performance scaling
        criticisms than of what some (from the php and python worlds) might see as a
        'free-for-all', non-uniform coding paradigm (where primarily examples always
        seem to live outside the mod_perl "citadel").

        <digression>
        But then, that's probably why I'm here to begin with... I'm a creative
        type and I think perl and it's gigantic community fit many fast moving
        business and artistic cases and requirements very well. Good democracy's may
        have access to many freedoms, but simply imposing their own standards as
        necessary at their own discretion.
        </digression>

        But to be honest, I've not been digging for recent, technical criticisms
        lately from the outside.

        So solid talking points should address maintainability as much as performance;
        inextricably linked as they may be. In fact, I believe mod_perl (and it's
        prudence's) does seem to help address many of those persistent
        maintainability arguments and in convincing ways... and this is from my own
        experience.




        On Wednesday 28 September 2005 17:27, Fred Moyer wrote:
        > Hi,
        >
        > Yesterday I was asked about the viability of mod_perl ( and more
        > generally Perl web based solutions ) in scaling web based applications.
        > I have my own personal experiences with scaling mod_perl, and am well
        > versed with the best practices, case studies, and methodologies out
        > there in scaling mod_perl web apps.
        >
        > So I was able to cite a few large sites from memory which used mod_perl,
        > and also in hopes of bolstering their confidence, I said something to
        > the effect that I would wager my career that mod_perl could scale with
        > the best of the other technologies out there, and I meant it. I cited a
        > couple of links [1] post conversation, and blogged about this today on
        > use.perl [2].
        >
        > So in wanting to improve upon my advocacy pitch the next time this
        > question is asked of me, I would like to ask the list for advice on
        > talking points. And also experiences with a situation like this when
        > you are asked to compare mod_perl's scalability to some other technology
        > in an apples vs oranges comparison. I've looked through the list
        > archives and didn't find much in this area, if anyone has a link to
        > relevant content please send it my way.
        >
        > Many Thanks,
        >
        > Fred
        >
        > [1]
        > http://perl.apache.org/outstanding/sites.html
        > http://www.masonhq.com/?MasonPoweredSites
        >
        > [2]
        > http://use.perl.org/~Phred/journal/26908
      • Steven Lembark
        -- Fred Moyer ... Because it s perl, not a substitute :-) One thing that means is that all of the Test::* modules are avaiable
        Message 3 of 3 , Oct 23, 2005
          -- Fred Moyer <fred@...>

          > Hi,
          >
          > Yesterday I was asked about the viability of mod_perl ( and more
          > generally Perl web based solutions ) in scaling web based applications.
          > I have my own personal experiences with scaling mod_perl, and am well
          > versed with the best practices, case studies, and methodologies out there
          > in scaling mod_perl web apps.

          Because it's perl, not a substitute :-)

          One thing that means is that all of the Test::*
          modules are avaiable for valdating your modules
          and their output. If that doesn't seem like much,
          take a look at the recnet O'Reilly Programmers'
          Notebook on Perl Testing: from use_ok through
          data structures to doc coverage you can write
          automated tests.

          OK, so what?

          - How many sites have crashed in history due to
          lack of regression testing? Why the lack of
          regression tests? Usually they "take too long"
          or "are too complicated." Welp, with Test::Foobar
          you can use "prove blah" to run the tests singly.

          Combine this with FindBin::libs and a symlink
          or two and you can regression test a whole suite
          of modules in a single "make test". Better yet,
          you can "make test install" and be sure that the
          stuff doesn't go into production until it passes
          muster.

          Perl also makes it easier than any other language
          I know of to mix heavily-OO code with functinal
          blocks. This gives you the best of both worlds:
          OO for re-usable code where it works and functional
          blocks at the top levels. The Test::* modules
          support both styles of coding well, which allows
          for complete coverage.

          Take a look at the Phalanx project's results for
          a good overview on the wonderfulness of testing
          in all its forms.

          All of this applies to Perl in general, but bringing
          it to web development is what mod_perl can do for you.

          --
          Steven Lembark 85-09 90th Street
          Workhorse Computing Woodhaven, NY 11421
          lembark@... 1 888 359 3508
        Your message has been successfully submitted and would be delivered to recipients shortly.