Conversation

Notices

  1. No debate this morning? I'm fine with that... Keeps my blood pressure down

    Wednesday, 09-May-12 08:20:31 UTC from web
    1. @purplephish20 You missed it actually.

      Wednesday, 09-May-12 08:21:05 UTC from web
      1. @toksyuryel @purplephish20 I don't think a discussion about databases is going to raise his blood pressure anyway.

        Wednesday, 09-May-12 08:22:16 UTC from StatusNet Desktop
      2. @toksyuryel What language were you going to write the app in anyway. I feel pretty confident saying "not PHP"

        Wednesday, 09-May-12 08:24:01 UTC from StatusNet Desktop
        1. @ceruleanspark It's totes .Net

          Wednesday, 09-May-12 08:24:54 UTC from StatusNet Desktop
        2. @ceruleanspark Python

          Wednesday, 09-May-12 08:27:25 UTC from web
          1. @toksyuryel Damn. I was sure it was gonna be basic

            Wednesday, 09-May-12 08:30:31 UTC from StatusNet Desktop
      3. @toksyuryel @ceruleanspark Nice one... Imma compensate by drinking coffee...

        Wednesday, 09-May-12 08:24:12 UTC from web
    2. @widget Is there such a thing as a python opcode cache?

      Wednesday, 09-May-12 08:36:09 UTC from StatusNet Desktop
      1. @ceruleanspark It does cache its bytecode, but even still the slowness is architectural. It's possible to dynammically overload the entire language during runtime, so it doesn't actually know what any particular sequence of instructions actually does until they're being executed. It's possible to achieve tremendous speedups by implimenting the language differently; cython is extremely fast, and stackless is quite good as well. But these alternate implimentations come at the cost of being several language versions behind.

        Wednesday, 09-May-12 08:44:10 UTC from web
    3. @widget I messed around with Pyramid for a bit before diving into learning SQL. Compared to PHP Python should still be a performance improvement, especially with the additional performance improvement of using PostgreSQL and nginx. It should not perform any worse than statusnet.

      Wednesday, 09-May-12 08:39:19 UTC from web
      1. @toksyuryel Now there's a ringing endorsement right there. That's one for the homepage. "Not worse than Statusnet"

        Wednesday, 09-May-12 08:43:49 UTC from web
        1. @ceruleanspark Gotta start somewhere :)

          Wednesday, 09-May-12 08:45:22 UTC from web
          1. @toksyuryel Will I still be able to obsess over sub-second enhancements in page render speed. This is important because it makes me feel needed.

            Wednesday, 09-May-12 08:47:09 UTC from web
            1. @ceruleanspark There'll always be room for improvements^^ you'd have to dig into the Python and SQL code though. PostgreSQL helpfully provides the EXPLAIN and EXPLAIN ANALYZE commands for performance fetishists to go nuts optimizing queries with, and it's even possible to write trigger functions in C.

              Wednesday, 09-May-12 08:50:01 UTC from web
    4. @widget It would be easier, but it would also be exceptionally inefficient and monsterously slower. Just because something is easy to do doesn't mean it's the right way to do it.

      Wednesday, 09-May-12 08:47:16 UTC from web
    5. @widget Python is a terrible web server.

      Wednesday, 09-May-12 08:56:43 UTC from web
      1. @toksyuryel They're also not very filling.

        Wednesday, 09-May-12 08:58:42 UTC from StatusNet Desktop