Use R you fucking monkeys

  • OgdenTO [he/him]
    ·
    2 months ago

    There's a joke in here about pirating SPSS. And then R, like a pirate.

  • chickentendrils [any, comrade/them]
    ·
    2 months ago

    SPSS and SAS are atrocious. For a weakly typed programming, I'll reach for some well-known python packages.

    I'm slightly wary of weak typing nowadays tho. I've seen at least two times where weak typing lead to processes not throwing exceptions and producing slightly weird results in huge data sets. I'm probably biased against R because these happened to be R lang processes. Both were not detrimental, fucked up optimal pircing in vampiric industries (like self-storage). Of processes I've inherited though, good tests and artifact validation (that have worked and caught things) were more commonly written in Python.

    • RNAi [he/him]
      ·
      2 months ago

      I have to do a very easy job, which only takes a lot of time and effort because I'm forced to interact back and forth with SPSS and SPSS stored data. I can't understand these people.

    • Speaker [e/em/eir]
      ·
      2 months ago

      Is there anything especially good that's strongly typed? Spark is fine, but I'm one of 5 people I've ever encountered who's actually using the typed dataset APIs or leaning into type safety as a feature, and doing so puts me at odds with basically every book anyone learning Spark is likely to read.

      • comrade_pibb [comrade/them]
        ·
        2 months ago

        Python is strongly typed. It's also dynamically typed

        That's different than weak typing

        • Speaker [e/em/eir]
          ·
          2 months ago

          The qualifier "especially good" excludes Python. Static typing would be preferred, but not required.

            • Speaker [e/em/eir]
              ·
              2 months ago

              I'm already writing Scala, Kotlin feels like it'd be a lateral or even negative move. 😄 I'd be curious to see what some nontrivial Spark in Kotlin looks like, though.