At work the major platform is the JVM, I say that rather than Java because there is most definitely a growing trend away from Java to achieve better individual and team productivity. Scala has played a role in this but it appears that the current leading choice is now Kotlin, which I have to say I do like a lot. Additionally there is interest in non-JVM languages with Go and Rust being the two front runners. For small tools I have tended to rely on Python since the mid 90’s, although I have become worryingly comfortable writing shell (zsh) scripts.

I do like to use Jupyter for experiments, and for documenting ideas where the/an algorithm is a part of the process. I have in the past also used Mathematica (now the Wolfram Language) for similar tasks, and still love the “batteries included” capabilities of the language.

At home, Racket is my favorite small tool language, taking the place of Python for a lot of tasks. I have also started to use Rust more as well. One personal project that has been very rewarding is the development of a set of Racket packages implementing some core Machine Learning (ML) algorithms. These packages are not intended to be production ready, I wanted to learn more about how ML processes worked and decided the best way was from the ground up. The result is a set of simple, pedantic implementations that are written to be readable rather than to scale.

Development Environment

Most of my development is done old school, command line style, on macOS and Linux. I use Amazon Linux at work, Ubuntu at home and have more than one MacBook Pro. I have weened myself of Eclipse, and while I have gone back to Emacs for a lot, I do use IDEA for Kotlin, Python, and Rust development.

I keep the development environment across these machines approximately in sync using a dotfiles repository in GitHub.

The sweet ML development machine under the desk at home is based on an HP Omen desktop with 64Gb memory, 8Tb disk, 512Gb SSD, an NVIDIA GTX 1080i and an NVIDIA Titan V, running Ubuntu for GPU training and inferencing. I use the same large monitor at work, and at home, a Dell UltraSharp U3415W 34-Inch Curved. I have also a Advantage keyboard and a XYZ mouse at home and work with a XYZ KVM switch at home connected to the three machines.

Programming Languages I Have Known

Yes, I collect these things; I am fascinated by the different approaches that languages take, the evolution among language families, and the specialization and generalization they support. I also periodically scan eBay and others for manuals and books for the more obscure or fossilized languages. Unfortunately some of these only exist as online, usually PDF, scanned manuals.

Because someone paid me to

  • Ada 83/95
  • BASIC Family
    • MS/Sanyo Basic
    • Visual Basic
  • COBOL (yes, really)
  • C Family
    • C
    • C++; Glockenspiel, Zortech, MSVCC, GCC, CLANG
    • C#
    • Objective-C
  • Erlang
  • FoxPro
  • JavaScript
  • JVM Focused
    • Java
    • Kotlin
    • Scala
  • Lispy Languages
    • Clojure
  • Mathematica/Wolfram
  • Modula-2
  • Perl
  • Progress 4GL
  • Python (from 0.5, and built first non-Unix port)
  • Rust
  • Query/Constraint Languages
    • OCL
    • SPARQL
    • SQL; DB2, MS T-SQL, Oracle PSQL, Postgres

Mostly for fun, or education

  • Assembler
    • 6502
    • 8088
    • 80x86
  • BASIC Family
    • BBC Basic
    • Commodore Basic+
  • Eiffel
  • Esterel
  • Forth
  • F#
  • Lispy Languages
    • Racket
    • Scheme
  • ML Family
    • Haskell
    • Miranda
    • OCaml
  • Occam
  • Pascal Family
    • Modula-3
    • Oberon
    • Turbo Pascal
    • VAX Pascal
  • Prolog
  • SNOBOL
  • Smalltalk, including Squeak

Related Posts