If we do not understand both the data and the models completely, it becomes very difficult to spot problems in the software we use to work on them: unexpected behaviour arising from software bugs may be mistaken for a peculiarity in either of them. It is then crucial that we minimise the chances of this happening by applying all the best engineering practices we have at our disposal. — The Pragmatic Programmer for Machine Learning.

Dissecting the GZIP format. Very interesting read for those interested in compressing techniques. #clang

If you need an easy way to convert MS docx to a PDF from the command line, don’t forget that lowriter (from Libre Office) has option to convert any document on the fly.

Stick with the mainstream & boring unless a competing alternative that is so much simpler and/or more powerful, and that has an acceptable learning curve, so that it will justify leaving the flock, exists.

Best resume I read in a while.

All mainstream, general purpose programming languages are (basically) Turing-complete, and therefore any programme you can write in one you can, in fact, write in another. There is a computational equivalence between them. The main differences are instead in the expressiveness of the languages, the guardrails they give you, and their performance characteristics (although this is possibly more of a runtime/compiler implementation question). — Why Haskell?

Plotting data in the terminal with gnuplot.

LSP: the good, the bad, and the ugly.

♪ The Cure · A Night Like This

Recorded live at the Roskilde Festival 2019.

Zsh Tricks to Blow your Mind.

Pesto tart with goat cheese and pine nuts. It wasn’t too bad a job after all.