< a quantity that can be divided into another a whole number of time />


June 10, 2021

I recently subscribed to a free Bnc4Free account in order to keep track of my IRC session. I already save daily log, but since I’m not running Irssi 24h/24, I miss part of the day, especially in the US. I know I should be setting up the whole stuff on a dedicated server at home — and I will do, one day, but I was too lazy the past few weeks. Anyway, it’s been working well and I now get a buffer playback saved in my logs as well. I’ve been using IRC for 15 years or so now, mostly as a casual user (ghost reader or private chat session), but it fits well in my setup.

The Book of Statistical Proofs is “a centralized, open and collaboratively edited archive of statistical theorems for the computational sciences.” It looks like the future Wikipedia for serious stats users. Let’s see how it evolves. On a related point, I also like Modern Statistics for Modern Biology, by Susan Holmes and Wolfgang Huber. I probably mentioned it a few times on this blog or the micro-blog, but it really is a nice resource for R practitioners (mostly base R and ggplot2, as it should be).

Sometimes I wish great authors provide RSS feed for us, human beings, who like keeping an archive of their readings. Lately I came across Infinite Ink which has very interesting articles to review, and a beautiful web UI. Unfortunately, no link to add to my RSS reader, and I’m doomed to not forget about this very elegant site. I now have more than 500 feeds in my personal OPML file. Feel free to grab what you find interesting there. And yes, Dirk, there’re still people reading others’ blog, especially yours.

Keep up the good work, Dirk, I’ll be listening. Also, no idea what happens on Seth Brown’s side, but his site keeps changing again and again. Sadly, no more blog posts available at the time of this writing.

My Emacs -> Vim switch is going well. For the coding part, I’m left with vim-polyglot (+1) and the LSP stuff, which works more (Python and Rust) or less (Haskell and R). This reminds me of John D. Cook’s famous talk about DSL languages. Yes, we need DSLs, and the tools around. What’s the issue with the tooling around the language? LSP clients are generally quite limited in this respect (with maybe a few exception, like VS Code). I believe ESS and Intero are the best ways to interact with R or Haskell, whatever is meant by “interact”. And yes, I know what a REPL is. Unfortunately, there’s nothing close to that in the LSP ecosystem yet. I’m left with R’s languageserver, which sucks at rendering online help via the K key because of rmarkdown issues (I reverted back to v0.3.9 because of that), and neoterm to get a “live REPL” or sort of. Same for Haskell’s tools, BTW. All in one, LSP tooling looks like you all get “goto”, “hover”, and “doc” for free, but barely more, especially when code actions are not implemented.

And yes, plants are growing well in case you were asking.

readings self

See Also

» May in review » March on the road » Lurking on the web » The unquantified self #16 » The unquantified self #15