25 May 2020

[Food] My sister and I have never really had a 'default café', somewhere nearby and easy to get to which does good food. In the autumn of last year, The Music Pantry opened in Wedmore, about 12km from us. It's an exclusively vegan place, laid-back, with really friendly owners, Heather and Liam. They love music, there's always an interesting playlist on the go, with record sleeves and music-related art decorating the place. They were holding live events too — hopefully they will again so I can go along. But for now, they have begun to offer their excellent vegan roast dinners to take away, along with a small selection of groceries. Hannah and I decided that this would be a weekly ritual and that we'd use it to have a weekly picnic together. Yesterday's was at Ashton Windmill. Very pleasant.

[Haskell] I continue to mess about with Haskell. This week, items from FP Complete's blog — via Planet Haskell — came up in my RSS reader. In particular, the concise Crash Course to Applicative Syntax — as well as a post it links to — is a nice reminder of much of what I've been reading about in the last few weeks. I keep reading references to arrows. They are a generalisation of monads, but beyond that, I have no real idea of what they are. I really want to spend my Haskell time in the IDE as much as I can for the next fortnight or so, and then I'll re-join the study of how the abstraction of computations appear in Haskell.

On the topic of staying in the IDE, my trail-building program continues to grow, but I'm keen to experiment with something more algebraic next. In October last year, Tom Nicholson challenged me to implement a basic Inductive Logic Programming system in Haskell as a learning and exposition exercise. This is an excellent focus as an intermediate goal, I think. While there are logic-themed libraries available for Haskell, I wanted to get a feel for how a simple search through a space of subsets, rules or even functions would work — a search that a symbolic machine learning system would do. The other day I noticed that Monday Morning Haskell did a post about implementing Q-learning in Haskell, and I'm sure there's much more out there in the direction of implementations of machine learning systems in Haskell.

[Beekeeping] It's been a big week for my beekeeping. I saw one swarm on Thursday and, after a crash course over the phone with a local beekeeper, I dealt with its recapture myself. By the time Sunday came around and I got a call about another swarm, I knew what I was doing! There seem to be swarms all over, because today we got a third swarm, but we had nowhere to put it! Even so, we now have four colonies in our little apiary. Fingers crossed that they stick around.