w/e 2024-02-18
Yesterday we both went on a one day “Hobby Chainsaw” course at a local college. We’ve inherited a chainsaw, and often have some chunky wood to dispose of, but didn’t have the courage to use the thing.
It was a fun day and really helped. It hadn’t occurred to me that we’d take chainsaws apart but that helped with demystifying them a great deal. We learned how to replace a broken pull start cord and how to file the chainsaw sharp again, among other things.
After lunch we went out and got sawing. The start of this was the most confusing because the procedures for starting the machines, using the choke, starter cord, trigger, etc, and turning the safety lever off/on, were all second nature to the teacher (who was great otherwise) but meant nothing to my head. I had no mental model for the thing and struggled with the sequences.
But once we got going it was all good. Really useful to have someone watch you cutting and point out how to do it better, and what was fine, compared to what was potentially dangerous. It all feels a lot safer now than before I’d touched a chainsaw.
A bit like scything – so many people remark how dangerous it looks and be careful you don’t chop your feet off lol, when it’d be really difficult to get your feet, and I’ve never once cut myself on the blade at all (so far).
§ This week I looked into how to forecast the best days to plug in our electric car to charge using power from our solar panels.
I found pvlib, a python library for “simulating the performance of photovoltaic energy systems and accomplishing related tasks”. It’s pretty complex and assumes a lot of knowledge about PV systems. Its documentation isn’t super friendly and the example code wasn’t entirely reassuring. I’m sure it’s possible to just “generate a forecast for panels of this spec, at this location” but the docs don’t easily show that. After a couple of hours of fiddling I decided I didn’t want to learn about Global Horizontal Irradiance, turbidity, the Davies & Hay model…
I’m surprised I couldn’t find blog posts with examples of how to do this kind of thing, which is the case for most stuff I want to do. I did find a Polish one which uses a module that no longer seems to exist (pvlib.forecast) but otherwise I came up blank.
I did ask ChatGPT which initially seemed promising but, unsurprisingly, the code was riddled with basic errors, and also used modules that no longer existed, and life is too short.
(To be fair: this week I did use ChatGPT to successfully answer a couple of very simple coding questions that I couldn’t easily google answers for, given how swamped google is with chaff these days.)
So I searched for APIs that would do all the complicated stuff for me.
Forecast.Solar looks like a nice API, with a “run by one guy” vibe. For free you get a forecast for today and the next day. Incremental levels of annual subscription get you further and more detailed forecasts and more sets of panels (we have two, at 90 degrees to each other).
Solcast also looks useful and I signed up for a free account to give that a try. I initially had fun knocking up a quick chart using PHP to fetch and cache the data (you only get 10 requests per day on the free tier, so caching it was useful) and Chart.js to draw a chart. Really nice to hack away on a quick little thing rather than a big project. Web development can be fun!
But ultimately, of course, web development was not fun, because I got frustrated working out how to do things with the JavaScript, thanks to overly terse documentation and most examples online being for earlier, incompatible, versions of libraries than those I was using. Plus, although you can see a 72 hour forecast on Solcast’s site, the free API only gives you 48 hours.
At this point it’s less satisfying but much easier to use these things they have now called “weather forecasts” and looking for the yellow suns.
§ We watched both short seasons of The Newsreader, an Australian drama (on iPlayer) about a 1980s TV newsroom, which was entirely new to me. It was really good, better than I expected, especially the first season. Some good performances, especially Anna Torv, and it was refreshingly much less pompous than US TV-newsroom-focused dramas. The only anchor who’d have been remotely likely to gravely intone how important their broadcasts are to the nation’s democracy was the one portrayed as old and out of touch.
§ Still enjoying the novelty of having a machine grind the morning coffee for me. Hopefully the guilty feeling that this a decadent step too far from the honest toil of hand-grinding will wear off soon.