I was asked to create a simple website for a Dog Walking and Pet Sitting Service in Leiston, Suffolk.
It's a static one-page website, rendered with HuGo. As well as being customised with semantic markup, the site sports a contact form and links to the business Facebook page.
We've got a Bosch Serie 4 washing machine (specifically, a WAN24100GB).
It recently stopped draining and instead just sat with the pump running.
On most front-loading washing machines there's a cover providing access to the pump housing so that blockages can easily be removed.
Unfortunately, sometimes the blockage pokes through and jams this cover, preventing it from being unscrewed.
This post details the process of disconnecting the drum outlet hose from the pump housing in order to remove a blockage and free the cover back up. Although performed on a Bosch, the process should be similar for most washing machines.
I am not, by any stretch of the imagination, a fitness fanatic. I always assumed that I'd start to exercise more with age, but rather than getting round to it I... err... got round.
That said, I'm not entirely ambivalent and I've tracked my daily steps ever since my Watchy got me back into the habit of wearing a watch (though they're now collected by my Bip).
My mindset towards fitness is perhaps best summarised as "I could do better, but there are other things I need to do".
A little while back, in pursuit of an easy win, I bought an under-desk exercise bike which, as the name suggests, is basically a pair of pedals attached to a resistance wheel.
Because it's always there, I sometimes find that I've unconsciously started pedalling.
The bike that I chose came with a cycle computer, which displays the revolution count, RPM and calories burnt. However, being a standalone device those figures are completely untracked, with no way to see what kind of benefit the bike was delivering over time.
In order to make these metrics available, I decided to use a Raspberry Pi Zero to build a wifi-connected cycle computer which writes stats into InfluxDB for visualisation and tracking in Grafana.
In 2014, I wrote a small Static Site Generator (SSG) in PHP.
At the time, Hugo had only seen a few public releases and both it and Jekyll were written in languages (Go and Ruby, respectively) that I wasn't quite as familiar with.
Both were more complex than I felt I needed, although there probably was also an element of NIH Syndrome: it was as much for the challenge as to satisfy a need.
Although originally written for use on a small .torify.net blog, in 2019 I started using it for other purposes and, in 2020, I migrated my recipes site over.
One of the advantages over Hugo and Jekyll was templated input: when invoked, the SSG prompts the user for a sequence of inputs in order to generate and store structured HTML into a sqlite database.
Recently, I decided that it was (past) time to look at migrating away from my ageing SSG.
The migration was performed by scraping the HTML site that the SSG had generated, so in this post I'm going to talk about the process that I used to consume HTML and convert it into Markdown ready to be fed into a different SSG.
I'll also talk about a headache that I ran into with lxml when attempting to extract raw HTML strings.
One of the features that Mini's include is a Tyre Pressure Monitoring System (TPMS).
These use sensors in the wheel and are supposed to warn if a tyre's pressure gets too low. Unfortunately, they're known to be a little temperamental and can throw false alarms, which won't reset until the warning is manually cleared.
This short post details how to clear the warning and tell the car to recalibrate against current tyre pressures. It should work on all Minis from 2007 onwards (the photos are from a 2012 Mini One).
Last year, after we ordered an Xbox, Amazon tried to leave us with the bill despite it being stolen from within their fulfilment chain.
Whilst that wasn't, in any way, Microsoft's fault it also wasn't the best introduction to the world of Xbox.
Admittedly, the (replacement) Xbox was happily received when unwrapped on Christmas day, even if the enthusiasm didn't quite manage to offset the stress and irritation that had characterised the prior weeks.
Since then, the xbox has seen regular use but, crucially, in a way that's avoided me having to use the Xbox store at all.
Unfortunately, that changed this week, with the experience instilling a need to write about Microsoft and their apparent inability to arrange inebriation in a brewery.
As the web has aged, the issue of Link rot has increased, with links silently breaking as sites and services fall by the wayside.
Although it's always been a problem, it feels like it's been more visible recently. From people abandoning Twitter to behemoths like Google deciding to turn off once popular link shorteners, there are likely going to be a ton of links that turn dead.
Link rot is particularly pervasive on blogs of a certain age, with links that once provided useful context now leading nowhere or, worse, to somewhere completely different. As an easy example, I used to write on a revenue-share writing site, but the domain helium.com has since been acquired by a crypto company: any links that I, or anyone else, might have made to the many posts there are as dead as a very dead thing.
Recently, Terence Eden asked bloggers in the fediverse how they deal with bad cases of link rot within their own corpus:
Beyond updating affected posts to note that links have broken, I didn't have a good answer to this: after all, what can you do when there's no suitable alternative?
But, it did motivate me to actually get started on a related project which had been turning over in my head: Although there's not much that can be done about linkrot that's already happened, there are things that we could do to help protect ourselves against future breakage.
In this post, I talk about setting up a link-rot insurance mechanism by automating the preservation of (almost) anything that I link to from within a blog post.
I recently decided to actually do something about an idea that I've been turning over for a bit.
I'll write a proper post on that later, but in order to do it, I needed to be running some form of link archiving solution (i.e. I needed some form of self-hosted Wayback Machine).
After a bit of looking around and experimentation with various data archival solutions, I decided to self-host an instance of Linkwarden. If you've not seen it, Linkwarden have a demo instance that you can log into (although it seems to periodically go unreachable - presumably whilst the demo resets itself)
This documentation details how to deploy Linkwarden into a Kubernetes cluster
