Earlier this month I mentioned on Mastodon that I was replacing a Docker-based local development environment at my day job with a Nix-based one, orchestrated with overmind and a justfile. There was quite a lot of interest in particular in how overmind and just could be used to replace a container / compose-based local development. While I can’t share the details of the significantly more complex migration I did at my day job (yet!

The last few months have seen a flurry of changes on major social media websites like Twitter and Reddit with regards to API access. This has also resulted in a number of new competitors seeing rapid growth and becoming new hubs for online discussion. Let’s start with the first point: you will always be able to save Reddit comments and Tweets to Notado by highlighting and saving a text selection. This doesn’t require any API access or scraping, and it will always work because it only relies on what is visible in your browser.

Earlier this evening I finally had the opportunity to watch Spiderman: Across the Spiderverse, which I enjoyed a lot. I have been reflecting on how my experience of films like these (films with teenage protagonists) has changed with age. Even in my mid-to-late 20s, I would often identify clearly with the teenage protagonists, but over the last few years I have increasingly started identifying with the parents of the protagonists. Spiderman: Across the Spiderverse was no exception to this recent trend.

The data model for bookmarking and highlighting services past and present can generally be distilled down to the following: * URL * Title * Scraped Content * User Highlight * User Annotation * ... Other Metadata The bookmark itself is tied to a URL, and anything else related to the bookmark, such as the title, the scraped content (if the service scrapes on your behalf), highlights and annotations are stored as additional metadata linked to that URL.

For the past three years I have maintained a browser extension that I publish to both the Firefox and Chrome web stores. I recently received an automated email from the Chrome web store overlords who had decided I was violating the “Use of permissions” policy by requesting the cookies permission “without using it”. Here is a snippet from the manifest.json for my extension: { "permissions": [ "activeTab", "contextMenus", "notifications", "cookies", "https://<SPECIFIC_WEBSITE_RUN_BY_ME>/" ] } I request the cookies permission for https://<SPECIFIC_WEBSITE_RUN_BY_ME>/ so that I can transparently use the cookie set by that website when the user logs in to send authenticated HTTP requests from the extension.

I have wanted to redesign my website for a while now. In fact, I have wanted to redesign my website ever since I saw Nathan Wentworth’s website. If you haven’t seen it before, check it out. To this day it remains my favourite personal website on the internet. My favourite feature on his website has to be the regularly updating feeds; music from https://last.fm, recent posts from https://mastodon.social, and recent bookmarks from https://pinboard.

In my mid-twenties my body was in pretty rough shape. I was suffering from chronic pain and had terrible mobility to the pain where my quality of life and mental state were being severely adversely impacted. I was also terrified at the idea of entering a gym. I worked with an excellent chiropractor in Birmingham (UK) who helped to provide significant pain relief and subsequently referred me on to a physical therapist.

The past few weeks I have been putting off updating my personal website, which has sat untouched since October 2019. I’ve done quite a lot of interesting work since 2019, so finally this past weekend, I forced myself to sit down, clone the repository, and update it. My personal website is nothing fancy; just a static website built with Hugo and a community theme, and hosted on S3. This should be simple, right?

While watching the implosion of Twitter in real-time over the past week, I noticed more and more people in my Twitter feed sharing links to their Mastodon accounts. During that time, https://instances.social was posted to Hacker News as a tool to help people find a Mastodon instance to register with. My initial experience with this was quite poor. Mastodon's onboarding story needs so much work. Still can't pick a server. Tried https://t.

User comments and discussions on the internet have a pretty bad rep these days, to the point where many people have just opted out of online commentary completely, going as far as using a combination of DNS filtering and browser extensions to remove them entirely from their browsing experience. I, on the other hand, have spent the last couple of years refining ways to separate signal from noise in the cacophony of online user commentary, as I have come to believe that the most valuable knowledge I acquire on a daily basis comes from commentary and discussion rather than publishing.