Last time I reposted a link to my blog on the fediverse my site went down. It's a small server but it should be able to withstand a bunch of get requests. I realised I probably don't need to verify a user's signature on every request, since it's only relevant when posting to the inbox route.
So now I've reduced request time by 100-fold, let's see if there's an improvement.
Hello, world. I've transcended "Using my blog to post about creating blogging software", now I am doing "Using my ActivityPub site to post about creating fediverse software".
So far it's extremely basic: Actors and Articles are represented, but no Create activity yet. The actor is followable but don't even think about sending me a like, note, or reply yet. If you want to ask any questions probably best to message @paul@hades.town for now.
That's all! Thanks.
Last week's design sprint was super interesting and productive, everyone did a great job coming up with ideas and designs, and I'm looking forward to reaching the point that we can start building a product.
Most of my thoughts this week have turned to packing boxes, as we'll be moving house next week. I'm not quite sure how I feel about it - it will be sad to say goodbye to our current home, but as the time gets closer I am getting more excited about owning my own home and creating a place that I know is going to be more permenant than anywhere I've lived in the last decade or two.
Adding to my links collection this week: Storygraph, a promising alternative to Amazon's Good Reads. My own personal boycott of tech giants is not so much to damage those companies in any way (it's only a drop in the ocean anyway), but to support small projects and help ensure there are alternative products and ecosystems for people who want a choice.
This month, a scientific study from Germany found that search engine results are getting worse, to nobody's surprise, in part to AI.
I've observed this decrease in result quality happening over the last few years (decade, even) as have many others in indie web circles. As with many problems on the modern web, there is a lack of intelligent curation and moderation that is being exploited for profit, and in my opinion it'll be hard if not impossible for search engines to do anything to remedy it in the long term.
But we are starting to see people experiment again with alternative forms of web navigation: manually curated directory listings, link aggregators like Lemmy, and the resurrection of the webring. I've added a links section to my website and will be starting to populate it with useful and interesting resources as I discover them, but what I'm really hoping for is to see someone create a federated discovery service inspired by StumbleUpon.
Next week I'll be joining my colleagues in a design sprint, which will be an interesting experience. We'll also be starting work on the RSS feed that I mentioned in my last post. And plenty of code refactoring, including dropping our storybook-based design system in favour of a combination of Figma files and hardcoded component demo pages within our main codebase.
With the Christmas period now a distant memory, what is there to look forrward to in January? Actually we have a couple of winter celebrations coming up that our household like to celebrate, in our small ways.
This first is Wassail, an old English folk celebration to ensure a good havest for the new year. It's celebrated on 17th of January, Old Twelfth Night (as it would have been 12th night in the Julian Calendar) by drinking mulled drinks, singing songs, and hanging toast on trees.
The second is Burns Night on January 25th, a Scotish celebration remembering poet Robert Burns. Again, celebrated with song, drink and food. Although since we're not Scottish this is a smaller celebration for us; mainly a chance to try our hand at making a vegetarian haggis and enjoy a glass or two of whisky.
At work this week I've suggested we add an RSS feed to the site. In an era where so many people are becoming fed up with social media algorithms, many users have started to seek ways to curate their own news feeds and digital social networks, and it's a quick win for us to add this kind of open data feed to our site. Plus it's a timely coincidence, yesterday being the anniversary of the tragic death of RSS inventor Aaron Swartz.
The lack of sunlight at this time of year really gets to me. It makes me feel heavy and slow. But we've had nice things to keep us occupied; the dregs of the Christmas period means plenty of snacks to finish off, the jigsaw to complete, and the decorations and tree to keep us feeling cozy... until tomorrow!
I tried a 7am swim yesterday, which left me in a good mood for the day, and remarkably awake. Unfortunately today the tiredness has caught up with me, but at least it's only Friday.
As for work, it's a slow easing in for this week. I'm trying to reduce the quantity of Javascript we send to the client, eventually with the goal of removing our React dependency completely. I think it's possible, and it will be nice once done - but it's going to be a long process!
Generating a URL for the HTML view of a data model is not a trival operation.
A URL represents a reference to a resource, but often it also has context.
For example:
http://example.com/posts/test-post http://example.com/authors/paul/test-postBoth of these links might display the same resource, but the second one might have a link back to "see more of paul's posts".
Another example:
http://example.com/films/batman http://example.com/screenings/2024-06#filmThese both refer to the same data model, but in very different contexts. In fact, the first refers to a keyed entry in a list, whereas the second is a sub-property of an entry in a different list.
So in order to generate a link we need some extra context. For the first URL, we can generate this by knowing "we want the films list, and the entry under batman" For the second, "we want the film data from an entry in the screenings list, specifically the entry for june 2024".
In code form this might look like this:
Subscribe with RSS
getURL('films', { id: 'batman' });
getURL('screening-film', { screeningId: '2024-06' })