I know that at least two human beings read my blog via RSS, so I’m writing this post to issue an apology if I’ve just pumped your feed reader full of my old blog posts.
The reason for this is that I’ve done some updates to the site structure and as part of that I’ve changed the way that the permalinks for posts are created. There are full details in the site changelog, but essentially I’ve changed it so that posts exist under a new /posts/ section and not directly underneath the site root at /{YYYY}/{MM}/{DD}/. My understanding is that this will register as new posts to most feed readers and I offer my sincere apologies for polluting what is probably an otherwise ergonomic and curated space for you.
- changelog.txt (.txt, 4KiB)
This was the result of me mulling things over for a long time. I want to keep the site as flat as possible and don’t have a lot of different content types; and the date-based permalinks kept things consistent from how the site used to be powered. However I read something recently — although I’ve lost the link, sorry — which stated this to be harmful because every section of the url should be resolvable to a page. This meant I should have been having pages return for each component of the date: /{YYYY} creating an index of all posts for that year, /{YYYY}/{MM} creating a list of all posts for that month etc.
Back when this site was dynamic, this was how it worked. You could visit a url for a year, month, or day and get the content for that period of time. I could probably configure Hugo to generate these pages but I think that would be trying to make the content fit the old structure instead of making the structure fit the content.
Further to this, I was also brushing up against the limitations of having a single /projects page which contained links and descriptions for my technical projects. I wanted to migrate this to make use of Hugo’s list pages to have an index of projects each with their own pages. This lends itself to also delineating posts into their own list page as well to keep things consistent.
I’ll also be setting up some redirects in the web server to try and minimise the amount of linkrot I’ve created as well as the amount of annoyance in the RSS feeds!