I have a habit of revamping my website every few years. Since starting on Blogger some ... 19 years ago I've moved between platforms and redesigned the layout several times over.
This time I decided to go for a static site build, which means not requiring managing WordPress plugins and updates, or running PHP on my server. It'll also perform a lot better if I get another big hit, like The surreal experience of my first developer job which sat at the top of Hacker News for a whole day.
My regular site revamping has been documented over the years...
I decided I needed a new website template so I made one
— Version 3 and Problems with IE (2009)
I have to make a series of decisions - where do I take the blog, and how do I get it there? Do I migrate to another blogging service - and if so, temporarily or permanently? Do I write my own temporary fix? Do I write my own blogging platform - and if so, would it work in the same way Blogger did?
— Goodbye Blogger? (2010)
I wonder how long before I get bored of this template and create or find another!
— Blogging about blogging (2020)
The 2010 post also included a prediction which seems to have come true:
Ideally, in time, I will write my own Content Management System which will manage my entire site, not just the blog. It will work exactly how I want it to, and it will be perfectly suited my own personal requirements - bespoke for my own purposes. PHP may be the language I write this in - it may not. By the time I get round to doing it properly I might be writing in a completely different language.
About that...
The project
For now I've kept the layout the same, but I built a markdown based static site generator. I absolutely could have used one of the many frameworks that do exactly this, but I just really fancied having a go at writing my own in Python. It has a blog index, an archive, archives by year and month, tags, an XML sitemap and an Atom feed. It also supports Twitter cards and general OpenGraph tags for other social media sites such as Mastodon:
I didn't spend very long building the new site, and I enjoyed working on it. It's quite satisfying building something of your own, and it's much easier to get done quickly with ChatGPT and a wealth of open source tools available.
I used Chameleon, Pydantic, markdown, pyyaml and BeautifulSoup.
The source code for the site generator is available on my GitHub. I called it Beemo and I've released it to PyPI. I'll try to make it more generalised to be useful to others.
Content
I wrote a script to pull all the posts from my WordPress site via its JSON API, keeping their titles, slugs and tags along with the content. I used pandoc to mangle the HTML content; BeautifulSoup to extract the images; and imagemagick to resize them. I ran the same script on my abandoned Tooling Tuesday blog, a separate WordPress instance, to bring those posts in too.
I also copied the source of some of the article I wrote for opensource.com which stopped running a couple of years ago. I had another look through my old Blogger account, and found a few really old posts I wanted to preserve, like Athletics? and My Received Files so I've ported those over too.
BBC
One thing that struck me while looking back at the content is the number of old posts tagged with bbc or bbc news, either written about or referencing BBC News articles, or about BBC TV shows.
One was about an event I was involved with which BBC R&D wrote up on their blog (and looking now I recognise some of the names on that site); one was about a time in 2007 when my parkour group was filmed for a regional TV show which was broadcast in 2008; another in 2009 where we were auditioned for a TV programme about intelligence stereotypes. One time I had an eventful kayaking trip which was covered by BBC News. I even posted about the death of Ronnie Corbett in 2016, and related this to GitHub.
I was a student in Manchester when Media City was being built, and the BBC were beginning to set up a new base there. I loved the idea of getting a job there once I finished university. I didn't see anything relevant at the time and for some reason I hadn't been looking at graduate schemes. It wasn't until I happened to spot a vacancy in BBC News Labs in 2019 that I finally put in an application, and started in January 2020, initially in London. Later I relocated back to Manchester and ended up with Media City as my base. So I got there eventually!