This is a prototype post. The interesting part isn’t moving a website from one host to another. The interesting part is changing how the site is built and maintained.
WordPress has served the blog well, but my actual workflow is increasingly somewhere else: VS Code, Git, Markdown and pull requests. So the obvious question became: why shouldn’t the blog work the same way?
What I want from the new version
The requirements are deliberately boring:
- write posts in Markdown
- preview them locally
- submit changes through pull requests
- validate the site before merge
- deploy automatically
- keep the LARSINUS visual language consistent with the other projects
- collect basic traffic statistics without turning the site into an analytics project
That’s enough.
What I don’t want to rebuild
I don’t need to recreate WordPress feature for feature. I need the features I actually use.
Search, RSS, categories, metadata, redirects and analytics are useful. A giant plugin ecosystem is not a requirement by itself.
Why this is running in parallel
The existing blog stays where it is while this prototype runs at new.larsinus.com. That gives me somewhere to experiment without turning a design exercise into an outage.
If the workflow feels better, the old content can be migrated gradually and the domain can move later.
And if it turns out WordPress was quietly doing more useful work than I realised? Then I’ve learned that before touching production. That’s a pretty good outcome too.