It seems like a long time ago when I originally launched this site. I can’t really remember when it was now, but the wayback machine traces it back to 2001, when the site was hosted on catch22.uk.net. HexEdit and my other software were in development for a considerable time before that even, sometime from the mid 90’s I think.
For the first few years, catch22 was built with simple statically-generated HTML that was uploaded to my ISP’s free web hosting. I had no web authoring tools available and everything was done in a text editor. The term ‘blog’ had only started to emerge, and it would be years before GitHub came into existance.
To ease development of the site, I devised a way to use templates which helped separate the content from rest of the layout and navigation. I made a simple macro syntax to help me make the source more modular. Eventually I learnt about ASP and ‘server side includes’ and for a while the site ran quite happily like this.
Drupal
At some point in the late 2000s I must have decided that Drupal was a good idea. It wasn’t. I learnt a lot about Drupal, PHP and database hosting, but keeping the site maintained and up-to-date with security fixes was a never-ending world of pain. Although the promise of content management and editing directly on the site was a real draw at the time, Drupal was a terrible choice for a personal blog in hindsight. It’s just so heavyweight and unweildy, and… need I mention PHP? What a disaster.
Full circle
Some time in 2018 I migrated the site to Jekyll. It was back to being a static site again, simpler but more modern this time, with markdown content nicely separated from the layout. At last it was stable again and needed little maintainence. I never really got into Jekyll due to the Ruby toolchain that was just a bit fragile and well, quite slow really. But it did the job.
Always behind the curve
Is that such a bad thing I wonder? This site was founded on that concept, where ‘plain old’ win32 and C were a bit of a relic even then, and unweildly frameworks like MFC ruled the roost.
The site is now running after yet another refresh, this time with Hugo, another static-site generator. But this one feels different. It’s stable. Hugo is super-fast, and the tool itself is self-contained and although the learning curve is a bit steep, it just works.
After the nonsense of CMSs and all the complexity these entail, it feels like we’ve gone full-circle. Static sites have been a thing, for a while now. I’m late to the party, again. There always someone at the bleeding edge making new frameworks, and abstractions, obfuscations, and things that just don’t need to exist. We’ll see a static CMS gain traction at some point, you can count on it.
Simple is always better.