In a recent Changelog episode Matt Biilman and Chris Bach talked about a technology pattern called JAMstack and how they used it to implement the new version of the Smashing Magazine website.

The idea here is that you generate most of your markup statically using something like Hugo or Jekyll and add dynamic components through JavaScript and the use of simple APIs. This somehow reminded me of what I’ve been doing for a while here and more recently also on PyVideo.org. There we generate all the primary content pages using Pelican out of a bunch of JSON files. The one dynamic component here is the search engine, which we’ve integrated with JavaScript backed by a search API written in Go.

But that’s just the beginning, as Matt and Chris have shown. They and their team at Netlify have created a whole toolbox of services ready for this kind of web application. Among them is a CMS and an e-commerce service. With these you can basically store all your content inside services and have a build-step compile it out into static pages, basically replacing what a Git repository has been for us with a custom content service API-layer.

But these are not the only services available out there. There are, for instance, some commercial ones like Contentful, which provides a CMS similar to what netlify-cms does. If you prefer on-premise solutions, there are things like Hexo (a blogging system) and Ponzu (another simple CMS implemented in Go).

Especially for content-heavy sites, JAMstack looks like a great approach. You basically have cronjobs or events that are triggered when the content in a repository or CMS changes and that statically rebuild the affected parts of the website. Everything else is loaded dynamically from APIs and displayed through JavaScript. Sure, this doesn’t work for every kind of website, but for news sites, online shops etc. I can definitely see its appeal 😀

I already have a couple of ideas for GoGraz where that might come in handy. Same goes for the Meet-The-Meetups.org where we are already fetching some of the content from the meetup.com API. More on that, though, when I actually come around implementing it 😉