Blogging with Hakyll

Written by Eric Rasmussen on April 11, 2013.

The biggest barrier to actually maintaining a blog isn’t having ideas, or even writing them down, but having to manage content. Typically content hosted on a web site somewhere, protected by a password, navigable through a user interface that requires clicking… oh, lots of clicking. It’s usually something along the lines of:

1. Go to the admin page for my blog or CMS
2. Login with a password
3. Click around to create a new post or find a saved draft
4. Try to make the rich text editor do what I want
5. Give up and use html
6. Save and publish the changes

It’s all a lot of work, and I am finding it an increasingly unsustainable workflow when I spend much of my day in emacs or at the command line. For comparison, here’s my workflow for open source software contributions:

1. Write something in a text editor
2. Run some commands in a terminal

With hakyll, now I get:

1. Write something in a text editor
2. Run some commands in a terminal

There are no shortage of static blog generators now that they’re in style, but hakyll stood out for me because it has a clean DSL-like feel, I’m at home with haskell, and pandoc is amazing. I plan to write blog posts primarily in markdown, but I also may use ReST or LaTeX, and hakyll’s default pandocCompiler can handle them all (and a number of others; see the pandoc website for more).

My other goal (not specific to hakyll) is making this a truly open source blog, in the sense that you are welcome to clone it and fork it. I’d prefer you don’t steal all my posts of course, but you are completely free to use my site.hs file however you wish.