Last time I updated my blog, I was using Clojure static site generation built on top of stasis. This served me quite well. Over the past week or so, I've been hacking on a new blog (and also have a new site in progress that will be up after I move to Florida). This blog is put together using hexo and Tufte CSS.

Why?

My old blog engine worked well for the most part. The major wrench in the works was actually the dependency on Emacs as an exporter. It prevented me from doing a lot of fancier things with the content because templating by string concatenation is a pain in the ass. It also would routinely break with updates to org-mode.⊕To be frank, my most significant gripe with Emacs these days is how difficult it is to maintain a statically versioned config. Once installed, things don't update; but whenever I re-run the setup (which happens more often than I want to admit when I'm hopping back and forth between machines), things will inexplicably break because a MELPA package updates.

I really liked the Clojure piece: that bit was very pleasant to work with. I didn't take the time to understand how some of the pieces *cough*optimus*cough* worked, but it still did exactly what I wanted.

Ditching org-mode

I ultimately ditched org-mode entirely, which was rather disappointing. The problem ultimately was that nothing supported it, and I got sick of rolling my own workarounds when Markdown covers 99% of my use case and the edge cases are covered by inline HTML.Yes, I know ox-md exists and will let me export to Markdown, but there isn't much point in exporting from org when it is giving me relatively little. The amount of fiddling required for the more advanced org features to render the way I want them to is too much for me.

I really like org-mode, so this was a tough sell for me. I even went through and created an org AST parser in Clojure (using the output of org-element as input) using Enlive to transform it to HTML, but it was finicky as hell and I knew that I would not want to update it when the output of org-element changes next.

Really, I could have plugged in markdown instead of org as the parser for my existing blog and gotten away with it. But no. That adventure is over for now; I have other adventures that are consuming what was formerly fiddling-with-blog-engine time.

Tufte CSS

I fell in love with Tufte CSS as soon as I saw it. I don't know if it is actually a great choice for my blog, but I'm gonna give it a shot! The highlights of it are that it has excellent font design, is incredibly simple, and has this lovely concept of margin notes. Margin notes are really simple in concept but I've never seen them on a blog before. I am rather fond of asides and frequently littered my posts with parentheticals containing them. I believe that margin notes are better suited for this.

No More Comments

I never really had issues with my Disqus comments, but I also never had much use for them either. Nobody commented. They provided no analytics and I doubt that I'd have used them anyway. If people want to comment on a blog post, they can email me, or tweet at @emallson.

Why Hexo?

I could describe some of the things I like about it, but honestly: it was the first batteries-included static blog engine for Node that I came across. It is doing everything I want for right now, so I'm unlikely to change it for the moment.

In Conclusion...

This is one part of my effort to update my site as a whole. Updating the style of my blog is an important piece. I have updated my main page as well, and am debating whether I should stick with Bootstrap or go with Tufte. I feel like I could accomplish a lot with that margin to give more info and character to specific events, but we will see. We will see.