I should probably just provide a quick refresher on how the whole mess of my blogging works, seeing as I’ve got a fairly complex, non-intuitive system.

First, I write posts in John Gruber’s Markdown. I used to use mg for this task; recently, I saw the light and decided to try to find a better text editor for this – and thus discovered Mou, a fantastic, lightweight editor for OS X. Perfect to run on elspeth.

Next, I pass it through a horrifying Ruby script which deals with tags. I really could rewrite this as a Jekyll plugin – already, it requires most of Jekyll – and I really need to do something about improving it anyway; the performance of it, while not terrible, could be better. Then I feed it into Tom Preston-Werner’s Jekyll, with a few patches, some of my own creation (faster_lsi, bspic and pandoc-ruby) – smart elements come out of Pandoc’s much nicer Markdown transformer, and image injection is done by my own Liquid extension.

I also build my startpage, which uses a set of XSL transforms. I don’t use it very often, now, after I switched to building my own Chromium.

This all then gets rsync‘d to my local test web server, then out onto the web.