Hey folks,

Having spent a few hours (heh) going through my blog and converting and modifying it for jekyll, I think it would be appropriate to list some general directions for improvement of the software. Jekyll is pretty good and it’s certainly more than usable, but I think there is enough room for improvement, and some additional features that would be rather interesting to explore. Because I think it’s really easy for us to all develop our own forks and work in relative isolation, I worry that there’s limited space to have a conversation about how to best improve the platform. Here’s a starting point:

  • Performance Jekyll, particularly for larger sites can take a considerable time to run.

    I’m not familiar enough with ruby to be able to speak to any areas of improvement in terms of code optimization, though I’m not sure that this is the best way to approach this issue. Rather, I think it would probably be best to figure out ways of gracefully preventing jekyll from recompiling pages that haven’t changed every time there’s an update. The possibility that makes the most sense to me at this point would be to include “masks” in the config file, that jekyll can skip over unless run with an --all option.

    Another option might be to develop a way to cache/store certain kinds of output files that require a lot of data to crunch (I’m thinking mostly of things like archives and of other long post loops.) Maybe developing some sort of internal index?

  • I’m (preliminarily) looking at using rpeg-markdown as a another markdown option, which should both allow the extended markdown syntax to work (woot!) and run much faster than maruku (and hopefully rdiscount). It seems difficult to install, but would generally be worthwhile.

  • Increase usability While jekyll just works, and the documentation is reasonably effective given the audience; there isn’t any automatic workflow for jekyll use. Having a few shell scripts and/or a “using jekyll” tutorial that explains day to day use would be quite helpful once you get started.

    As I finalize and put it through it’s paces for a few days I’ll get more clear about this, but, in general I’m thinking of things like: using git hooks to generate the content automatically on commit/etc.

    It would be valuable to write up how to use rsync as that would likely make more sense in some situations. Same with running tasks in cron.

  • Template Directory: Most of the jekyll powered blogs that I’ve seen have been pretty good about publishing the repositories for their sites. This is a good thing indeed, but given how the templates work, I think it would be good to collect various templates together for ease of access and educational proposes.

  • Make Categories More Flexible: While the current blog categories support is great for adding “multiple blog” support, it makes it difficult to interact/loop posts that aren’t in different categories. I’d like to have arbitrary site.key.value objects that work like site.categories.CATEGORY. There are ways to program around this, but they aren’t pretty. This would make jekyll even more incredibly powerful, but I think it’s a much lower priority.

    If you use jekyll, I’d love to hear what you’d like to see from the jekyll in the future. A number of these options (rpeg-markdown, usability/workflow documentation, template directory) are things that I can work on (and will, for sure), but a few of them are beyond me and I’d love to help folks do some of this work, if possible.

Onward and Upward!

(ps. my _posts directory is about 6.2 megs, with a file count that’s a bit south of 1,300 files, so if you wanted to assert that I’m pushing the system a bit hard, that’d be totally reasonable.)