After taking a break from working on Harlie, today I cranked up Emacs and dove back in. And was immediately struck by the need to add a few more comments. When folks say you need comments even if you're the only one to be working on it, they're definitely right.

It's only been 4 weeks since I worked on the code, and some bits are almost incomprehensible. Another thing that struck me was the need to have a bit of a tidy-up of some of the code. However, that may turn into a complete refactoring project, so I'm going to leave it for now whilst I'm hacking deep into the code.

At the moment, there are still no long-term plans as to what the bot will do, or trivial things like documentation. I think my solution to this will be as follows:

1. Get drunk.
2. Bash out some random ideas and plans. The feature list will include the word 'cool' at least once. …per line.
3. Sober up. (optional)
4. Wonder what I was on when I wrote the documentation
5. Pray for a miracle to happen (such as advances in neural control, so I can just think of it and it appears in my editor.)

(About step 4 is where I'd get VC funding if I were using ${this_weeks_web_2_0_language}.)

Short term, however, goes a little like this (in no particular order):

• Perform a little code polishing.
• Fix the parser to allow modules to easily add extra commands.
• Add code to support modules, the loading and unloading of in particular.
• Document, document, document!
• A script to grab all the dependencies.

If you want to help out with debugging, programming and otherwise developing Harlie, please keep an eye on the bug-tracker (http://www.furryhelix.co.uk/bugs) and the gitweb interface (http://www.furryhelix.co.uk/git). If you want to send patches that fix a bug, attach them to the relevant bug. Other patches and so forth can be sent to harlie-dev@furryhelix.co.uk.