Roughly in order of precedence (the closer to the front of the file, the more likely it'll be done before next release): * URL snarfers need to run one at a time, in order for the _snarfed queue to work properly. * We could use some better log formatting -- there's no reason "supybot:" needs to be there, especially when we've disabled plugin-specific logfiles. * Karma needs to use invalidCommand so it can be nested. * We need to store the key for channels in the registry and have an easy way for plugins to send joins to channels based on their registry configuration. * We need to note when bans expire on a channel and send the unban, that way plugins can use the ircdb ban stuff and not worry about sending or scheduling unbans. * Filters need to be added to the logs during testing so many logs (useless during testing) won't occur. * We should probably add a "hello" command to make things more compatible with Eggdrop, since we've been replacing many eggdrop bots lately. * We should be able to set the log level for plugins individually. * Rbot has a "remind" plugin that seems pretty cool, we should get one as well. It might do well as a command in the Later plugin. * MozBot has a "list" plugin that seems pretty cool, we should get one as well. Problems that involve a lot of tedious minutiae, but really need to be done at some point: * ircutils.{isChannel,isNick,toLower} should be static methods on each Irc object, since they can vary per-network. That's not the tedious part; the tedious part is having every other place that uses them use those methods instead. Hard problems that won't get done until someone really wants to have some fun: * Redundant relaying -- having more than one bot handle relaying, where a secondary one will automatically take over if the first becomes incapacitated. * Have the bot detect when other bots are in the channel responding to the same prefix character, and do something based on that. * Unicode support. Basically, we'd have to have a way to set the encoding for the server, and use unicode throughout internally. The real issue is how much it would affect the current code to switch it over to unicode, and what kind of burden it would put on plugin authors to deal with that issue.