Change of plans
So after some field tests with an early version of my erlang bot, I found some major problems in my idea.
Having the bot sit in a noisy IRC channel, just collecting words between 5 and 10 letters in length, and just counging occurrences, and then looking at the list, made me realize that the idea was lacking a lot of important things.
The word list ended up to be a lot of unimportant filler words like “rough, could, maybe, thought”, a bunch of nicknames, and even some web domains.
I could never worked out something interesting with this. Also relying on third party libraries to do certain tasks (like reading and answering through twitter), seemed to be risky, regarding my bad experience with various databases.
New oportunity
I’m glad that I hang out in this twitch.tv channel that happened to have some kind of arbitrary point system, everyone was asking for having a real bot for it. And as I already had a basic bot, knew how to do it, and wanted the bot to connect to twitch anyway, I figured I might have this little side-track and implement the point system.
So I did, rather quickly. And impressively with hot code reload, which I just love using erlang.
The basic point system turned out to be quite nice. I even added some more things like a highscore list, and a lookup and all the kink.
Now, as there is plenty of #learnbruary left, with a new open project, I just make this my main project for now, implementing interesting and useful options into the twitch IRC bot. Nothing new, but a learning experience anyway.