Skip to content

TRAC.md

Latest commit

 

History

History
72 lines (52 loc) · 3.11 KB

TRAC.md

File metadata and controls

72 lines (52 loc) · 3.11 KB

TRAC - Bot::BasicBot::Pluggable::Module::Brownian

Issue, to-do, and wish list, tracking file for the Brownian development project.

ISSUES

  • (smpb, 25/11/2001) Brownian does not parse punctuation on the end of factoids correctly.

TO-DO

  • (smpb, 24/11/2011) Deprecate this file and move all of this to a proper platform. Say, Github.
  • (nfn, 25/11/2011) Bot::YadaYada::RSS is still spammy as hell on the logs. Fix this.

WISH LIST

  • (nfn, 11/05/2012) Implement a mechanism to store messages for users who are away. Use the "tell X about Y" interface and, if Y is not online, store the message and send it to her when she comes back on.

  • (nfn, 03/04/2012) Implement a URL collector module and give people access to it via the web.

  • (nfn, 03/04/2012) Add a twitter account to the bot and teach it to tweet when we ask it to and also on specific "special" occasions.

  • (nfn, 30/11/2011) Add an alarm/reminders module to the bot.

  • (smpb, 24/11/2011) The bot's very own internal karma meter:

    • He dislikes people that 'botslap' him often
    • He likes people that 'botsnack' him (but not too often!)
    • He "feels" a kinship to those who have taught him more
    • Boilerplate replies (like greetings) should vary according to these metrics
    • (nfn, 25/11/2011) The replies should be modulated by the user's recent interactions in the form of botslaps and botsnacks. In the extreme the bot should stop responding to the user after being heavily abused

  • (nfn, 25/11/2011) Implement a 'substitute' module that catches phrases like 's/this/that/' and replies with 'user meant "latest user's phrase with substitution'

  • (smpb, 25/11/2011) Implement NIALL, the Non-Intelligent Artificial Language Learner, within Brownian:

  • (smpb, 25/11/2011) Probability associated to replies

    • Brownian should recognize keywords in the middle of sentences (not just the beginning)
    • If the keyword is found on a very active conversation, Brownian's tendency to disrupt should be lower
    • Likelihood of reply should be high to someone recently joined or that hasn't spoken in a long time
    • Direct mentions ("brownian: blah") should have a reply rate of 100%

  • (smpb, 25/11/2011) Pipe unrecognized queries to a "magic 8-ball" or an engine powered by NIALL, removing lame stock replies like "Search me, bub."

SOLVED ISSUES

  • (smpb, 24/11/2011) At every single restart of the bot, the RSS module parses all the feeds again and floods the channel with notifications. -- Solved by (nfn, 25/11/2011)

  • (nfn, 25/11/2011) Implement a logging module. -- Solved by (nfn, 23/03/2012)

  • (nfn, 25/11/2011) Implement a 'backlog' module, allowing people to read the backlog for the last N [minutes|interactions] before they joined the channel. (See logging module) -- Solved by (nfn, 04/04/2012)