You will need to update your infobot.config and infobot.users. See the example files. -- This requires perl 5. You should be able to start up just by running infobot. If you are using macperl, you will (currently) have to make one minor change (because $^O didn't work for me under os8). The infobot uses parameter files, typically in the params directory, to set up with. It treats anything on the command line as a parameter file and tries to load it. If you are using macperl, you will want to set the value of the macperl parameter to 1 in the two given param files (in the 'files' dir). By default, the infobot uses the IRC setup. This may change. NOTE that the irc version has no output by default; you'll have to turn up the debug level to get more. to start up the infobot, just invoke it from the command line. infobot --- A note on forms: you can prepend the tag to values in the db to just get a reply with no extra info. e.g. x => y then when you ask 'x?' it will just reply 'y' instead of something like "i think x is y". --- * extending the bot by adding your own code try to keep your changes inside src/myRoutines.pl so that you can easily just replace this file when there are new revs. this is called just after some of the normalization stuff in urlProcess.pl. take a look at the file for details. basically, if myRoutines returns non-null, it's taken to have 'caught' the event. you can just return '' to let the rest of the processing go at it. --- * update_db & dump_db update_db is a little perl script that will take a flat ascii file and make an infobot-style db out of it (currently just a couple of dbm files). You'll want to make 2, even if theye contain only 1 element each. it will also simply add (and overwrite entries in) existing dbs. This is especially nice if you don't allow updates to the databases through IRC and just want a collection of permanent factoids. update_db where is an ascii file like (in the case of an url-style infobot): infobot => at http://www.cs.cmu.edu/~lenzo/hocus.html one key => value per line. In the current setup, you need an is-database and an are-database, mainly for legacy reasons about representing plurality and being able to give the proper form. in the infobot-is.txt file and infobot-are.txt files you have examples. these are a fine starting point: 1.1 update_db infobot-is.txt infobot-is update_db infobot-are.txt infobot-are 2. Now you need to edit infobot to set up your bot. Don't forget to set the path to perl properly at the top and make it executable. 3. Then try running infobot. At present, there's a bunch of VERY BAD code in it, so don't use the -w switch unless you want to fix a bunch of things and mail me. Eventually, kill it and then you'll probably want to crontab it. included is a sample crontab and the script that you will need to edit. dump_db will just make a flat ascii file out of the named db, e.g. dump_db infobot-is Both update_db and dump_db take an optional switch, -m, which tells them what DBM module to use. You'll need to specify this if you set DBMModule in your config file in order to get the bot to use something other than Perl's default. Eg, update_db -m DB_File infobot-is.txt infobot-is dump_db -m DB_File infobot-is good luck, and mail me! kevin lenzo@cs.cmu.edu ps - i am just releasing this _now_ instead of waiting to fix everything. If you use this and like it, or even if you don't, please mail me! --- thanks to: You, for getting this and using this. Especially if you mail me and let me put you on the mailing list. lenzo@cs.cmu.edu special thanks to: steve orens (sorens) for being a tour-de-force beta bomber yo for working with script. this is a big one! amug and everyone there for hosting the undernet url #macintosh for dealing with url through his troubled childhood tris for being an early guinea pig jadin for pointing out the @verb bug... fixed in 0.17b chucky burnett for tons of stuff