FoFbot by jseakle

This is a Twitter bot that retweets friends of your friends that you
don't already follow.

I wrote it because I was inspired by @tinysubversions' many Twitter
bots to try one of my own, and it was the first toy idea I thought of
to learn the API with. The original is running at @fofbot.

It turned out to be really cool! It has found me a good handful of new
people to follow, and many, many cool links and conversations I never
would have seen otherwise.

It should be noted that following one is probably not a good idea if
you already spend too much time on Twitter.

INSTALLATION

This is a pretty straightforward ruby script that I run from a
cronjob. Any other kind of scheduler is obviously fine too, though if
you turn on any of the error reporting, you might want to do something
other than puts it to STDOUT if your scheduler doesn't mail that to
you like cron does.

Steps:

1. Install the 'twitter' ruby gem from https://github.com/sferik/twitter

2. Make a new Twitter account for your bot. This requires a new email
account.

3. Obtain your API credentials from dev.twitter.com by logging in and
clicking "My Applications" and plug them in to the strings at the top
of the script.

4. Obtain your bot's and your own twitter ids,
http://gettwitterid.com/ is a fine way to do that. You could also use
the gem. Plug them in to the indicated places in the script (your
bot's appears twice).

5. Uncomment any of the error reporting messages that you desire.

6. Set up a cron job or similar to run the script. Mine looks like:

    */30 * * * * ruby -W0 /home/jake/fofbot/fofbot.rb

which makes it run every half hour. Depending on how many people you
follow, and how many people they follow, it's conceivable that you
might run into the Twitter rate limit if you run this too frequently.

ISSUES

There are lots of dumb things about this script, because I hacked it
out quickly for myself. Feel free to send me pull requests, or even
just nag me for changes that would actually matter to you. Things that
stand out:

* All the user-specific data should be pulled out to a separate file
  so that a runnable script can be checked into this repo. If you plan
  to pull ever again, you should probably run an untracked copy to
  avoid this.

* The random choice of tweet should be done without replacement.

* All the numbers of retries are based on "empirical evidence", ie.,
  completely arbitrary.

Have fun, and let me know if you actually end up using this/doing
anything else cool with it!