@fewerror is a Twitter (and Telegram) bot in the @StealthMountain genre. If you follow it, it will correct you when you say “less” but should have said “fewer”. It is 100% accurate all of the time.

Here's a piece reflecting on three years of @fewerror.

So much praise, it now lives in a separate file.

Thanks to the unstoppable @aparrish for pointing me in the direction of TextBlob in her post on the making of @VoynichTechNews.

This thing expects to fetch authorization credentials from environment variables. I source a file like this:

export CONSUMER_KEY="..."
export CONSUMER_SECRET="..."
export ACCESS_TOKEN="..."
export ACCESS_TOKEN_SECRET="..."

To get values for those variables, why not follow Allison Parrish's instructions for everywordbot? You might alternatively find get_oauth_token.py useful for ACCESS_TOKEN and ACCESS_TOKEN_SECRET.

It originally used the statuses/filter streaming API to receive all tweets containing the word less, and replied to one at most every two minutes. Unfortunately, it was quickly banned for “sending multiple unsolicited mentions to other users”. (I'm not sure how @StealthMountain escapes the same fate.)

These live in the cbdq directory.

• @gnuerror: an incoherent activist for an incoherent age.
• @xbotsdoney: Yet another sixpenny automaton, furnished promptly as a paean to the service which powers it.

Inspired by a series of books for young children. More details in this blog post. source, unnecessarily large grammar.