Respond with a newly generated tweet when @notsorrygame is mentioned on Twitter

[mogul]

We need to resolve how we actually respond to incoming tweets.

• How is our response triggered?
• Are tweets generated on the fly, or pulled from a pre-generated set?
• Do we log the request/response anywhere?
• Do we have to include the addressee's handle in length limits?

[jcombos]

These are good questions. Here are my thoughts:

Trigger: It's probably easiest for the bot to respond to anyone who
tweets "@notsorrygame" but if we wanted to limit it so that it doesn't
tweet at people just for mentioning it, we could try to limit the trigger
to tweeting "sorry" at @notsorrygame. Overall, I'm for the easiest option,
though.

Tweet generation: If it's fast, we could have the tweet generated on the
fly, but I could also see a daily generation of tweets that are pulled from
randomly each time one is needed. I generally favor the latter since it
could handle high traffic requests a little better, but I leave it to you
to determine what is the easiest to code.

Request/response log: I'm assuming that the tweet at the twitter handle
and it's response will both be logged like normal tweets in it's feed, so
people could always look at it read all the tweets its sent out. If that's
the case, would we need to log anything beyond that?

*Addressee handle: *Not only do I think we need to handle the addressee's
handle (so the tweet response goes to the right person) but we're also
going to need to set aside space for the sorry card generated too.

Speaking of which, any ideas on how to pre-pend the card info?

On Mon, Jul 6, 2015 at 12:51 PM, Bret Mogilefsky notifications@github.com
wrote:

We need to resolve how we actually respond to incoming tweets.

• How is our response triggered?
• Are tweets generated on the fly, or pulled from a pre-generated set?
• Do we log the request/response anywhere?
• Do we have to include the addressee's handle in length limits?

—
Reply to this email directly or view it on GitHub
#4.