Attempt to get "nobodys around" message to only affect newcomers

[jywarren]

This is not right... but it's a first stab!!! Help!

towards #108

[pdurbin]

@jywarren I really appreciate you jumping in here and trying to improve the situation. The idea of only showing the "nobody's around" message to newcomers sounds great but I'm confused about the "remember for 1000 seconds" concept. I guess I was thinking all of the non-newcomers would be stored in a csv file or whatever and they'd never see this message again. Probably my thinking is influenced by "nicks.csv" at https://github.com/shaunagm/WelcomeBot

Update: I just read through #108 (comment) about the second condition and I think it's making more sense. I need to ponder this a bit more. I still think I would just go for a file called regular.csv or whatever. This would be a list of non-newcomers for whom this "nobody's around" message is simply noise. I don't love how plosbot always insists on having the last word, no matter who talked last. It's troubling. 😄

[jywarren]

"remember for 1000 seconds" is because once someone's spoken up, how do we distinguish them from someone who's been around for a month or a year? I was thinking we could see if they've interacted in the chatroom longer than 1000 seconds ago -- i.e. about a half hour. If they have, we consider them a non-newcomer (the period could be longer, for sure).

This could be refactored into a longer term solution as part of #101, where we remember when someone joined -- but I'm looking at a short term solution here so we don't get blocked by #101.

[jywarren]

And if it helps to try to constrain the goal of the idle behavior here -- it's maybe narrowly to ensure that newcomers don't meet silence on their first interaction in the chatroom -- what do you think @ryzokuken ?

[ryzokuken]

@jywarren I think the general solution (one you've implemented) of show the message to everyone once, and never again sounds good as well.

[jywarren]

Would that work? I guess so! What about this fix, do you think you'd have some time to help solve this one? I'm not as familiar with the code as you... any pointers?

[ryzokuken]

@jywarren I'd love to. I remember doing a ton of refactoring work to take this one on. Let me take a stab at this.

[jywarren]

Hi @ryzokuken -- do you think the new MemoryBehavior can enable a relatively quick solution here? Thanks!

[ryzokuken]

It can be, actually. With a few changes, the existing behavior will work well enough.