AgileVentures / Paironauts

To find a pair partner any time!
MIT License
7 stars 5 forks source link

Pairing lobby - Don't merge #15

Closed yakryder closed 6 years ago

yakryder commented 6 years ago

Just want to open this up for visibility.

yakryder commented 6 years ago

@bcb37 I've thought about it some more and I don't think we need nameable users in order to make the first iteration acceptance test of users seeing other users pass.

"A user in the pairing lobby can see other users in the pairing lobby": In its rawest most unrefined sense it means I can see the other individuals who are also in the pairing lobby, not that I can identify them beyond their being distinct from each other. It can't just be binary -- yes there are other users, or no I'm alone. But I don't need to know their names. It is sufficient that the other users are countable, which depends in turn on their being distinguishable from each other. Which they already are through our Presence implementation.

Rather than trying to shoehorn user naming into this feature, can we wait to do that until we actually add fully-fledged users?

I would also happily hear the counterargument that a pairing lobby without a tangible notion of a user is building with one of the fundamental building blocks missing.

yakryder commented 6 years ago

digression

For communication purposes in a single lobby session, it's also very important that the identity of users present in the lobby not change. We already have that sort of covered -- we can assign them their phoenix_ref as name, or we assign them a random non-duplicable name. If we wanted to ensure that users have both human-readable names and they never experience the unlikely [given a random name sampling of sufficient size] experience of having someone named Mary leave and another distinct someone named Mary enter right after, we could write some logic keeping track of the unique users seen by all the users present in the channel and that we never pull a name from that list unless it is the same participant coming back (probably identified by IP). Though that seems torturously overblown.

There is something very intuitively appealing about a no-sign up service that anonymizes everyone but then comes to a collaborative and fruitful end. The ways of seeing that positively probably diverge sufficiently from the perspectives of the internetz that we should really just have real users, but I like the idea of it.

/digression

yakryder commented 6 years ago

Stale

yakryder commented 6 years ago

To be crystal, we want and will use this functionality. But I don't want stale PRs on the books and I don't think it's ready to be merged.