Reevaluate meeting times

[cjihrig]

We've recently had people step down from the TSC, and new members have joined. Let's revisit our meeting schedule to see if the current times are still best for the collective group.

[sam-github]

I'm in North America, west coast, so the 5am meeting doesn't work well for me, the other two work well.

[ChALkeR]

Based on current data, the timeslot difference would not be very large -- one timeslot would be shifted by one hour, one other would be shifted by one or three hours, one would remain the same (the timeslot at 16 UTC seems rather stable atm).

I'll post more info prior to the meeting next week or once the data gets updated.

[mcollina]

Maybe we can move to two timeslots instead?

[ChALkeR]

@mcollina That would be possible (e.g. 15, 20 seems viable atm), but it would be slightly less ideal (less attendance probability for some people, but 15 20 has just two people at 48% with everyone else above 50%).

What would be the reason for preferring two instead of three? Is it more convenient for everyone?

The problem with the two timeslots is that with two timeslots, there will certainly be at least two people who will never be at the same meeting together. Three timeslots kinda solves that, bumping the minimum interaction to 20% atm.

Currently, 15,16,20 looks like the best option for three slots and 15,20 looks like the best option for two.

Let's wait a bit for some more input, though =).

[mcollina]

Let’s keep the three slots then!

[ChALkeR]

Status update: after the recent changes in data, the optimal seems to now be 15, 16, 19.
Let's wait a bit more =).

[ChALkeR]

Unless the data changes, the proposal would be 15, 16, 19.
ping @nodejs/tsc @gabrielschulhof @Fishrock123 once again about the data =).

[ChALkeR]

No data updates over the week.

[ChALkeR]

No data updates since May 20.
Let's go with 15, 16, 19 =).

[ChALkeR]

@Trott @mhdawson @nodejs/tsc Do we still need to confirm those via vote, or can we start using those starting with the next meeting? Ref: #713 (comment)

[Trott]

@Trott @mhdawson @nodejs/tsc Do we still need to confirm those via vote, or can we start using those starting with the next meeting? Ref: #713 (comment)

5 upvotes and no other participation wouldn't be enough for a vote, but since there's no one opposed, it's enough for consensus. I say we call it done!

[targos]

+1. Do we start with 15 this week?

[mhdawson]

To avoid confusion I'd suggest we use the current scheduled time (9PM UTC) and I'll update for all future meetings.

[mhdawson]

Calendar updated for meetings started next week, closing issue.

[Fishrock123]

Updated mine but it did not change significantly.

[ChALkeR]

@Fishrock123 Thanks! It looks like that didn't change the optimum.

[ChALkeR]

The data was updated once more just now :-).

Two alternatives now:

times: [ 15, 16, 19 ]
  participation: { avg: .756, min: .39, stdev: .1584, fair: .9536, n5: 17 }
  percentages: 95,95,95,94,94,80,80,80,80,80,80,70,64,60,59,59,59,39
  interaction: { avg: .568, min: .0658, stdev: .1761, fair: .657, n5: 102 }
  meetings: { probs: [ .998, .996, 1 ], avg: [ 13.56, 13.12, 14.22 ] }
times: [ 15, 18 ]
  participation: { avg: .753, min: .484, stdev: .1736, fair: .9635, n5: 14 }
  percentages: 95,95,95,94,94,94,87,80,72,72,72,72,72,65,48,48,48,48
  interaction: { avg: .562, min: .019, stdev: .1975, fair: .533, n5: 82 }
  meetings: { probs: [ .998, .998 ], avg: [ 13.56, 13.58 ] }

I propose to keep the 3-slot variant (15/16/19).

[mhdawson]

+1 to sticking with 15/16/19. That is what the calendar now reflects.