Review meeting date times ?

[mhdawson]

May be time to review the day/times for the TSC meetings. #808 (comment)

@targos sounds like Wednesday is not good for you at all. @nodejs/tsc are people open to looking to move to another day? From past conversations I think we'd tried to keep all instances on the same day is that still the case or should we consider different days in addition to different times?

[ChALkeR]

Everyone, please also update the valid at column even if your timeslots were not updated so that we can tell how actual the data is =).

[mhdawson]

I wonder if we should also discuss a different day based on the comment by @targos. Possibly a timeslot on Thursday?

[gabrielschulhof]

Thursdays would work better for me as well.

[mcollina]

I'm good on both Wednesdays and Thursdays.
I would prefer to avoid Mondays, Tuesdays and Fridays.

[targos]

Thursdays work for me. If we do multiple days we'll need multiple spreadsheets

[Trott]

FWIW, Wednesdays are much better than Tuesdays or Thursdays for me.

[mhdawson]

I've added another sheet for Thursday. I have a feeling we are going to need to use 2 days.

[ChALkeR]

Given the current data, it's likely that we will have either 2 or 3 (out of 3) timeslots on Thursdays instead of Wednesdays.
We don't have about half of the data confirmed yet. @nodejs/tsc ping =).

[mhdawson]

@ChALkeR at this point we probably have the data we are going to get. Can you pull generate a recommendation from the data?

[ChALkeR]

@mhdawson data updated (on the third page of the sheet).

Seems like three main alternatives (times not mentioned on a purpose -- those are present in the table, but I don't want the specific timeslots to affect the generic decision we make here).

Here, percentages are expected participation of each person, in %.

Option 1. Three timeslots (Wed/Thu)

  participation: { avg: .694, min: .39, stdev: .1527, fair: .9549, n5: 17 }
  percentages: 94,93,93,93,89,70,68,64,64,64,63,63,63,63,63,63,63,48,39
  interaction: { avg: .485, min: .0757, stdev: .1724, fair: .597, n5: 87 }
  meetings: { probs: [ .958, 1, 1 ], avg: [ 11.58, 14.21, 14.27 ] }

One person at ~40%, one at ~50%, 10 at or above ~60%, 2 at or above ~70%, 5 at or above ~90%.
Average ~69%.
Interaction at ~49% average.

The downside here is that it brings one person significantly below 50% (to 40%) participation.

Option 2. Two timeslots (Wed/Thu)

  participation: { avg: .764, min: .48, stdev: .219, fair: .9764, n5: 12 }
  percentages: 95,95,95,95,95,95,94,94,94,94,94,72,48,48,48,48,48,48,48
  interaction: { avg: .577, min: .0189, stdev: .255, fair: .438, n5: 66 }
  meetings: { probs: [ 1, 1 ], avg: [ 14.78, 14.27 ] }

7 persons at ~50%, one at ~70%, everyone else (11) at above 90%.
Average ~76%.
Interaction at ~58% average, but particular probabilities are decreased.

The downside here is interaction form -- in particular, there are going to be more pairs of people who are very unlikely to ever be present on the same call.
I.e. it's more segmented.

Option 3. Two timeslots (Thursday only)

  participation: { avg: .749, min: .48, stdev: .214, fair: .9763, n5: 13 }
  percentages: 95,95,95,95,95,95,94,94,94,94,72,57,57,48,48,48,48,48,48
  interaction: { avg: .554, min: .0189, stdev: .247, fair: .452, n5: 76 }
  meetings: { probs: [ 1, 1 ], avg: [ 14.21, 14.27 ] }

6 persons at ~50%, two at ~60%, one at ~70%, everyone else (10) at above 90%.
Average ~75%.
Interaction at ~55% average, but particular probabilities are decreased.

Has the same problem as Option 2 -- there are 6 people each of whom would be able to attend only one meetings, and they segment into two groups, so people from different groups won't be ever present on the same meeting.

[ChALkeR]

@mhdawson One more note: it seems that not everybody confirmed Thursdays yet, and I see at least one person whose data has the potential of shifting the optimal timeslots significantly there.

[Trott]

We should probably add availability data for @codebytere and @mmarchini and see if anything changes. (The other two current nominees, Ruben and Gireesh, are already in the spreadsheet. And the recent emeritus, Sakthipriyan, has been removed from the spreadsheet. So the spreadsheet is up-to-date otherwise.)

[Trott]

We should probably add availability data for @codebytere and @mmarchini and see if anything changes.

I'll email them the link along with the "please do not share this information" caveat.

[Trott]

Also, daylight savings time change happens in the U.S. this weekend, so everyone here will probably have to update their data again. Whee!