Node.js Technical Steering Committee, 20 Dec 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2017-12-20 Node.js Foundation Technical Steering Committee (TSC) Meeting

Description

https://github.com/nodejs/TSC/issues/447

A

Welcome to the node.js technical steering committee meeting for December 20th 2017.

A

Does anybody have any announcements they wish to make before we get to the agenda.

B

I've sent a pull request on a sink aerators, so if you want to tinker with it and so on, please give us my friend some feedbacks. It was a long waited feature. So now that note cv8 6.3 landed, we can actually have them.

B

Lady I pass it in the chat and reach notes. That's just about it. Okay,.

A

I'm, sorry, you said it wasn't a sink iterators here. Yes,.

B

A sink I traitors PR, yes,.

A

A

The first issue is build two. This is in the node repository for request, one four, seven, nine five update the minimum kernel version, two three 10.0 I, don't think this is I, don't believe this is urgent and we were hoping to have either Rubin Bridgewater or Gibson Fahnestock for this or maybe both of them. Neither of them are here so I would propose that we table this for the next meeting. Does anyone feel otherwise.

C

I'm fine tabling this just one thing that I think as a whole, we probably need to figure out is our compiler support story and well build has partial ownership over this as far as like, what's living in our infrastructure, I think we can maybe be a little bit more intentional about how we want to support, compile of theirs, which compilers we want to support and how that may change over LTS and whether or not we're gonna maintain support.

C

I know this is gonna be important, especially when we think of it like our dependencies and what compilers dependency support. I think this is something that would be good to be intentional about.

A

Pool well I tried to capture that in a single sentence in the Google Doc. If you want to expand on it, if I didn't accurately represent what you stated, please go ahead and edit that any other comments on this before we move on to the next issue or something we need to discuss about this.

A

Cool all right, in that case, we'll leave this on the agenda for next week, and hopefully we can talk about it there, although, as always, if we can get it done and github before then even better okay. So the next issue is all the remaining issues are in the TSC repo.

A

The next issue is issue number 4 for 4 in the TSC repo, it is to propose a new TSE meeting schedule. The current proposal is to keep it the same as the existing schedule, except for this meeting. The one we're having now will take place three hours later and that will be a significant improvement for five TSC members, including a few members in underserved timezones.

A

China and eastern australia, it will be a little bit worse for some folks in Europe, including two people who are on the call today for Anzhi and the tale, but I would like to tweak the schedule this way, starting with with January, and we don't have enough people here too right now. Today we have nine I. Think CTC members on this call one two three four five, six, seven, eight, nine, yes and we would need 11 to to make this change.

A

However, if we can, if we can get as if we can get as many eyes and nays and abstentions as possible, already noted, that would be wonderful by me. Does anybody else have something they'd like to say about this.

C

So, just to verify that I'm, following everything quickly correctly, because it's moved back and forth we'd now be keeping all the same times, except for this time, which would move forward three hours. That's right! This.

A

Time will will be three hours later than it currently is so so, for you do instead of noon, it will be three p.m. that makes sense.

C

The the only thing that I'm curious about and I know that we're doing this algorithmically and programmatically would be the time that is noon.

C

Like 12 o'clock, UT's t I know that that's a meeting that tends to get not a lot of people showing up I recognize that this is because we're optimizing for people who generally have meetings in not great times but I was curious, if perhaps that time might be flexible, just like literally asking the five or six people who are in inconvenient time zones. If that is the most convenient time for them, rather than just sticking to what's in the spreadsheet yeah.

A

That's a possibility: I had I had some other thoughts along those lines about which actually I'm gonna take it back to github. But um but you know, if yeah yeah, actually your your your proposal is a good one. I'm going to stop talking I, don't need to talk about this you're you're you're right! That's a good idea! Yeah.

C

Just like the sort of like my major concern would be that, since all of us are being reflected in the algorithm that perhaps that other time is just still not fair, go.

A

Ahead be kids, are you going to say something? Yes,.

D

Commercially in the yeah.

D

Optimal times there do you see there itself so.

A

Someone or changes they'll give make do that.

D

In the airport.

A

Yeah I can do that. There's a there's, a gist where I have the code, but I could I could check it into the insert repository, maybe create I should perhaps create a tools directory in the TSE repo and put it there. Does that sound like a good place for it.

D

Sounds good to me.

C

We also have note for utils. It could potentially live in that as well.

A

Or for the dead air I'm trying to take some minutes while I'm doing this and I need to capture miles as thought, which was that.

A

A

To work, okay, okay is a oh.

A

And we have a few more attendees arriving. This is good. They we.

E

Now have 11 people.

A

So actually, in theory, we could pass this if everybody's in agreement with it, although I don't know, if that's actually the case, so could I guess. First of all, I'd like to call for a vote. Does anybody second that or object I.

F

Mean maybe they need to vote on it but sure.

A

Okay cool, so uh so everybody in favor of moving this meeting forward three hours. If you could +1 that in the chat and if you're opposed to it, if you could minus one it and if you really don't care either way and you abstain, 0 it that would be really awesome. Okay,.

C

Rich, quick, low and Gibson just thing me he's available now, if we want to go back to that issue, sure.

A

Why don't you send him the once you send him the link to the to the meeting got it well, thanks.

A

And again sorry everybody for the dead air, recording votes right now and I'll be done in just a second. Hopefully.

A

And then we will move on to the next dish because we do not have enough votes to pass. This.

A

One two: three: four: five: six, seven, eight nine, so we're missing two people missing me and we're missing: Trevor I'm, guessing Trevor joined late enough that he probably does not want to weigh in on this at this time. Is that accurate Trevor? Oh plus one hey there, we go awesome, okay, cool all right.

A

What that means! Take it back to the issue and we'll go from there. Gibson has joined us okay. So unless anybody objects to it, I'm gonna say: let's go back to the build issue. So so we can use Gibson's time effectively. um So we're going back now again, unless somebody objects', please speak up or say something: no Chet we're going back now to node issue, one four: seven: nine five, which is update the minimum kernel versions. Three, ten zero and Gibson you're prepared to.

A

G

Can give a quick, quick summary, although please go ahead, hey so, basically, we have yet like, since we, since the ids, merge and stuff, we haven't really ever gone through and said: when do we stop supporting things, and how do we tell people that before you know it's too late? So basically we still support a an older version of the kernel and we need to work out.

G

When is the right time to drop it, and, of course, given that we going to support an LTS version for three plus years, we need to work it out well in advance. So start this discussion for no date and basically the point was that there's still going to be quite a few platforms, not least well, six, which are still in support for the whole lifetime of no date, so it was kept. The question is now: when do we drop? Do we drop four No ten?

G

Basically, or do you want to consider dropping it before no 10 goes out so.

G

Dates no pen goes end-of-life in April, 20-21 and apparently central six goes is around till the end of 2020.

G

So correct me: if I'm wrong, anyone, possibly rod but I, think that means we'd have about half a year where central six is end of life and no twists are supporting it, which of course means that if there are problems, then you know we start to run out of ability to provision machines or tests or even build in some cases.

G

So the question is: when do we? When do we learn this and do we land it? 14-0 do.

H

We have any dependencies other than Allah Beauty that have a specific kernel version and that needs to be supported.

G

I'm not sure anyone else notes it particularly for mv8, obviously do, but once you know once we we don't tend to update after we start right, we don't have many native dependencies I'm, not aware of anything else. That's stated. Let's say that.

A

And is there I mean apologies for not backing it more carefully in the actual issue, but is there consensus on the build working group to make this change and the and the dispute is, is an external to the build working group.

G

I'm not sure I'd say this consensus. Rod is not.

A

G

Organizing just a bit okay, great I, think there's still a bit of uncertainty about how to organize this and the bill. Workgroup, isn't sure whether it should be making recommendations or weights making decisions and TSE should be ratifying or what I'm. Basically, it's all a bit up in the air yeah.

H

More question: if so, if Libby V is the main dependency that has a kernel that is set on a kernel version, why is it that node would need to use a newer kernel version.

G

It wouldn't necessarily so part of this is we can continue to build on an older version of the kernel than we say we support, because then, if you have to update we've already told you like when 10 goes, you know when we start running 10. We can start an old version and bump it up. As long as we are clear to people that this is what we say, you should be running on right.

G

So if we say don't use central 6 for no 10, we can still building it for a while, and then that gives people extra time to move off, but they know that they run an older version and updates might not continue to work, but even minor or patch updates. But.

H

That's that's where I'm getting confused, if the only the only component that has a strict kernel version is Libby V in Libby I mean unless we're planning on updating, P, V V right, so he's gonna be at 2.6 for forever. So what what do we gain by a so I think.

G

The things we tend to hit are more like errors when you new things, don't work anymore. So, like you, have to update GCC, and you know that needs a requires and you are the bugs and the see that people have hit.

G

There are problems. We just can't provision machines anymore, like to build things because their end of life, I think, is also a question of there's different scene, the minimum we could theoretically support and what we should recommend for people right. So what the thing this isn't necessarily everything you can use. So we have an experimental level. If you look at the supported platforms, this is Tier one which is you know. You should be completely safe, running this and there's an experimental version which is earlier versions of the kernel. Don't.

A

Want like Center US v at tier one.

H

Okay, so what I hear is that the the bumpkin internal version is almost like being used as a catch-all for wanting annual, newer Lipsy and that stuff, you said, I, don't understand why by each one and we dealt can't be dealt with independently, where we say: okay, now we're just going to support a newer, gilepsy or so not.

G

H

G

This, if we wanted to take- and you are so GC for nine- is what we say you- we need to build. No ode, 10, well, Nodine upwards, that doesn't you can't get that by default for 32-bit linux on central 6, which is the I think the only thing we have, that is still the oldest kernel version, which we need to build on, so you'd then have to either start compiling your own current or your own GCC or like doing other stuff they don't. They do tend to tie together.

G

What you can get on this OS is, you know which Jesus nips Tennessee with us, which CCC, which toolchains that makes sense.

F

To me to move it off of here, one support also I mean I, wonder if they shouldn't if this shouldn't actually be the build working groups decision, because I mean I feel like we're kind of at their mercy for doing this work anyway. So, if they're not able to maintain an older test machine because they don't have the time and we're kind of stuck with that anyway. So what's the point, that's making that decision necessarily right, yeah.

G

Maybe the answer is for the TRC to say: Pontifical workgroup. Unless you know and then, if people have objections, then we can kind of raise it but other than that yeah and maybe a general policy is what's really needed like okay. Well, when this go end-of-life, it goes into experimental and then we support as long as we can and then when it breaks, then we stop. You know, maybe that's the kind of thing we should be doing so. I can propose that if that makes sense.

A

It looks like that's getting support in the chat window on the right there.

A

G

A

Does that give you a clear path forward? Yes,.

G

Let's say yes well, I.

A

Have I have a favor to ask him really sorry: I forgot that I wasn't just supposed to be chairing the meeting, but also supposed to be taking notes. So I didn't take any notes during that conversation, so they diem yep. The Google Doc I would super appreciate it and I'm really sorry about that. Okay,.

G

Sounds good all.

A

Right anything else, Gibson or anything else, anybody else on this topic.

A

Great, let us move on.

A

So right next topic- and hopefully I'll- remember to take notes this time rather than make somebody else do. It is.

A

Tse issue number four for three mentorship initiative: ramping up. This is Michael Dawson's issue and he put it on the agenda and I propose tabling it until he's back, although it really just seems to sort of be informational, there's a link in that issue, and you can just follow that link so you're aware of what's going on.

A

Does anyone have? Does anyone want to talk about this, or should we table it for two or three weeks from now when Michael Dawson's back I have.

C

Some really quick stuff to add so I, don't know if anyone's had a chance to read through it. I took like a quick look at the process and talked to Dan Shaw about it as well. My thoughts on it right now we're mostly that the process seems really well designed and thought through. The person who put this together is someone who's been involved in outreach E for a while, which is one of the programs that we've done mentorship with the biggest thing to look at quickly is you you can see.

C

There's a graphic and kind of the process is talking about like which projects within node want to work in the mentorship program and identifying the mentees, so it kind of went through Dan and did an exercise. What that may look like and the main working group that was identified as good for working good for mentorship I thought was going to be benchmarking, but I.

C

Think for those of you who are in working groups that may want to have mentees, maybe streams for you Mateo or build for Gibson, who, still on the call it'd, probably be good to review. This document and see how it could be successful, I know originally as a TSC. One thing that, like I had suggested to us, was having a mentorship program where all new collaborators get mentors, and so that's like one example of idea, and then tea and mentors, and so just in general I have a feeling like as the TSE.

C

If we want this to be successful, we should make sure that we're engaging in this process early. That's.

A

I think I took better notes this time, which is to say that I actually took notes. um Any other comments on this, hey, I, think I, don't think, there's any reason to take it off the agenda. I, don't think we want to do that without Michael, saying okay anyway, since there's the one who put it on so we'll leave it on for next week and probably skip it then too, because I don't think Michael make it X with either.

A

But yes, let's look at that github issue and be aware of what's going on, um okay, the next issue. Well, there's the standing issue of strategic initiative updates, so I'm, just gonna read these off.

A

The moderation working group is still trying to work out how we get the necessary privileges in the near term to actually perform moderation, tasks and the in the github organization. So you can see the discussion in the node.js, slash admin, repo and issue 33 I'm likely to put that issue on the TSC agenda soon to try to get some definite resolution one way or the other okay, so Michael reports, Michael thousand reports that v8 is 6.4, is ready to land on master. It's still not part of the stable Channel scheduled for January.

A

Twenty third, two thousand 18 miles reports that document guidelines for account. Oh wait! No, that's not one miles reports, that's a whole other issue. Okay, any other strategic initiative comments or reports from anybody. I.

F

Think for under the streams thing, the the moving to or the allowing for a experimental work on async iterators, with streams kind of falls under that sort of broad thing, as that will be playing a future part in that development.

A

Thanks Jeremiah I'm, not sure I caught that or how to summarize it, but if you, if you could hop in the Google Doc I, just type the words Jeremiah : streams and if you could just sort of flesh that out that was they'll be great. Thank you very much anything else. Anybody about to take initiative updates, yeah, two quick ones from.

C

Me if you take a look at core quests under the TSC one, that's open right now is an initiative. Refactor I'm gonna, just notice this initiative wrong, so I fixed that is now properly spelled initiative refactor. It restructures the way that the initiatives are laid out, they're no longer late at a table, and it introduces a new concept of a stakeholder. The idea of a stakeholder being you know someone who is actively involved in an initiative who is not the champion.

C

This was an external request from someone who was specifically asking about how they better figure out. You know who are the stakeholders on specific initiatives like who are the people who speak with authority or, more so like who? Should they be reaching out to collaborate with and originally when the initiatives came out? I was thinking that maybe it didn't make sense for the only for the champions to only be TSE members and introducing the stakeholders here, you know, allow other collaborators and potentially non collaborators to also you know join.

C

So it was requested of me to create a better definition of what a champion and a stakeholder is as well as you know, just kind of following through on that. So just kind of to anyone in here who is a champion right now and if you have people who you think might be good stakeholders for the initiatives you're working on. Please add a comment and I'll add them to that update and then also you know when I update the copy in there with some definition of what a champion in a stakeholder is. Review of.

C

That would be really appreciated.

C

Well, thank you um and I guess sorry routes, one more thing quickly to the room. If anyone could answer, do you have a problem with stakeholders and not being committers? No.

A

I have no problem with that.

F

As long as they're doing something towards a and not just sitting there and like poking at people telling them what to do, mmhmm.

C

Yeah the only like 50/50 I had on that would be like I almost feel like. If we feel comfortable enough nominating someone to be a stakeholder, why wouldn't we want to nominate them to be a collaborator and committer.

B

It could be a path to that um I. Think sometime. Okay, let me make a very quick example: I would probably like Doug Wilson and somebody or somebody else from the Express team as a stakeholder in HD Beach, oh okay, and if we have that back in the time it would have been helped in some of the promise we ended up having later on when we ended up all the compatibility and stuff. So that's an exact example.

B

You know any stakeholder that don't want to be a collaborator, but still wants to be our stakeholder in a major feature, because for my way or another they depend on it.

C

Okay, excellent I'll work that into the definition and I think that makes a ton of sense. Thank you. Everyone.

F

So from the chat, since this is pretty relevant to this I think so densha asks do. We need to invite the individual who's having and a mentorship initiative to a tasting tse meeting for feedback on I guess, like updates I.

F

Think the question there is, if someone is, is leaving a strategic initiative. That's not a TSU! Remember what do we do? Do we invite them I.

A

Don't think it's I think I think the way these treaty initiatives are currently laid out. Champions need to be PSC members, but I could be wrong about that miles. Go ahead, yeah.

C

I was gonna, say that was something I was originally going to ask in the process. But currently all the people who have taken on champion roles have been TSE members to me. That seems like it's weird, because in the same way that being a stakeholder can be a path to being a collaborator, ice-t being a champion being a path to the TSE, but I'm not convinced right now that someone could be an effective champion without being a committer, because how are you actually moving the initiative forward?

C

You and I have somewhat of a similar inkling regarding stakeholders, but I don't want to block on that so I'm comfortable personally, with the champion being a non TFC member as long as they are a collaborator, but I think that that would have to be with the consensus of the TSE.

A

So a champion does not have to be someone doing work champion. Just has to be someone on the TSC who actually champions the initiative, and so if the TSE needs an update from someone who is not on the TSC, it is the champions responsibility to invite that person as an observer, I, don't I would opt to punt this to the Q&A segment.

A

At this point, we have a bunch of things still to go through and a private segment and I know that I know we're running out of time, but I also don't want to I wanted to be mindful of the time, but I also want to squash conversation. So because of the time, let's.

C

Kick this to the issue, I think we've talked about enough enough about it today, github.

A

Seems like a good place for it? Okay, so there are there's. A next issue is TSE issued three one, one which is document guidelines for accountability, expectations for all members of the TSE and Kham Kham, my levels, you put it on the agenda. Do you want to let us know what sort of decision or action you're looking for here.

C

Yes, so we originally created this as a way of thinking about increasing accountability. There has been some work done for this as far as like in the admin repo, we now have remember expectations doc, but I don't feel like this was fully followed through, as I believe. Part of what we wanted to see done here was improved moderation.

C

Guidelines for for leadership, which we don't currently have. That was something that was potentially gonna, like I thought was being tasked to the moderation group, but they have not. The moderation team is not taking this up, so I thought that it would be a good time for us to revisit, and you know, come up with what are the guidelines of accountability for leadership and in our team?

C

Are we satisfied with where we are right now and if not, but what do we need to do and who are going to be the stakeholders we're going to move that forward?.

A

Or use this, are you? Are you able to means there's some? So this sounds like a discussion. Is there? Is there? Is there a decision or an action here or.

C

So I think I think the decision or action that we need to do is a committee right now is just if people feel like we're good, then maybe there isn't more work to be done. I personally am pushing that. We need to revisit this and specifically around the moderation and so I think what we need to do as a committee. The actionable bit for today is decide. Is this something that we want to officially task?

C

The moderation came with and if we're not officially tasking the matter, the moderation team with it that I would like to see us more actively pursuing this in the new year. Okay,.

A

So, on behalf of moderation, since um let me see here I'm just saying if anybody in the moderation team is chiming in, but not in our slack okay, so in general the moderation team does not see itself as a group that creates policy.

A

That said, a lot of members are very happy to to recommend policy either as individual more as individuals and as a group. So what I can do is I can take this back to the moderation group and see if there's something it see if it's something they want to take up or not that's what I can offer yep.

C

That works for me and and maybe I can just pose this to the TSA as well and I know that I've been putting it like. I was under the impression when we spun up the moderation team that this particular thing and recommending policy around it was one of the most important things that the moderation team was supposed to be doing, and so perhaps I was mistaken and if other people think I'm mistaken that's great. But if other people were under the same impression.

C

This would be a good time to kind of chime in and maybe let me know.

F

That seems was taking to me because I am quite under the impression the moderation team is not there to particularly make policy, except for like internally, so that they can deal with issues that rather actually deal with issues that we already have. Policies for okay, I mean maybe I'm wrong there too, but that was my understanding.

C

If that's the case, then I think that the TSE needs to pick this up as an initiative rich as the moderation champion. I would like to ask if this is something that you can follow up on. I.

A

Yes, yes, yeah, there's multiple prongs here, there's like figuring out if the TSA thinks were we're good, as is cetera et cetera, but yeah I will I will take that on I'll leave this on the agenda and I'll I'll try to at least summarize something that will be either actionable or just here's some information, everybody and go from there and more.

C

Than happy to be a stakeholder and help drive this initiative.

A

All right sounds good. Any other comments on this topic before we move on to the last agenda item.

A

Now that I'm unmuted I'm gonna say that again, um okay, great uh last issue is when I put on the agenda and it's TS CSU number 284, not all members of the github or are members of the members team. If you look at that issue, Michael Rogers left a comment a long time ago explaining the members team and based on his comment, I believe we should remove people from the members team if they are not in any other team.

A

I'm gonna do that unless there are objections, although I won't do it until the community committee has has time to weigh- and has you know to put this put this on their meeting agenda- I itemize, some reasons in the in the issue. Why we shouldn't be shy about this? It's super easy to revert.

A

You can add them again. We want to lock down the members team a little bit because it gives access to the moderation, repo and some other stuff like that. So if anyone has concerns or objections, now would be a very time to raise them or drop a note in that github issue and I'll stop talking so someone less talkative. The me can weigh in if that's something.

C

I'm not sure and less talkative than you. This is something the automation team could maybe write about, for they could crawl all of our teams, make a list and then give us a list of all the people who are a members that are not in any other teams.

C

That seems like a useful tool that we could have and something that we could just do as a periodic audit.

A

I, imagine that something they could do.

E

Instead of having something regular, could we do it once and then make the members team top-level team, which is the parent of all the other teams, so everyone will inherit the membership.

E

Then nobody would be directly member of the team, but just yeah member of it because they are member of other teams loved.

C

It I'm a huge fan of that seems like a very strategic initiative. Oh.

D

Look these certain members.

A

I'm, sorry Nikita, can you say that again.

D

The decision is I think 16 of membership holding Network music stellar. Member of some other teams.

A

I'm not sure I, understand the question. I'm really sorry.

D

H

F

And Soviets suggested from the YouTube chat that I get a bat with Probot could probably monitor the mentorship continuously. But it sounds like we won't really need that.

A

Yeah I think Nikita is gonna, be typing up the question in the chat or something like that. But while we're waiting for that.

A

Perhaps it's time to check out what specially, since there's only 10 minutes left check to see if there's any questions QA in the IRC or youtube chats or anywhere the Twitter's or anywhere else that we tend to take questions from while we're looking for those I'll mention that the the upcoming meetings for the entire project can be found in noches org slash calendar tomorrow there is a nodejs and API team team meeting.

A

They apparently meet weekly at the exact same time, then there's nothing until Tuesday when there will be a build working group meeting and you can check the again check. Node org slash calendar to get the time in your local timezone, of course, with the time zone change, it might not be Tuesday, it might be Monday, but you know Wednesday or something but there's also a Diagnostics workgroup meeting a week from today. The TSE is meeting again a week from today and with that I'll stop talking and see.

A

If we have any questions, you have any questions.

F

There is a reasonably large discussion.

F

About the TSC and moderation working group from Disha and tyranny, but I'm not really sure. If there's a question in here or something we can really address,.

A

I'm gonna go back to Nikita's question, says: I was asking how that would work with outside collaborators.

A

um You know I, don't know if there are outside collaborators in the members group, because then there would be up there.

D

A

F

D

They membership members.

A

I'm so I'm really interested in this discussion. You mentioned Jeremiah, but let's, let's not try to I, mean.

F

You probably look at it yourself if you went to the YouTube livestream.

F

Give me a second, so tyranny asked if Forest and Tracy could come to the next meeting or forest end or I. Think Tracy come to the next meeting to discuss the moderation working groups and if this, but there is there's a meeting already, which tisha pointed out from, was that October 25th I think T's T, meaning where, for us in a comprehensive overview of the moderation working with it already includes not.

F

So if anyone wants to check that out, that's probably already described in detail there.

F

Another not there's one comment, which is moderation. Policy, like governance, is one of those hard people issues that seems the constant consistently falling off the TSC sphere of interest, which is probably kinda true.

A

Okay, I move that we adjourn this meeting and move to the private segment and the any objections gonna get a second.

A

F