Stream: project leads (public)

Topic: roadmap in 2020


nikomatsakis (Jan 14 2020 at 14:40, on Zulip):

So I really want to get this roadmap RFC up and done. =)

hey @Eric Huss, I saw the cargo in 2020 post, and it got me thinking. Well, first, it's a good read, and it got me wondering what the role of the roadmap should be. I think it shouldn't just be each team "stapling together" their plans -- more like identifying themes and goals that the teams should interpret in their own ways.

nikomatsakis (Jan 14 2020 at 14:41, on Zulip):

Toying with that I started reworking the draft points I had before

nikomatsakis (Jan 14 2020 at 14:42, on Zulip):

So far I have these, though I think the last few are not right (too narrow)

nikomatsakis (Jan 14 2020 at 14:42, on Zulip):

Notably, "trait system" isn't really right I think -- there's something else I'm looking for -- maybe "fundamentals"?

simulacrum (Jan 14 2020 at 14:42, on Zulip):

I think IDE falls as a subbullet of "Follow-up on in-progress designs and efforts"

nikomatsakis (Jan 14 2020 at 14:42, on Zulip):

A lot of things do,

nikomatsakis (Jan 14 2020 at 14:42, on Zulip):

which is ok :)

nikomatsakis (Jan 14 2020 at 14:43, on Zulip):

but I was hoping to add a few more themes, the idea being

nikomatsakis (Jan 14 2020 at 14:43, on Zulip):

to help in categoring which in-progress efforts...

simulacrum (Jan 14 2020 at 14:43, on Zulip):

hm, sure, that makes sense, though perhaps doesn't quite jive well with the main roadmap

simulacrum (Jan 14 2020 at 14:43, on Zulip):

e.g. it may be worth leaving the which to the individual teams

nikomatsakis (Jan 14 2020 at 14:43, on Zulip):

to be clear, I agree that bullet should be removed

nikomatsakis (Jan 14 2020 at 14:44, on Zulip):

probably the last two bullets, maybe even Async I/O

nikomatsakis (Jan 14 2020 at 14:44, on Zulip):

but I would like to add a few other themes that replace them I guess is what I mean

nikomatsakis (Jan 14 2020 at 14:44, on Zulip):

but maybe we don't need to, I don't know

nikomatsakis (Jan 14 2020 at 14:44, on Zulip):
nikomatsakis (Jan 14 2020 at 14:45, on Zulip):

that feels pretty vague

nikomatsakis (Jan 14 2020 at 14:45, on Zulip):

still, I think the 2nd bullet is very on point

simulacrum (Jan 14 2020 at 14:45, on Zulip):

I think that may not be a bad thing

simulacrum (Jan 14 2020 at 14:45, on Zulip):

but also -- it feels pretty concrete

nikomatsakis (Jan 14 2020 at 14:45, on Zulip):

it's kind of like "ok, we put up with various problems because we wanted to focus on fundamentals, but we need to come back to those 'intermediate' problems"

nikomatsakis (Jan 14 2020 at 14:45, on Zulip):

things that develop at scale

simulacrum (Jan 14 2020 at 14:45, on Zulip):

not in what we do, but why we want them

nikomatsakis (Jan 14 2020 at 14:45, on Zulip):

I think this fits a lot of my intuitions right now

nikomatsakis (Jan 14 2020 at 14:46, on Zulip):

and notably I think one of the challenges about those 'intermediate' problems is that sometimes they don't impact a lot of people, but those people who they impact, they impact a lot

nikomatsakis (Jan 14 2020 at 14:46, on Zulip):

maybe there's a bullet to be crafted around that?

simulacrum (Jan 14 2020 at 14:46, on Zulip):

hm, maybe

nikomatsakis (Jan 14 2020 at 14:46, on Zulip):

cargo features kind of feel like this --

nikomatsakis (Jan 14 2020 at 14:47, on Zulip):

we've got the basic feature, we've gained experience, but we know that there are various problems that arise in practice

simulacrum (Jan 14 2020 at 14:47, on Zulip):

refocusing on the "5%" so to speak?

nikomatsakis (Jan 14 2020 at 14:47, on Zulip):

I guess that's a form of "polish"

simulacrum (Jan 14 2020 at 14:47, on Zulip):

i.e., instead of expanding the 95%

nikomatsakis (Jan 14 2020 at 14:47, on Zulip):

refocusing on the "5%" so to speak?

to a certain extent :) I'm not sure I'd phrase it quite like that

simulacrum (Jan 14 2020 at 14:47, on Zulip):

Sure, yeah

nikomatsakis (Jan 14 2020 at 14:47, on Zulip):

I guess that what "polish" and this "fix smaller problems" have in common is

nikomatsakis (Jan 14 2020 at 14:47, on Zulip):

sometimes it's a laundry list of things

nikomatsakis (Jan 14 2020 at 14:48, on Zulip):

no one of which is the "key" thing

nikomatsakis (Jan 14 2020 at 14:48, on Zulip):

I guess two concrete examples are #project-inline-asm and #project-safe-transmute

nikomatsakis (Jan 14 2020 at 14:48, on Zulip):

that sort of fit this to my mind

nikomatsakis (Jan 14 2020 at 14:49, on Zulip):

both I think also fit the "in progress" design bullet

nikomatsakis (Jan 14 2020 at 14:49, on Zulip):

but they also fix real problems, esp. problems that production users hit, and/or "systems problems"

nikomatsakis (Jan 14 2020 at 14:49, on Zulip):

(sorry, thinking out loud a bit)

nikomatsakis (Jan 14 2020 at 14:49, on Zulip):

anyway maybe I should just run with those 3 bullets to start, I agree they're sort of easy to defend

nikomatsakis (Jan 14 2020 at 14:50, on Zulip):

I had this text:

nikomatsakis (Jan 14 2020 at 14:50, on Zulip):

Structure of the roadmap

The roadmap this year is based around a few central goals. These goals
are intentionally rather broad -- they are meant to be interpreted
throughout the year by the various teams, as they make decisions about
what to pursue. In some cases, we have identified and called out
particular areas of focus, such as the 2021 Roadmap and Async I/O, but
in others we have left the choice rather broad.

simulacrum (Jan 14 2020 at 14:51, on Zulip):

I think running with the first three bullets is the right move

nikomatsakis (Jan 14 2020 at 14:51, on Zulip):

I am increasingly of this mind

nikomatsakis (Jan 14 2020 at 14:51, on Zulip):

maybe I just needed someone to push me over the edge :)

nikomatsakis (Jan 14 2020 at 14:51, on Zulip):

"less is more"..

nikomatsakis (Jan 14 2020 at 14:51, on Zulip):

plus I dont' want to stall on this forever

simulacrum (Jan 14 2020 at 14:51, on Zulip):

yeah, I think we need to get something out in the next two weeks

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

I have a 'self goal' to have something open by end of week

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

so I can focus on the things I really want to be doing :P

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

side note that I want to revisit the roadmap procss

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

this year doesn't feel very good at all

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

one thing I think might be useful

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

that I saw in prior roadmaps

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

maybe I'll try to emulate it

nikomatsakis (Jan 14 2020 at 14:52, on Zulip):

is to lay out the course of the year

nikomatsakis (Jan 14 2020 at 14:53, on Zulip):

and in particular I'd like us to start thinking about when we will be making rust 2021 decisions

nikomatsakis (Jan 14 2020 at 14:53, on Zulip):

and things like "start survey"

nikomatsakis (Jan 14 2020 at 14:53, on Zulip):

I think this year we were all kind of adapting and it was hard, but we can do better

simulacrum (Jan 14 2020 at 14:53, on Zulip):

yeah -- I think for a lot of those decisions it would be great to put them on the calendar

simulacrum (Jan 14 2020 at 14:53, on Zulip):

especially because we need to line up resources

nikomatsakis (Jan 14 2020 at 14:53, on Zulip):

OK, thanks, that was helpful

simulacrum (Jan 14 2020 at 14:54, on Zulip):

I would be happy to help with the draft, too, fwiw

simulacrum (Jan 14 2020 at 14:54, on Zulip):

I probably don't have the energy to create it

simulacrum (Jan 14 2020 at 14:54, on Zulip):

but fleshing things out, yes :)

nikomatsakis (Jan 14 2020 at 14:54, on Zulip):

OK, I'm going to push it to my repo, I'll add you as a collab perhaps?

nikomatsakis (Jan 14 2020 at 14:54, on Zulip):

or could use hackmd

nikomatsakis (Jan 14 2020 at 14:54, on Zulip):

I can't decide :)

nikomatsakis (Jan 14 2020 at 14:54, on Zulip):

I like being able to edit in emacs

nikomatsakis (Jan 14 2020 at 14:54, on Zulip):

and/or go offline

nikomatsakis (Jan 14 2020 at 14:54, on Zulip):

so I guess I'll go w/ repo :)

simulacrum (Jan 14 2020 at 14:54, on Zulip):

repo is fine by me :)

nikomatsakis (Jan 14 2020 at 14:59, on Zulip):

Reading back over the text, one thing I think is "missing" here a bit (and maybe that's fine, it can come up in follow-up blog posts or other things) are any hint of specific areas, such as error handling, Async I/O, traits, where we expect to focus. That might also make some of the text harder to write. Well, I guess we'll see.

nikomatsakis (Jan 14 2020 at 15:00, on Zulip):

anyway I threw up a super rough draft in the roadmap-2020 branch on my repository

simulacrum (Jan 14 2020 at 20:01, on Zulip):

I am going to try to spend at least an hour in a bit hacking on that

Kyle Strand (Jan 14 2020 at 20:25, on Zulip):

I think it would be reasonable to list the specific technical areas of focus in the "in-progress efforts" section.

Kyle Strand (Jan 14 2020 at 20:27, on Zulip):

One oddity is that the three items seem to be in entirely different categories. I almost wonder if "governance" and "technical" are more like two separate dimensions/domains of focus that the project must have every year, and the road map should just acknowledge that before specifying the specific concerns _within_ each domain to highlight in the road map.

Kyle Strand (Jan 14 2020 at 20:31, on Zulip):

Or, perhaps, the three dimensions, listed in the same order as your bullet points above, would be:

simulacrum (Jan 14 2020 at 20:53, on Zulip):

I think I explicitly don't want to talk about the specific things we aim to focus on

simulacrum (Jan 14 2020 at 20:58, on Zulip):

@nikomatsakis I pushed up a few paragraphs on the 2 bullet points you didn't have anything for yet -- I feel like I want to add something about the themes but am not feeling particularly in the flow so might come back to it

nikomatsakis (Jan 14 2020 at 23:09, on Zulip):

@simulacrum thanks, I made a few more edits, but I have to run now; I too am 'not feeling it' right now. I'd like to start weaving in quotes and things from people's posts, for example, but .. yeah.

simulacrum (Jan 14 2020 at 23:09, on Zulip):

mhm

nikomatsakis (Jan 15 2020 at 15:02, on Zulip):

@simulacrum I've been doing some edits this morning. I pushed some text around the organization section, take a look. I'm having "mixed feels" about the structure -- it feels like a bit of an abdication, I guess, of an "opportunity" to narrow our focus a bit more, perhaps?

That said, one of the things I want to do also is to start looking at the survey a bit more (have to figure out the best way to be publishing that "in-progress stuff"), because I'm wondering if looking through it there might be some hints at what else we can say.

I guess it feels like we should be able to give some guidance to help people select which in-progress efforts to prioritize =)

simulacrum (Jan 15 2020 at 15:03, on Zulip):

Hm yeah

simulacrum (Jan 15 2020 at 15:03, on Zulip):

Let me take a look

nikomatsakis (Jan 15 2020 at 15:03, on Zulip):

I also included a yearly calendar

simulacrum (Jan 15 2020 at 15:03, on Zulip):

I agree that it seems good to be more specific

simulacrum (Jan 15 2020 at 15:04, on Zulip):

But I'm not sure how to do that without getting too fine grained

nikomatsakis (Jan 15 2020 at 15:05, on Zulip):

Yes, I know.

nikomatsakis (Jan 15 2020 at 15:05, on Zulip):

well one thing I was thinking

nikomatsakis (Jan 15 2020 at 15:05, on Zulip):

first off, I was re-reading @boats gem of a post on "Organizational Debt"

nikomatsakis (Jan 15 2020 at 15:06, on Zulip):

in any case I was thinking that the "improve organization" section might be easily expanded to some separate bullets

nikomatsakis (Jan 15 2020 at 15:06, on Zulip):

I really like the idea of trying to keep these goals general and more of "problem statements"

nikomatsakis (Jan 15 2020 at 15:06, on Zulip):

regarding the more technical things, this partly why I wanted to look into survey

nikomatsakis (Jan 15 2020 at 15:07, on Zulip):

Some initial notes from skimming what @Jasun Carr has done so far:

nikomatsakis (Jan 15 2020 at 15:07, on Zulip):
nikomatsakis (Jan 15 2020 at 15:08, on Zulip):
nikomatsakis (Jan 15 2020 at 15:08, on Zulip):
nikomatsakis (Jan 15 2020 at 15:08, on Zulip):
nikomatsakis (Jan 15 2020 at 15:09, on Zulip):
simulacrum (Jan 15 2020 at 15:11, on Zulip):

I think the first bullet is a problem we're working towards. I see the website internationalization as a stepping stone towards it, and more broadly as we work to make the teams better suited to distributed work and so forth

nikomatsakis (Jan 15 2020 at 15:11, on Zulip):

( Personally, I feel like compilation times / IDEs / error-handling / async remain super important themes, with another being kind of long tail of "systems features". The last one I am having a bit of trouble quantifying, except that I think it has come up repeatedly in terms of things like adoption from companies. )

nikomatsakis (Jan 15 2020 at 15:11, on Zulip):

I think the first bullet is a problem we're working towards. I see the website internationalization as a stepping stone towards it, and more broadly as we work to make the teams better suited to distributed work and so forth

right, well, it might be worth highlighting "outreach and integratation of global community" or something

nikomatsakis (Jan 15 2020 at 15:11, on Zulip):

it's kind of just a separate bullet I guess

simulacrum (Jan 15 2020 at 15:11, on Zulip):

Hm so I would actually say that we should not focus on it as a top-level itrm

nikomatsakis (Jan 15 2020 at 15:12, on Zulip):

I guess that one other fact is I do want to be somewhat "unambitious"

simulacrum (Jan 15 2020 at 15:12, on Zulip):

But rather try and make sure that the organizational work allows for that to happen

simulacrum (Jan 15 2020 at 15:12, on Zulip):

I don't want to bite off more than we can handle

nikomatsakis (Jan 15 2020 at 15:12, on Zulip):

and I do think that the "fairly high level" structure is a good way to sidestep the challenge of not knowing quite how many resources we have to provision

nikomatsakis (Jan 15 2020 at 15:12, on Zulip):

Yes, I feel like there should be some middle ground somehow

nikomatsakis (Jan 15 2020 at 15:13, on Zulip):

maybe it's worth going back a step

nikomatsakis (Jan 15 2020 at 15:13, on Zulip):

why did I list, e.g., error handling and async ?

nikomatsakis (Jan 15 2020 at 15:13, on Zulip):

Error handling is interesting because I see it as "followthrough" but of a differnet kind

simulacrum (Jan 15 2020 at 15:13, on Zulip):

I see the potential team structure changes as a good way to prepare for inclusion of more countries and so forth

nikomatsakis (Jan 15 2020 at 15:13, on Zulip):

in short, I think there are a number of problems we've known about for some time but put off trying to solve

nikomatsakis (Jan 15 2020 at 15:13, on Zulip):

cargo's features are another example

nikomatsakis (Jan 15 2020 at 15:14, on Zulip):

I think there is some value in circling back to some of those long-standing problems and trying to take a fresh ook

nikomatsakis (Jan 15 2020 at 15:14, on Zulip):

specialization probably fits in this bucket

simulacrum (Jan 15 2020 at 15:14, on Zulip):

Yes. I agree. I am not myself 100% sure that we should do so this year

simulacrum (Jan 15 2020 at 15:14, on Zulip):

Like - I want to

nikomatsakis (Jan 15 2020 at 15:14, on Zulip):

Well, I think it really depends on what we've got

simulacrum (Jan 15 2020 at 15:15, on Zulip):

Maybe we can identify these as long standing issues that are good test cases for RFC redesign

nikomatsakis (Jan 15 2020 at 15:15, on Zulip):

async is a bit different, in any case

nikomatsakis (Jan 15 2020 at 15:15, on Zulip):

it's more like "don't let this stall out"

simulacrum (Jan 15 2020 at 15:15, on Zulip):

Sure

nikomatsakis (Jan 15 2020 at 15:16, on Zulip):

Yes. I agree. I am not myself 100% sure that we should do so this year

I mean, we sort of are, right? at least the cargo roadmap that @Eric Huss laid out included attack on some long-standing problems

nikomatsakis (Jan 15 2020 at 15:16, on Zulip):

it's a bit weird though to say "one of our problems is long-standing problems"

simulacrum (Jan 15 2020 at 15:16, on Zulip):

Well, yes. I'm not saying that things aren't shaping that way

nikomatsakis (Jan 15 2020 at 15:16, on Zulip):

without trying to name them :)

nikomatsakis (Jan 15 2020 at 15:16, on Zulip):

(but when phrased as goals it makes sense)

simulacrum (Jan 15 2020 at 15:16, on Zulip):

I'm saying that if we go for those problems I don't think we'll have much time left

simulacrum (Jan 15 2020 at 15:17, on Zulip):

Which is fine! But we should be cautious about expecting to work on organizational challenges at the same time

nikomatsakis (Jan 15 2020 at 15:18, on Zulip):

I'm pondering :)

nikomatsakis (Jan 15 2020 at 15:18, on Zulip):

I definitely agree about trying to avoid being too ambitious

nikomatsakis (Jan 15 2020 at 15:19, on Zulip):

let's put this question a bit on hold for a second and look at 'improve projection functioning, governance, and visibility". I tried to make that a bit more specific by adding this text:

nikomatsakis (Jan 15 2020 at 15:19, on Zulip):

(curse Zulip's markdown for not handling that correctly, grr)

nikomatsakis (Jan 15 2020 at 15:19, on Zulip):

I'm wondering if those bullets themselves are kind of "goals" to highlight

nikomatsakis (Jan 15 2020 at 15:19, on Zulip):

i.e., replace the single bullet with those bullets

simulacrum (Jan 15 2020 at 15:19, on Zulip):

I think they are

simulacrum (Jan 15 2020 at 15:20, on Zulip):

But I like our 3 bullets :)

nikomatsakis (Jan 15 2020 at 15:20, on Zulip):

(or perhaps create sub-bullets)

simulacrum (Jan 15 2020 at 15:20, on Zulip):

Sub bullets seems good

simulacrum (Jan 15 2020 at 15:20, on Zulip):

I think these are the major prongs/problem statements

simulacrum (Jan 15 2020 at 15:20, on Zulip):

(in this area)

nikomatsakis (Jan 15 2020 at 15:21, on Zulip):

Side note, I've been reading this book Messy and it seems pretty relevant to all this

nikomatsakis (Jan 15 2020 at 15:21, on Zulip):

recommended :)

nikomatsakis (Jan 15 2020 at 15:22, on Zulip):

I'm finding the middle bullet hard to convert into "goal form"

nikomatsakis (Jan 15 2020 at 15:23, on Zulip):

the problem statement is "too few people"

nikomatsakis (Jan 15 2020 at 15:23, on Zulip):

although I suspect it is "trying to do too much"

nikomatsakis (Jan 15 2020 at 15:23, on Zulip):

it seems like the real problem is that we have trouble scaling our initiatives to what is available and planning around that?

simulacrum (Jan 15 2020 at 15:23, on Zulip):

I think also root cause of first bullet

nikomatsakis (Jan 15 2020 at 15:24, on Zulip):

i.e., trying to do too much is a root cause of first bullet?

nikomatsakis (Jan 15 2020 at 15:24, on Zulip):

if so, I agree

simulacrum (Jan 15 2020 at 15:24, on Zulip):

i.e. we don't have people because people don't know what to do

nikomatsakis (Jan 15 2020 at 15:24, on Zulip):

ah

nikomatsakis (Jan 15 2020 at 15:24, on Zulip):

I also agree

nikomatsakis (Jan 15 2020 at 15:24, on Zulip):

but I think part of that is that coordinating people is hard

nikomatsakis (Jan 15 2020 at 15:25, on Zulip):

that's kind of a "mentoring" / "path to membership" problem

nikomatsakis (Jan 15 2020 at 15:25, on Zulip):

it's complex :)

nikomatsakis (Jan 15 2020 at 15:25, on Zulip):

and project groups will solve ALL OF IT :P

nikomatsakis (Jan 15 2020 at 15:25, on Zulip):
simulacrum (Jan 15 2020 at 15:26, on Zulip):

Heh, I hope so :)

simulacrum (Jan 15 2020 at 15:28, on Zulip):

I do think doing too much exacerbates the problem

simulacrum (Jan 15 2020 at 15:28, on Zulip):

i.e., hard to put it all on one page or so

simulacrum (Jan 15 2020 at 15:28, on Zulip):

but I also think that our high parallelism has advantages

simulacrum (Jan 15 2020 at 15:28, on Zulip):

(OTOH, even I don't usually know where I can help out, if I wanted to -- and that means that probably no one knows :)

nikomatsakis (Jan 15 2020 at 15:29, on Zulip):

yes to all the above :)

nikomatsakis (Jan 15 2020 at 15:30, on Zulip):

I can't quite tell if a goal of "get more people paid to work on rust"...

nikomatsakis (Jan 15 2020 at 15:30, on Zulip):

it's definitely out of our control :) but it's relevant

nikomatsakis (Jan 15 2020 at 15:30, on Zulip):

well, not fully

simulacrum (Jan 15 2020 at 15:31, on Zulip):

I think it should not be

nikomatsakis (Jan 15 2020 at 15:31, on Zulip):
nikomatsakis (Jan 15 2020 at 15:31, on Zulip):

woah, sub-bullets worked

nikomatsakis (Jan 15 2020 at 15:31, on Zulip):

in any case, I can't tell what XXX should be

simulacrum (Jan 15 2020 at 15:31, on Zulip):

because in some sense right now I don't think if people were paid, it is clear that they would be able to "find work"

nikomatsakis (Jan 15 2020 at 15:32, on Zulip):

Well, I'm thinking a bit about what e.g. the Parity post said

simulacrum (Jan 15 2020 at 15:32, on Zulip):

Maybe "ensure follow-through on design work"

nikomatsakis (Jan 15 2020 at 15:32, on Zulip):

which I still feel captures a problem (and one more general than companies)

nikomatsakis (Jan 15 2020 at 15:33, on Zulip):

“We, too, have team members who are interested in helping on specialization or fixing the aforementioned bugs. However, it’s often unclear whether the work is worthwhile. To a business, it is hard to argue that one might spend a month or two working on a new feature without any assurance that the approach taken would be accepted.”

simulacrum (Jan 15 2020 at 15:33, on Zulip):

Yes, I think that fits into the "find work" problem, in some sense

nikomatsakis (Jan 15 2020 at 15:34, on Zulip):

I think people misread this as "we want guarantees", but I see it more like "we want agreement that we'll follow through and solve the problem, but we're willing to do the work"

simulacrum (Jan 15 2020 at 15:34, on Zulip):

I agree that there's two sides to it -- identifying problems and matching those to 'plausible acceptance'

nikomatsakis (Jan 15 2020 at 15:34, on Zulip):

(and I think the same dynamic definitely arises with volunteers)

simulacrum (Jan 15 2020 at 15:34, on Zulip):

Yes, I think I agree on the latter

simulacrum (Jan 15 2020 at 15:34, on Zulip):

i.e., that it is not guarantees

simulacrum (Jan 15 2020 at 15:34, on Zulip):

I think the same happens with volunteers

nikomatsakis (Jan 15 2020 at 15:35, on Zulip):
nikomatsakis (Jan 15 2020 at 15:35, on Zulip):

maybe that?

simulacrum (Jan 15 2020 at 15:36, on Zulip):

hm -- I think "scale" is not quite the right verb :)

simulacrum (Jan 15 2020 at 15:36, on Zulip):

but it kind of matches, hm

nikomatsakis (Jan 15 2020 at 15:36, on Zulip):

maybe just no second bullelt :)

simulacrum (Jan 15 2020 at 15:37, on Zulip):

I think it is important to call this out

nikomatsakis (Jan 15 2020 at 15:37, on Zulip):

so the problem is a "too few hands"

nikomatsakis (Jan 15 2020 at 15:37, on Zulip):

the solutions really are two fold, I guess

simulacrum (Jan 15 2020 at 15:37, on Zulip):

i.e., it is a key problem

nikomatsakis (Jan 15 2020 at 15:37, on Zulip):

or three

nikomatsakis (Jan 15 2020 at 15:37, on Zulip):
nikomatsakis (Jan 15 2020 at 15:37, on Zulip):
nikomatsakis (Jan 15 2020 at 15:37, on Zulip):
nikomatsakis (Jan 15 2020 at 15:37, on Zulip):

maybe just call out those second two as things we should try to do

simulacrum (Jan 15 2020 at 15:37, on Zulip):

I'm not sure that "too few hands" -- maybe "too few matched hands, many unmatched hands"

nikomatsakis (Jan 15 2020 at 15:38, on Zulip):

well

nikomatsakis (Jan 15 2020 at 15:38, on Zulip):

I don't thikn it's that simple

nikomatsakis (Jan 15 2020 at 15:38, on Zulip):

I think a lot of times the problem is "reviewing and mentoring" bandwidth more than anything

nikomatsakis (Jan 15 2020 at 15:38, on Zulip):

it's not so much that there's a lack of volunteers around

nikomatsakis (Jan 15 2020 at 15:38, on Zulip):

it's that trying to help structure efforts is exhausting, and there are relatively few people with the time to do it

simulacrum (Jan 15 2020 at 15:38, on Zulip):

hm, so maybe you misunderstood :)

By matched I mean to -- language team time, review/mentoring bandwidth, etc

nikomatsakis (Jan 15 2020 at 15:39, on Zulip):

so maybne that's actually what we should say

nikomatsakis (Jan 15 2020 at 15:39, on Zulip):
simulacrum (Jan 15 2020 at 15:39, on Zulip):

Yes, I think that's good

nikomatsakis (Jan 15 2020 at 15:39, on Zulip):
nikomatsakis (Jan 15 2020 at 15:39, on Zulip):

yes, feeling good about that

simulacrum (Jan 15 2020 at 15:40, on Zulip):

(and we can expand on why)

nikomatsakis (Jan 15 2020 at 15:40, on Zulip):

that clearly also covers "help people get paid"

nikomatsakis (Jan 15 2020 at 15:40, on Zulip):

right

simulacrum (Jan 15 2020 at 15:43, on Zulip):

Maybe we should try to circle back to why we want to tackle error handling etc

i.e., it's still not completely clear to me other than just "long standing problem"

simulacrum (Jan 15 2020 at 15:43, on Zulip):
simulacrum (Jan 15 2020 at 15:43, on Zulip):

I guess it fits under the 2nd bullet

simulacrum (Jan 15 2020 at 15:44, on Zulip):

to be honest these three bullets are looking very ambitious

nikomatsakis (Jan 15 2020 at 15:45, on Zulip):

well it all depends

nikomatsakis (Jan 15 2020 at 15:45, on Zulip):

I don't think we'll "vanquish" these problems

nikomatsakis (Jan 15 2020 at 15:45, on Zulip):

but I think we can make strides towards them

simulacrum (Jan 15 2020 at 15:45, on Zulip):

Sure, yes -- I guess -- it feels like we're tackling three huge bears

nikomatsakis (Jan 15 2020 at 15:46, on Zulip):

they're sort of "evergreen" :)

nikomatsakis (Jan 15 2020 at 15:46, on Zulip):

we could be more narrow, e.g., "form a foundation" or whatever

nikomatsakis (Jan 15 2020 at 15:46, on Zulip):

"introduce a staged rfc process"

simulacrum (Jan 15 2020 at 15:46, on Zulip):

I think this is good for the high level goals

nikomatsakis (Jan 15 2020 at 15:46, on Zulip):

but maybe it's better to just say

nikomatsakis (Jan 15 2020 at 15:46, on Zulip):

" we know that we can't 'solve' these problems "

nikomatsakis (Jan 15 2020 at 15:47, on Zulip):

"but we should be trying to take steps in each of these areas and make improvements"

simulacrum (Jan 15 2020 at 15:47, on Zulip):

But we should clarify that we intend to make progress -- i.e., it is not our intent to finish/solve/declare done

simulacrum (Jan 15 2020 at 15:47, on Zulip):

yeah, that language sounds good

nikomatsakis (Jan 15 2020 at 15:48, on Zulip):

another thing

nikomatsakis (Jan 15 2020 at 15:49, on Zulip):

something we haven't really done much of

nikomatsakis (Jan 15 2020 at 15:49, on Zulip):

is interim reports that talk about

nikomatsakis (Jan 15 2020 at 15:49, on Zulip):
nikomatsakis (Jan 15 2020 at 15:49, on Zulip):
nikomatsakis (Jan 15 2020 at 15:49, on Zulip):

I think that we should be doing more of this

nikomatsakis (Jan 15 2020 at 15:49, on Zulip):

i.e., I sort of think it would be the norm that the "goals" are the sorts of things that are rarely if ever "done"

nikomatsakis (Jan 15 2020 at 15:49, on Zulip):

but it's more towards the end of the year that we talk about what we have done

nikomatsakis (Jan 15 2020 at 15:50, on Zulip):

maybe that belongs in the calendar

nikomatsakis (Jan 15 2020 at 15:50, on Zulip):

the original north star RFC included plans for this

nikomatsakis (Jan 15 2020 at 15:50, on Zulip):

just hard to find the bandwidth sometimes :)

nikomatsakis (Jan 15 2020 at 15:50, on Zulip):

but I feel it's sort of imporant for the Rust organization

simulacrum (Jan 15 2020 at 15:51, on Zulip):

We should make it part of the calendar

simulacrum (Jan 15 2020 at 15:51, on Zulip):

Maybe November?

simulacrum (Jan 15 2020 at 15:51, on Zulip):

i.e., we decide on edition in October

simulacrum (Jan 15 2020 at 15:51, on Zulip):

and close out the year in Novemeber

simulacrum (Jan 15 2020 at 15:52, on Zulip):

also a good fit for survey planning perhaps

nikomatsakis (Jan 15 2020 at 15:52, on Zulip):

no later than that for sure

nikomatsakis (Jan 15 2020 at 15:52, on Zulip):
nikomatsakis (Jan 15 2020 at 15:52, on Zulip):

that is what I wrote so far

simulacrum (Jan 15 2020 at 15:52, on Zulip):

I think putting it earlier is hard

nikomatsakis (Jan 15 2020 at 15:52, on Zulip):

I think November is probably good

simulacrum (Jan 15 2020 at 15:52, on Zulip):

hm so September is not "Start" for the survey then

nikomatsakis (Jan 15 2020 at 15:52, on Zulip):

Why not?

simulacrum (Jan 15 2020 at 15:52, on Zulip):

we need to start in august I think if we want results by october

simulacrum (Jan 15 2020 at 15:53, on Zulip):

it'll go live september

nikomatsakis (Jan 15 2020 at 15:53, on Zulip):

sorry, that's what I meant

nikomatsakis (Jan 15 2020 at 15:53, on Zulip):

good to call that out

nikomatsakis (Jan 15 2020 at 15:53, on Zulip):
nikomatsakis (Jan 15 2020 at 15:53, on Zulip):

this btw is super helpful

nikomatsakis (Jan 15 2020 at 15:54, on Zulip):

we should be doing more of this ;)

nikomatsakis (Jan 15 2020 at 15:54, on Zulip):

(maybe we should, e.g. on core-team repo or something, keep a "planning calendar" that project leads and teams can add to...)

simulacrum (Jan 15 2020 at 15:54, on Zulip):

I think we should also publish intermediate retrospectives April and August

simulacrum (Jan 15 2020 at 15:54, on Zulip):

(at least)

simulacrum (Jan 15 2020 at 15:55, on Zulip):

or push for teams to produce documents of that kind

nikomatsakis (Jan 15 2020 at 16:02, on Zulip):

yes

nikomatsakis (Jan 15 2020 at 16:02, on Zulip):

I think most realistically

nikomatsakis (Jan 15 2020 at 16:02, on Zulip):

we will need someone to go do "interviews" with the teams to get this info

nikomatsakis (Jan 15 2020 at 16:02, on Zulip):

and that's ok

nikomatsakis (Jan 15 2020 at 16:02, on Zulip):

actually the "async interviews" model is maybe a good one

nikomatsakis (Jan 15 2020 at 16:02, on Zulip):

I've found it pretty scalable and enjoyable

simulacrum (Jan 15 2020 at 16:03, on Zulip):

Yes -- I think incorporating ecosystem/community is a good fit for the checkins

simulacrum (Jan 15 2020 at 16:03, on Zulip):

(ultimately, that's somewhat who the roadmap is for)

nikomatsakis (Jan 15 2020 at 16:06, on Zulip):

two intermediate retrospectives feels like maybe too much

nikomatsakis (Jan 15 2020 at 16:06, on Zulip):

I guess it depends how much work it is

nikomatsakis (Jan 15 2020 at 16:06, on Zulip):

I'm just worried about overcomitting :)

nikomatsakis (Jan 15 2020 at 16:06, on Zulip):

Ps, i'm working on the text still

nikomatsakis (Jan 15 2020 at 16:06, on Zulip):

coming along nicely I thnk

simulacrum (Jan 15 2020 at 16:08, on Zulip):

hm, maybe

nikomatsakis (Jan 15 2020 at 16:08, on Zulip):

are there any posts from this year that talked about design process.... I think @Yoshua Wuyts probably, but I already quoted them...

nikomatsakis (Jan 15 2020 at 16:08, on Zulip):

two intermediate retrospectives feels like maybe too much

maybe one in the summer?

simulacrum (Jan 15 2020 at 16:08, on Zulip):

well -- so I was thinking that the first is basically the all hands retro

simulacrum (Jan 15 2020 at 16:08, on Zulip):

and the second is "the summer" :)

nikomatsakis (Jan 15 2020 at 16:09, on Zulip):

ok

nikomatsakis (Jan 15 2020 at 16:09, on Zulip):

I was debating about exactly that, an "all hands retro"

simulacrum (Jan 15 2020 at 16:09, on Zulip):

I'm not sure if we should talk much about it, but I think putting it on a calendar is the right thing

simulacrum (Jan 15 2020 at 16:09, on Zulip):

now, mind you, this might just be a post that basically says "here's a bunch of links to paper docs"

nikomatsakis (Jan 15 2020 at 16:10, on Zulip):

we should definitely be doing that

nikomatsakis (Jan 15 2020 at 16:10, on Zulip):

I thnk we didn't last year because it..was particularly draining

simulacrum (Jan 15 2020 at 16:10, on Zulip):

some about design: https://spacekookie.de/blog/rust-2020-the-rfc-process-and-distributions/

simulacrum (Jan 15 2020 at 16:11, on Zulip):

yeah last year was pretty draining

nikomatsakis (Jan 15 2020 at 16:12, on Zulip):

that is actually good :)

nikomatsakis (Jan 15 2020 at 16:12, on Zulip):

I like spacekookie's rather flippant first paragraph :)

nikomatsakis (Jan 15 2020 at 16:12, on Zulip):

PS @Aidan Hobson Sayers just opened a PR, so I'm pinging them to mention we've been chatting a lot here :)

nikomatsakis (Jan 15 2020 at 16:14, on Zulip):

ah but looks mostly easy to integrate

spacekookie (Jan 15 2020 at 16:14, on Zulip):

:D

nikomatsakis (Jan 15 2020 at 16:19, on Zulip):

OK, I pushed the changes we've been discussing @simulacrum

nikomatsakis (Jan 15 2020 at 16:19, on Zulip):

I'm going to go merge @Aidan Hobson Sayers's PR

simulacrum (Jan 15 2020 at 16:23, on Zulip):

I think this is looking really good

simulacrum (Jan 15 2020 at 16:27, on Zulip):

I need to run, but other than some typos and so forth -- I feel like I would personally be ready to signoff on this

simulacrum (Jan 15 2020 at 16:27, on Zulip):

I can try and polish it up a bit and we should try to bring it to core team at meeting today, as well as circle it around among project leads

nikomatsakis (Jan 15 2020 at 16:30, on Zulip):

I agree it's starting to feel like something that could be published

nikomatsakis (Jan 15 2020 at 16:30, on Zulip):

at minimum it's not embarassing

nikomatsakis (Jan 15 2020 at 16:30, on Zulip):

good, this has been weighing on me

nikomatsakis (Jan 21 2020 at 20:46, on Zulip):

I've made a number of edits to the draft and hope to open it soon, though the major sections remain unchanged

simulacrum (Jan 21 2020 at 21:03, on Zulip):

@nikomatsakis :thumbs_up: from me on shipping

Last update: Jan 28 2020 at 01:50UTC