Rust Peers list?

[Nicholas Nethercote]

Hi,

An idea came up in Austin of having a "Rust Peers" list.

Because Rust expertise is unevenly spread, sometimes a patch author or
patch reviewer might want a Rust expert to review a patch from the point of
view of whether it is reasonable Rust code. (E.g. I did this in bug
1423840, where I got glandium to review the patch from a prefs module point
of view, and Manish to review the patch from a Rust point of view.)

Because this might be a common need, it seems worth formalizing. I have
been asking around and the following people have agreed to perform this
role.

- Alexis Beingessner
- Josh Bowman-Matthews
- Emilio Cobos Alvarez
- Manish Goregaokar
- Nika Layzell
- Cameron McCormack

All of these people have extensive Rust experience. (The list could also be
extended, but this is enough people to start.)

The question I have is: should this be a Mozilla module? I can see
arguments in favour and against it being a module.

- In favour: these people do reviews, and the modules list is the canonical
place for finding reviewers.

- Against: these would be opt-in reviews. If a patch author is confident in
their Rust ability, and the "normal" reviewer is likewise confident, then
an extra review from a Rust Peer would not be necessary.

We already have a "C++/Rust usage, tools, and style" module (
https://wiki.mozilla.org/Modules/All#C.2B.2B.2FRust_usage.2C_tools.2C_and_style)
but that feels different to me.

If we choose to make this a module, I'm not sure which section of
https://wiki.mozilla.org/Modules/All it would fall under.

If we choose not to make this a module, I guess this list of people would
instead be put onto https://wiki.mozilla.org/Oxidation.

What do people think?

Thanks.

Nick

[Kan-Ru Chen]

On Wed, Dec 20, 2017, at 6:52 AM, Nicholas Nethercote via governance wrote:
> The question I have is: should this be a Mozilla module? I can see
> arguments in favour and against it being a module.
>
> - In favour: these people do reviews, and the modules list is the canonical
> place for finding reviewers.
>
> - Against: these would be opt-in reviews. If a patch author is confident in
> their Rust ability, and the "normal" reviewer is likewise confident, then
> an extra review from a Rust Peer would not be necessary.
>
> We already have a "C++/Rust usage, tools, and style" module (
> https://wiki.mozilla.org/Modules/All#C.2B.2B.2FRust_usage.2C_tools.2C_and_style)
> but that feels different to me.
>
> If we choose to make this a module, I'm not sure which section of
> https://wiki.mozilla.org/Modules/All it would fall under.
>
> If we choose not to make this a module, I guess this list of people would
> instead be put onto https://wiki.mozilla.org/Oxidation.
>
> What do people think?

If we expect our Rust expertise to grow quickly (I think we should) this list would not be necessary. Therefore I support to put it onto https://wiki.mozilla.org/Oxidation.

Other the other hand, if they are going to collect and formalize best practices into a Rust style guide, they should probably be the "C++/Rust usage, tools, and style" module peers.

My 2c.

Kanru

[mhoye]

On 2017-12-19 5:52 PM, Nicholas Nethercote via governance wrote:
> The question I have is: should this be a Mozilla module? I can see
> arguments in favour and against it being a module.
>
> - In favour: these people do reviews, and the modules list is the canonical
> place for finding reviewers.

The Module system provides clarity of ownership and a clear escalation
path in cases of disagreement, which both seem like sound things to
invest in early as we expect Rust usage and participation to grow.

- mhoye

[Nicholas Nethercote]

On Wed, Dec 20, 2017 at 2:10 PM, Kan-Ru Chen <ka...@kanru.info> wrote:

> >
> > If we choose to make this a module, I'm not sure which section of
> > https://wiki.mozilla.org/Modules/All it would fall under.
> >
> > If we choose not to make this a module, I guess this list of people would
> > instead be put onto https://wiki.mozilla.org/Oxidation.
>

> If we expect our Rust expertise to grow quickly (I think we should) this
> list would not be necessary. Therefore I support to put it onto
> https://wiki.mozilla.org/Oxidation.
>

That's a good point. To restate and expand:

- We don't have "C++ peers" because C++ expertise is an implicit
prerequisite for being a peer in a module containing C++ code.

- We don't have "JS peers" because JS expertise is an implicit prerequisite
for being a peer in a module containing JS code.

- Ditto for other languages (e.g. Python for build system modules).

Rust is currently special because it's so new, but eventually it'll be like
the others:

- We won't have "Rust peers" because Rust expertise will be an implicit
prerequisite for being a peer in a module containing Rust code.

In other words, this list of people will become less necessary over time,
which is not what you'd expect for a Mozilla module. So I will not make
this a module and just put it on the Oxidation wiki page.

Thank you to all those who responded.

Nick