Re: Clarifications needed to 'Intent to ship' process

[Ehsan Akhgari]

On Wed, Apr 27, 2016 at 12:33 PM, smaug  wrote:

> On 04/26/2016 01:47 PM, Ehsan Akhgari wrote:
>
>> On 2016-04-26 1:02 PM, Mike Hommey wrote:
>>
>>> On Mon, Apr 25, 2016 at 10:52:02PM -0400, Boris Zbarsky wrote:
>>>
 On 4/25/16 10:34 PM, Mike Hommey wrote:

> Don't we already have that with superreviewers?
>

 Kinda, sorta.

 (How outdated is that list, btw?)
>

 Quite.  If we're talking about
 https://www.mozilla.org/en-US/about/governance/policies/reviewers/
 then of
 the 30 people on the list, I would say:

 * 10 or 11 are no longer actively involved.
 * 4-5 are not actively involved in web-facing bits.
 * 2-3 are involved in web-facing bits in a pretty narrow capacity.

 If I were coming up with an "API owners" group, I'd take some of the
 people
 from this list, but also probably a few who are not on this list

>>>
>>> Shouldn't we just kind of repurpose the superreviewers, update the list,
>>> and keep it fresh?
>>>
>>
>> The traditional job of superreviewers is very different though (and it's
>> mostly a historical thing anyway).
>>
>>
> Well, a traditional job for superreviewers is/was things like API reviews,
> which isn't that different
> from saying what APIs (or other features) should be exposed to the web.
>
> Updating superreviewer list and repurposing it a bit sounds reasonable to
> me.
>

OK, sure.  I don't have any strong arguments against it!  At least the
superreviewer list will start to become useful again! :-)

-- 
Ehsan
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[smaug]

On 04/26/2016 10:31 AM, Ehsan Akhgari wrote:

On 2016-04-25 10:58 PM, Boris Zbarsky wrote:

That said, note that a bunch of the items above lie somewhat out of a
specific module, so it's not clear to me that we want intent-to-ship OKs
to be module-specific.


Yeah, a bunch of stuff definitely crosses multiple module boundaries, so
I think relying on the module system for shipping intents will make it
unclear who will have the final say.

The model that Blink has been using is a few individuals who are API
owners  deciding whether
they can ship a specific feature.  Perhaps we can use a similar model.



They just seem to have surprisingly few owners. That might be an issue 
occasionally.
And in our case, who should have the power to say, "this is ok to ship". I just 
suggested peers since that
is something we have already, but I agree it may not be right option.


___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[smaug]

On 04/26/2016 01:47 PM, Ehsan Akhgari wrote:

On 2016-04-26 1:02 PM, Mike Hommey wrote:

On Mon, Apr 25, 2016 at 10:52:02PM -0400, Boris Zbarsky wrote:

On 4/25/16 10:34 PM, Mike Hommey wrote:

Don't we already have that with superreviewers?


Kinda, sorta.


(How outdated is that list, btw?)


Quite.  If we're talking about
https://www.mozilla.org/en-US/about/governance/policies/reviewers/ then of
the 30 people on the list, I would say:

* 10 or 11 are no longer actively involved.
* 4-5 are not actively involved in web-facing bits.
* 2-3 are involved in web-facing bits in a pretty narrow capacity.

If I were coming up with an "API owners" group, I'd take some of the people
from this list, but also probably a few who are not on this list


Shouldn't we just kind of repurpose the superreviewers, update the list,
and keep it fresh?


The traditional job of superreviewers is very different though (and it's
mostly a historical thing anyway).



Well, a traditional job for superreviewers is/was things like API reviews, 
which isn't that different
from saying what APIs (or other features) should be exposed to the web.

Updating superreviewer list and repurposing it a bit sounds reasonable to me.

-Olli
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Eric Shepherd]

I'd also love to take this opportunity to remind everyone, especially
our newer contributors and developers, to be sure to add the
"dev-doc-needed" keyword to the appropriate bugs for any changes which
should include updates to documentation on MDN.

See
https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Getting_documentation_updated
for details on how this system works to help us document all the good stuff.


*From:* smaug
*Sent:* Monday, Apr 25, 2016 1:19:16 AM EDT
*To:* dev-platform@lists.mozilla.org
*Subject:* Clarifications needed to 'Intent to ship' process

> based on couple of conversations we need some clarifications to
> 'intent to ship'.
>
> First, we aren't yet consistent enough to send 'intent to ship'
> emails. I think that takes
> just some time for patch authors and reviewers to get used to the
> process, that whenever there is some
> larger than minor web phasing API addition/removal being done, 'intent
> to ship' email to this list should be sent.
>
> Second, it isn't clear how we're supposed to react to the 'intent to
> ship' emails.
> I propose we require two OKs from the owners/peers of the relevant
> module (of which one could be given while reviewing the patch), and
> definitely no opposing comments from the owners/peers. But in case
> other people object... I guess we'll always have special cases and
> process can be
> improved when needed.

-- 

Eric Shepherd
Senior Technical Writer
Mozilla Developer Network 
Blog: https://www.bitstampede.com/
Twitter: http://twitter.com/sheppy
Doodle: http://doodle.com/the.sheppy

___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Boris Zbarsky]

On 4/26/16 1:02 AM, Mike Hommey wrote:

Shouldn't we just kind of repurpose the superreviewers, update the list,
and keep it fresh?


I think that would be pretty reasonable, yes.

-Boris

___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Ehsan Akhgari]

On 2016-04-26 1:02 PM, Mike Hommey wrote:
> On Mon, Apr 25, 2016 at 10:52:02PM -0400, Boris Zbarsky wrote:
>> On 4/25/16 10:34 PM, Mike Hommey wrote:
>>> Don't we already have that with superreviewers?
>>
>> Kinda, sorta.
>>
>>> (How outdated is that list, btw?)
>>
>> Quite.  If we're talking about
>> https://www.mozilla.org/en-US/about/governance/policies/reviewers/ then of
>> the 30 people on the list, I would say:
>>
>> * 10 or 11 are no longer actively involved.
>> * 4-5 are not actively involved in web-facing bits.
>> * 2-3 are involved in web-facing bits in a pretty narrow capacity.
>>
>> If I were coming up with an "API owners" group, I'd take some of the people
>> from this list, but also probably a few who are not on this list
> 
> Shouldn't we just kind of repurpose the superreviewers, update the list,
> and keep it fresh?

The traditional job of superreviewers is very different though (and it's
mostly a historical thing anyway).

___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Mike Hommey]

On Mon, Apr 25, 2016 at 10:52:02PM -0400, Boris Zbarsky wrote:
> On 4/25/16 10:34 PM, Mike Hommey wrote:
> >Don't we already have that with superreviewers?
> 
> Kinda, sorta.
> 
> >(How outdated is that list, btw?)
> 
> Quite.  If we're talking about
> https://www.mozilla.org/en-US/about/governance/policies/reviewers/ then of
> the 30 people on the list, I would say:
> 
> * 10 or 11 are no longer actively involved.
> * 4-5 are not actively involved in web-facing bits.
> * 2-3 are involved in web-facing bits in a pretty narrow capacity.
> 
> If I were coming up with an "API owners" group, I'd take some of the people
> from this list, but also probably a few who are not on this list

Shouldn't we just kind of repurpose the superreviewers, update the list,
and keep it fresh?

Mike
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Boris Zbarsky]

On 4/25/16 10:34 PM, Mike Hommey wrote:

Don't we already have that with superreviewers?


Kinda, sorta.


(How outdated is that list, btw?)


Quite.  If we're talking about 
https://www.mozilla.org/en-US/about/governance/policies/reviewers/ then 
of the 30 people on the list, I would say:


* 10 or 11 are no longer actively involved.
* 4-5 are not actively involved in web-facing bits.
* 2-3 are involved in web-facing bits in a pretty narrow capacity.

If I were coming up with an "API owners" group, I'd take some of the 
people from this list, but also probably a few who are not on this list


-Boris
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Mike Hommey]

On Tue, Apr 26, 2016 at 10:31:06AM +0800, Ehsan Akhgari wrote:
> On 2016-04-25 10:58 PM, Boris Zbarsky wrote:
> > That said, note that a bunch of the items above lie somewhat out of a
> > specific module, so it's not clear to me that we want intent-to-ship OKs
> > to be module-specific.
> 
> Yeah, a bunch of stuff definitely crosses multiple module boundaries, so
> I think relying on the module system for shipping intents will make it
> unclear who will have the final say.
> 
> The model that Blink has been using is a few individuals who are API
> owners  deciding whether
> they can ship a specific feature.  Perhaps we can use a similar model.

Don't we already have that with superreviewers? (How outdated is that
list, btw?)

Mike
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Ehsan Akhgari]

On 2016-04-25 10:58 PM, Boris Zbarsky wrote:
> That said, note that a bunch of the items above lie somewhat out of a
> specific module, so it's not clear to me that we want intent-to-ship OKs
> to be module-specific.

Yeah, a bunch of stuff definitely crosses multiple module boundaries, so
I think relying on the module system for shipping intents will make it
unclear who will have the final say.

The model that Blink has been using is a few individuals who are API
owners  deciding whether
they can ship a specific feature.  Perhaps we can use a similar model.

___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Boris Zbarsky]

On 4/25/16 1:47 AM, Jet Villegas wrote:

The "Intent to Implement" should help some of the concerns and allow for
comments. "Intent to Ship" usually means (at least for Platform Rendering)
that we'll be removing the #ifndef RELEASE flags and enabling preferences.
That is, by the time the "Intent to Ship" e-mail is sent, fundamental
issues should have already been resolved with module owners and other
parties.


It seems to me that "Intent to Implement" is mostly about "is this 
possibly a good idea at all?" and may well not involve a thorough API 
review (because the API may well change as a result of implementor 
feedback during implementation) nor a thorough consideration of web 
compat (because there is no data yet), what other UAs are doing (because 
it doesn't matter that much at that early stage), or the actual state of 
the feature in relation to our codebase (because it's not known yet).


An "Intent to Ship" should consider all of the above; it's a much higher 
bar.


That said, note that a bunch of the items above lie somewhat out of a 
specific module, so it's not clear to me that we want intent-to-ship OKs 
to be module-specific.


-Boris
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Re: Clarifications needed to 'Intent to ship' process

[Jet Villegas]

The "Intent to Implement" should help some of the concerns and allow for
comments. "Intent to Ship" usually means (at least for Platform Rendering)
that we'll be removing the #ifndef RELEASE flags and enabling preferences.
That is, by the time the "Intent to Ship" e-mail is sent, fundamental
issues should have already been resolved with module owners and other
parties.

--Jet

On Mon, Apr 25, 2016 at 2:19 PM, smaug  wrote:

> Hi all,
>
>
> based on couple of conversations we need some clarifications to 'intent to
> ship'.
>
> First, we aren't yet consistent enough to send 'intent to ship' emails. I
> think that takes
> just some time for patch authors and reviewers to get used to the process,
> that whenever there is some
> larger than minor web phasing API addition/removal being done, 'intent to
> ship' email to this list should be sent.
>
> Second, it isn't clear how we're supposed to react to the 'intent to ship'
> emails.
> I propose we require two OKs from the owners/peers of the relevant module
> (of which one could be given while reviewing the patch), and
> definitely no opposing comments from the owners/peers. But in case other
> people object... I guess we'll always have special cases and process can be
> improved when needed.
>
> Then there is also the case when relevant module doesn't really have
> peers, or not enough peers. I guess in that case OK from an owner/peer of a
> related module is fine too, or from a DE, or even from an active
> superreviewer (I know we aren't really using sr that much anymore).
>
> Note, this all is a separate thing from .webidl reviews, since the webidl
> may be written way before shipping, and reviewers for the .webidl are
> possibly different than those giving OKs to the overall API and its
> implementation.
>
>
> https://wiki.mozilla.org/WebAPI/ExposureGuidelines should be update once
> there is some agreement on what kind of process we want.
>
>
> comments?
>
>
> -Olli
> ___
> dev-platform mailing list
> dev-platform@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-platform
>
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform