>> Although we've seen bad bugs (and bad memory bugs in particular) come
>> from coders of all levels of experience, I personally am not at all
>> comfortable with the idea of inexperienced Firefox front-end coders
>> having review power in a module we intend to ship in Firefox.
>
> That's not the situation we have. Jed Parsons and I have both landed a good
> bit of code (in BrowserID module) for FX Desktop and FXOS+Gaia. So we have
> experience. We're not the world's experts, but we also know our limitations
> and have always consulted with other teams starting with initial designs
> (e.g. Dolske).
>
> Remember that an equally (if not more) important aspect of this module is
> the service integration piece, which is where we *are* indeed experts.

Indeed, and nobody is arguing that you shouldn't have full review
powers over the service parts of this project.  The argument is that
Firefox peers should at least sign off on this plan to create a new
Firefox module composed of people who, as I understand, have not
landed any code in desktop Firefox which is currently enabled.

Also note that two of the proposed module peers apparently do not have
any experience hacking on Firefox or Gecko at all.  (At least, I don't
see any commits in m-c with the names 'Mayo' or 'Hilaiel'.)

I would object strongly to making someone a peer of a new Gecko module
affecting desktop Firefox if their only experience on the desktop was
hacking on a module which was currently disabled, and of course we
wouldn't even consider giving someone review power in Gecko if they
had no experience there.  But I can't speak for the Firefox folks,
which is why I'm asking that you find someone who is willing to vouch
for you.

>> If one or more existing Firefox peers is willing to personally accept
>> responsibility for this module, I'd be OK with this proposal.  Do you
>> think that's a fair burden to meet?
>
> Given the above, and given the module's explicit callout that nay overlap
> will require full cooperation of other module owners, I don't think that's a
> fair burden to meet, if only because everyone is already incredibly
> over-committed.

I don't think that the "full cooperation of the relevant module
owners" criterion is sufficient.  If your intent is to get relevant
module owners to review your code, then just create a separate module
for the Firefox bits, make them owners and peers, and I will no longer
object.

But it sounds like your intent is to seek rubber stamps from module
owners/peers and merely "consult with [them when] starting with
initial designs".  If you're going to do that, all I'm asking is that
you find a Firefox peer who is willing to vouch for your proposal now,
by saying "In my opinion, Ben, Jed, Mark, and Lloyd know enough to
review patches in Firefox, and if they screw anything up, I will
personally ensure that things are righted in a timely fashion."

If the peer is confident that you guys aren't going to screw anything
up, this commitment requires zero work, so I don't buy the "nobody has
time" argument.

If on the other hand nobody is willing to take personal responsibility
for this proposed structure, then I'd think that's a strong sign of no
confidence from our experts, and I would continue to object to this
proposal.  But please keep in mind that I am not a peer of the module
ownership structure, so I'll be objecting from up in the peanut
gallery.  You'll scarcely be able to hear me from your seats.  :)

-Justin
_______________________________________________
governance mailing list
[email protected]
https://lists.mozilla.org/listinfo/governance

Reply via email to