Core Dev Application - Questions/Voting Thread

Tue Nov 7 09:58:37 UTC 2017

Hi Balint,

On Mon, Sep 18, 2017 at 05:36:46PM +0200, Balint Reczey wrote:
> On Mon, Sep 18, 2017 at 10:28 AM, Robie Basak <robie.basak at ubuntu.com> wrote:
> > I realise that others on your team may know you better and thus have
> > more out-of-band knowledge in considering your application, which is
> > fine. For me though, I think that your application as written falls
> > short of general expectations.
> 
> Before applying I made sure that I met every requirement listed for
> Core Dev [1],
> but I'm sorry if I'm failing to meet general expectations. If you
> could provide clear guidance about the general expectation and
> quantify to amount of work I should devote to showing my knowledge I
> would really appreciate that.

Sorry I haven't yet replied to this until now. I have felt that I owed
you a complete and detailed answer as I explained when I met you in
person in late September. Unfortunately I haven't been in one place for
very long since your email and so have found it difficult to catch up
with everything I'm blocking something on while being on the road, and
writing this up as fully as I think you are owed took quite some time.

I've written this from the point of view of the situation at the time of
your application in September to try and explain my reasoning as it was
then. Given the length of your involvement in Ubuntu, I appreciate that
the past two additional months is a significant additional amount of
time proportionally, and the situation may well have changed now. I
invite you to update us and your application with anything new if you
can and haven't done so already.

Applying for core dev is, I think, somewhat special. We don't have any
process for removing someone from core dev. A DMB decision to grant core
dev is effectively permanent and it is important that we don't make a
mistake in doing so. I'm sorry that your experience of this application
hasn't been great, but please understand that your request is to jump in
at the deep end, that this is fairly exceptional and that the DMB isn't
very used to handling this. I expect to hold any core dev applicant to a
higher standard than an applicant to any of the other uploading teams;
for an applicant requesting to go to core dev directly, even more so.

In your case, I haven't seen any technical shortcomings about your work
that would give me reservations about granting you core dev. However, I
feel that I have a very small sample on which to base a core dev
application with regards to Ubuntu-specific work, as opposed to one of
the other uploading teams where I am less demanding.

In my view, a cultural fit is also important. I want to be confident
that decisions you make as a core dev won't upset other developers.
Technical disagreements themselves are to be expected, of course; it is
how Ubuntu developers proceed in the face of a disagreement that I think
is important. And for Canonical employees, understanding and being able
to negotiate the distinction between Ubuntu governance and Canonical
priorities is particularly important. And these are also factors on
which I have little evidence to draw any conclusion.

I appreciate that doing work in Debian where possible is the right thing
to do, that you generally do this well, and that doing the right thing
in this way can seemingly make it difficult to gain the necessary
experience in Ubuntu-specific processes in order to meet the
requirements for an Ubuntu uploader application. However the DMB
generally works on a "need to unblock" principle. If fewer uploads to
Ubuntu are needed because uploading to Debian is preferred, then by
definition those uploads weren't blocking on not being an Ubuntu
uploader. The DMB will focus on what _is_ blocking, rather than what
isn't, and considers applications on that basis.

From a technical perspective, I'd expect a core dev to have:

A thorough understanding and experience in Debian packaging. I have yet
to look in detail, but it seems that this won't be a problem for you,
especially with Steve's endorsement :-)

A demonstrated understanding of Ubuntu-specific processes such as Ubuntu
package merges, SRUs, the release cycle, milestones and exceptions,
proposed migration, the main/universe distinction, MIRs and component
mismatches, handling transitions, and the operation of the seeds. I
don't necessarily expect detailed direct experience in all of these, but
I do expect to see direct and deep experience of at least some of them
and a general understanding of most of them.

In addition I would expect a core dev applicant to either: (a) be an
already experienced Ubuntu uploader through one of the other uploading
teams such as a packageset or MOTU; or (b) present an exceptionally
strong application to go direct to core dev. In the case of (a), the
previous upload history will allow me to assess the application in
detail. Failing that, a string of sponsorships and endorsements from a
wide range of existing uploaders that includes some respected names
would give me confidence in (b). Without either (a) or (b), I don't have
enough to give me the confidence in approving an application that I
think I need.

In your case, as you are in the Foundations team at Canonical, I don't
think (a) really makes sense for you, as there is no other uploading
team suitable for the work you do. So for you it must be (b). But there
are some ways in which I think your application is quite weak at the
moment; both in your application itself and also in your presence in the
Ubuntu development community:

You have only one and a half endorsements (Steve only speaks to your
technical skills and specifically excludes Ubuntu-specific processes and
sponsorships into Ubuntu) and none from anyone who isn't on your team at
Canonical. People tend not to give negative endorsements; instead they
generally decline to give an endorsement. So a lack of endorsements can
mean one of two things: 1) your peers don't think you should get core
dev because they don't think you're suitable; or 2) your peers aren't
negative but either haven't got round to endorsing you or haven't had
enough experience with your work to endorse you. Note that the DMB
cannot tell the difference between the two cases just by looking at a
tiny number of endorsements. Help us by getting a wider range of
endorsements! For a core dev application, especially for jumping
straight to core dev, I think you should aim for five (though I do
weight them based on who is endorsing, how diverse they are and how
positive they are).

Usually, by the time someone applies for core dev, DMB members who are
still active in Ubuntu development have already interacted with them
extensively. I barely see you interacting with other developers in
#ubuntu-devel or #ubuntu-release at all. I don't see a single email to
ubuntu-devel@ or ubuntu-release@ from you ever. This makes a decision
challenging for me: how do I know that your future interactions with
other developers, particularly in the face of some technical
disagreement, will be positive?

Six months with active and heavy involvement in general Ubuntu
development might be enough to demonstrate direct experience of the
release cycle; six months involvement in Ubuntu development in total
with seemingly little involvement does not. So it isn't that six months
isn't enough; it is that in your specific case, you don't seem to have
been involved enough in general Ubuntu development[1] that I have seen to
give me confidence that you have this knowledge or experience.

So right now, though I appreciate your technical skills, my vote is -1
for your application as I'd like to see more involvement and evidence of
Ubuntu-specific processes. I appreciate your contributions and would
like to see you get core dev eventually, but I don't think there's
enough evidence in this application as it stands right now.

Here are some objective actions I think you could take to change my
mind:

1) Activity in Ubuntu-specific development tasks, such as some items
from my list above, and an update to your application wiki page that
make these items explicitly evident; 2) Public collaboration with
existing Ubuntu developers; 3) Endorsements from existing uploaders for
the previous items.

Given the strength of Steve's endorsement on your techncial ability, I
think that a cycle's worth of significant contribution covering
Ubuntu-specific development tasks would be far more than enough. But I
don't see anything even close to that on your application right now. Am
I missing something?

I hope this helps, and I hope that I'll be able to +1 you for core dev
soon :)

Robie
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <https://lists.ubuntu.com/archives/devel-permissions/attachments/20171107/798a1bf9/attachment.sig>