Gentoo Archives: gentoo-osx

From: Finn Thain <fthain@××××××××××××××××.au>
To: gentoo-osx@l.g.o
Subject: Re: [gentoo-osx] Elections for Project Lead Positions
Date: Thu, 11 Aug 2005 05:26:27
Message-Id: Pine.LNX.4.63.0508111442101.16790@loopy.telegraphics.com.au
In Reply to: [gentoo-osx] Elections for Project Lead Positions by Lina Pezzella
1 On Thu, 11 Aug 2005, Lina Pezzella wrote:
2
3 > -----BEGIN PGP SIGNED MESSAGE-----
4 > Hash: SHA1
5 >
6 > We're happy to see feedback on the list in regards to our recent message about
7 > electing project leads. Hopefully there's been enough time for everyone to get
8 > their few cents in. At this point, we'd like to clarify what would change were
9 > we elected as leads, as well as iron out the details of the election process.
10 >
11 > Operations:
12 >
13 > 1) Development of General Team Policy - With only four active
14 > developers, it has been pretty easy to develop consensus on how we as
15 > a project treat different situations. Our recent agreement to switch
16 > to 'use userland_Darwin' instead of 'use ppc-macos' in ebuilds is an
17 > example of this. However, the project will not always consist of four
18 > people (it now effectively consists of six active members). Since we
19 > are a relatively new project, we can avoid some of the issues that
20 > plague the -core and -dev mailing lists by keeping a general policy
21 > page up to date from early on (I.E. now). This will also be of great
22 > help to new developers and potential contributors. It might not be
23 > critical now, but as we expand as a project into the future, it will
24 > become more and more important.
25
26 Good idea. Question is, will all developers commit to keeping it current.
27
28 > 2) Intraproject Relations - Some of us are self-motivated and like to do
29 > our own thing while others want direction. I am hoping that part of
30 > our problem with inactive developers comes from this lack of
31 > direction. Furthermore, if we all do our own thing without any
32 > central coordination, we will inevitably end up duplicating work
33 > (this has already happened more than just once). For both of these
34 > reasons, we think it might be a good idea to keep a listing of what
35 > each developer is working on with regards to the project (perhaps on
36 > the project page, which we all have CVS commit access to). For
37 > instance, Lina maintains a package in sci-biology and she maintains
38 > the sci-biology category for ppc-macos. We're not looking for a list
39 > of packages people are involved in porting each week; we're looking
40 > to keep track of long-term goals that each developer is working
41 > towards. This is also good for PR purposes, as well as in the
42 > recruitment/contributions department. (We don't feel that Bugzilla
43 > isn't suited for keeping track of each developer's current work in
44 > this way.)
45
46 CVS may not be a good way to improve communication, it might need to be
47 more convenient before people will use it.
48
49 > Strategic:
50 >
51 > 1) Recruitment - Let's face it, we need more devs, badly. To make the
52 > recruitment process more manageable and coherent for all parties
53 > involved, some sort of central recruitment liason(s) need to be
54 > available and announced as such to the outside world. While virtually
55 > any developer is able to recruit, it's not as easy as just announcing
56 > open positions or a need for help, and then watching the e-mails come
57 > in. We tried that, and it didn't work so well. Often the best
58 > applicants are those who don't apply themselves. Fabian, our most
59 > awesomest recent recruit, did not send us an e-mail; rather, we
60 > e-mailed him after seeing some of his activity in Bugzilla. It's a
61 > pretty safe bet to say that good recruitment involves a decent amount
62 > of administrative work and research. Recruitment isn't just about
63 > getting new blood into the team either, though; there's a long
64 > training process involved (both before and after application for
65 > developer membership), and for this to be effective, a central
66 > training body needs to be established for the outside world to be
67 > able to correspond with. We don't want to repeat the initial
68 > recruitment disaster -- it's not fair to the rest of the Gentoo
69 > community, and it's not fair to the new developer not to receive
70 > adequate attention and training. As such, we plan to keep the
71 > recruitment process slow and sure.
72 >
73 > 2) Public Relations - This is something we're not even near having
74 > enough of. The documentation is outdated, and profile changes often
75 > go unannounced until after the fact, if they're announced at all.
76 > We're really happy to see some traffic on this mailing list again,
77 > and this needs to continue to improve. In case not everyone noticed
78 > yet, we, along with a few other devs, have been trying to promote ML
79 > traffic as of late, as opposed to getting things done via IRC and
80 > keeping them unannounced otherwise. We feel that this is a great PR
81 > move, and more action like this in the future needs to be built upon.
82 > Specifically, changes to the profile and anything that would qualify
83 > subjectively as large progress should be announced to the mailing
84 > list -- it's important to keep users informed of changes, and
85 > progress is great for keeping interest in the project. For example,
86 > if a use flag is removed from use.mask, it should be publically
87 > announced, as should news on major packages, such as
88 > baselayout-darwin or perl. Having a lead responsible for this keeps
89 > things easier on everybody. Of course, any other developer that wants
90 > to announce their own progress notes is more than welcome to do so.
91 > ;-)
92
93 Absolutely. Unlike IRC, mailing lists represent searchable archives of
94 useful information and solutions to common problems.
95
96 However, this is not PR, just good communication. Good PR just isn't
97 sufficient in the open development model. This is something even apple is
98 still trying figure out.
99
100 > 3) Interproject Relations - In addition to PR, we need a designated
101 > liason to the rest of the Gentoo community. This becomes particularly
102 > important when other Gentoo projects need to interface with the Mac
103 > OS X project, and vice versa. For example, if a developer from
104 > another project has a Mac and wants to keyword their own packages,
105 > they need to be aware of project policies (such as the use of
106 > userland_Darwin). Having a designated person to go to for this sort
107 > of thing is easier than the current free-for-all situation. After
108 > all, this developer is helping us out; the least we can do is make
109 > things easy for him/her. On the flip side, there needs to be a
110 > designated person to deal with interproject conflicts. While this
111 > isn't exactly a fun job, it's better to have one unified face than to
112 > have split personality disorder. We are a project, not just a group
113 > of developers that occasionally talk to each other.
114
115 Having a dedicated role for this is misguided. IMHO, every gentoo-osx
116 developer should be continually involved with other parts of the wider
117 Gentoo developer community, simply in the course of their job. Also, every
118 gentoo-osx developer should be responsive to other members of the wider
119 Gentoo developer community, just to share the load.
120
121 To create such a dedicated role is to encourage the other devs to believe
122 they can work in isolation. It also bottlenecks access to the project
123 team, which is only really useful if you are trying to contrive some sort
124 of public image.
125
126 If this proposal is a reaction to one dev shooting his/her mouth off and
127 those views being mistaken as representative of the entire project, then
128 there are better solutions.
129
130 >
131 > 4) BugDay Participation (Operations && Strategic) - BugDay is a great
132 > way to meet new potential developers and to allow current developers
133 > to take time off from their current blocker-problem and tackle some
134 > easier stuff. We haven't been utilizing the potential here as much as
135 > we should be. We have a lot of bugs in bugzilla and we all know that
136 > it can be a bit overwhelming sometimes. We might make better
137 > progress if all of us agreed on a short list of things we wanted to
138 > tackle in a day and took some time once a month to work together on
139 > them. Pine and Pango come to mind as two recurring problems we might
140 > be able to solve in such a session. The purpose here isn't to solve
141 > large problems, but to solve wide-spread 'little' problems. It has
142 > the additional benefit of introducing current developers to fresh
143 > blood, and vice versa.
144
145 Isn't there an initiative to allow users and developers to vote on bugs?
146
147 > To summarize, we are not looking to dictate development or assign duties
148 > to developers. We recognize a need for someone to assume some of the
149 > administrative roles, and we are offering to fill them. This project has
150 > made a great deal of progress by allowing its developers free reign to
151 > pursue their individual interests. No point in fixing this; it isn't
152 > broken.
153 >
154 > If there are any questions about any of the above, please voice them! We
155 > are always open to suggestions.
156
157 I think yours is a great post because, while I may not completely agree
158 with every point, this stuff should be discussed.
159
160 -f
161
162 > In regards to actually voting, what does everybody feel on this? Should
163 > there be a formalized election (such as the use of 'votify'), or should
164 > there just be a public vote on IRC or via e-mail? Since this hasn't been
165 > done before within the Gentoo for Mac OS X project, we're breaking new
166 > ground here. Please make your opinions known! Also, how much time should
167 > be allowed for other developers within the project to announce
168 > candidacy?
169 >
170 > - --
171 > Hasan Khalil && Lina Pezzella
172 > Ebuild & Porting Co-Leads
173 > Gentoo for OS X
174 >
175 > -----BEGIN PGP SIGNATURE-----
176 > Version: GnuPG v1.4.1 (Darwin)
177 >
178 > iD8DBQFC+tNSNJ9STR9DbYERAtjQAJ0YctzJGppbD54+5CNPxuuChNUKaQCfaJUS
179 > xatTyO7tr1UlPXsOsxjf/BA=
180 > =a8im
181 > -----END PGP SIGNATURE-----
182 >
183 --
184 gentoo-osx@g.o mailing list