1 <!doctype debiandoc public "-//DebianDoc//DTD DebianDoc//EN">
5 <title>PROPOSAL: A mechanism for updating Debian Policy documents</title>
7 <name>Manoj Srivastava</name>
8 <email>srivasta@debian.org</email>
10 <version>$Revision: 1.6 $</version>
12 <copyrightsummary>Copyright © 1998 by Manoj Srivastava.
15 You are given permission to redistribute this document
16 and/or modify it under the terms of the GNU General Public
17 License as published by the Free Software Foundation; either
18 version 2, or (at your option) any later version.</p>
20 On Debian GNU/Linux systems, the complete text of the GNU
21 General Public License can be found in
22 `<var>/usr/doc/copyright/GPL</var>'. </p>
26 <heading>Introduction, and Administrivia</heading>
28 This is a proposal for creating a process through which the
29 Debian Policy documents are to be maintained and updated, it
30 sets forth the processes, and also calls for the creation of a
31 team responsible for the task of updating policy, however,
32 this team does not act as author or editor of Policy itself,
33 that is the task of the Debian Policy mailing list.
36 It should also be pointed out that this proposal itself doess
37 not call for the modificxation of the Policy documents
38 themselves. I would rather not rush into anything as serious
39 as modification of the formal policy documents themselves, and
40 I suspect that we would learn and refine this process in
41 practice. I would rather that the formal modifications be
42 deferred until after the kinks of this process have been
46 Another thing that bears mentioning is that this proposal is
47 only for the every day routine functioning of the policy
48 group. Traditionally, the policy group, under the aegis of the
49 Policy editor, worked on the basis of a consensus derived in
50 the group. This proposal merely removes the need of a
51 dedicated policy editor, and passes the debian packages that
52 contain the policy into the hands of a few people who no
53 longer exercise editorial control, and, paying homage to our
54 growth, relaxes the requirement for a consensus.
57 This is not supposed to change the way the group works, except
58 in minor detail. There are some policy changes are light
59 weight and can be decided upon within the policy group, by
60 near consensus. In most day-to-day cases, the Policy group
61 should and must be able to conduct Policy discussions and
62 amendments without the intervention of the Technical Committee
63 or other Constitutional issues. Only in cases of extreme
64 dispute (formal objection) should the intervention of
65 Constitutional bodies come into play. In any other situation,
66 the Policy group should be able to conduct business
67 unfettered. This is the only way we can continue to improve
71 <em>In the following, the term developer refers to registered
72 Debian developers.</em>
74 <p>A copy of this document should also be found at <url
75 id="http://master.debian.org/%7Esrivasta/policy/"></p>
77 <heading>Deadline for tabling the discussion</heading>
79 I decided to use the suggested "usual" period of two weeks
80 for this proposal. Therefore, this proposal needs to be
81 acted upon before August the 22nd, 1998.
85 <heading>People Seconding the Proposal</heading>
87 Well, since Michael Alan Dorman, Phil Hands, and Richard
88 Braakman have volunteered to serve on the policy maintainer
89 team, I think they have no objection to being
93 <p>Michael Alan Dorman <email>mdorman@debian.org</email></p>
96 <p>Richard Braakman <email>dark@xs4all.nl</email></p>
99 <p>Philip Hands <email>phil@hands.com</email></p>
106 <heading>Archives and Personnel</heading>
108 <heading>The policy maintainers team</heading>
110 I propose we select/install a group of people who have
111 access to the CVS repository for the Policy documents;
112 however, this set of people behave more like maintainers
113 rather than authors/editors. This group does not create
114 policy, nor does it exercise editorial control, Policy is
115 decided "upstream". The group that decides on policy should
116 be the group of developers on the Debian-policy mailing
117 lists, which is how it was always done; so the group of
118 policy maintainers have no real power over policy. Since
119 they would have access to the CVS repository I guess it is
120 desirable that the people so appointed be ``mature'',
121 however that is determined.
124 I think that since the policy maintainers have no special
125 powers, there is no need to restrict their participation in
126 the discussion. We do need to have at least 4-5 people on
127 the job, preferably closer to 8, so that policy does not
128 languish when any maintainer goes missing (we do need
129 vacations, you know, once in a while), and since little
130 creative power is vested in the maintainers, we do not need
131 a central control. And the archives of the list can be used
132 as a record of the action decided upon even if all
133 maintainers are away at some time.
137 <heading>The CVS Repository</heading>
139 There should be a repository set up on
140 <tt>cvs.debian.org</tt> for this, with the people on the
141 policy maintainer team having write access to it.
144 The repository should contain all the packages under the
145 control of the team, and also should have an area where the
146 weekly status document is kept; once the document is under
147 CVS, it should be a simple matter to script exporting the
148 document out to a place where the web server can serve it,
149 as well as create the weekly posting to
150 <tt>debian-policy</tt> and <tt>debian-devel</tt> mailing
151 lists. This document could also be kept under CVS then.</p>
155 <heading>Procedures and Processes</heading>
157 <heading>Proposing amendments to the Policy</heading>
159 Unlike before, when the policy editor gathered in issues
160 which, in his view, were candidates for inclusion in policy,
161 I propose that issues are brought up in the policy group,
162 and, if the initial discussion warrants it, any developer,
163 with at least two(?) seconds can formally propose as a
167 The rationale behind the requirement for seconders is that
171 Encourage people to test the waters on the policy
172 mailing list, and this could help create an proposal
173 with a better chance of success</p>
177 Prevent frivolous or ill conceived proposals from
178 wasting peoples time (If the proposal does not even
179 convince two developers, surely this is not ready for
180 inclusion in Policy?)</p>
185 The whole discussion process is meant to be light weight; If
186 you wish the proposals to be amended, talk to the proposer,
187 and get the amendment in. Or else, post an alternative, and
188 let the group decide which one is better.
191 If the process gets very contentous, and needs something
192 like votes on amendments and withdrawl of proposal, then
193 this is not the correct forum for this, and the procedures
194 outlied in the constitution should be followed.
197 This document is not suppoed to supplant the processes
198 outlined in the constitution, nor is it an end run around
202 <heading>Notifications and Status Reports</heading>
204 Periodically, possibly weekly, a summary of current policy
205 topics can be posted to the Developers mailing list, as
206 well as to the policy mailing list. Since the BTS is used
207 for keeping track of policy amendments, the list of
208 current amendments shall always be on the web.</p>
210 Amendments to policy that have been accepted by the policy
211 group shall also be part of the notification. (recently
217 <heading>Deadlines for Tabling Discussions</heading>
219 It has been observed in the past that discussions on the
220 mailing list devolve into endless arguments. In order to get
221 away from the debating society aspect, at the time of the
222 formal proposal, a deadline can be set (probably by the
223 proposer, since they are likely to have an idea how
224 contentious the discussion is likely to be) for ending
225 discussion on the issue, which should rarely be less than 10
226 days, and typically two weeks or so. I hope that a hard
227 minimum period need not be set, and that the proposers would
228 be reasonable, and not set too short or too long a time for
232 If a consensus is reached by the policy group, then the
233 maintainers shall enter the amendment into the Policy
234 document, announce the inclusion in the periodic report, and
235 release a new version.</p>
237 <heading>Extensions to Deadlines?</heading>
239 If a deadline is approaching, and the discussion s almost
240 concluded (in other words, it has not reached an impasse),
241 and the consensus on the policy group is that an extension
242 of a week would resolve the issues better, a one-time
243 extension could be granted. Care should be taken in
244 exercising this option, since abusing this would merely
245 postpone closures. </p>
249 <heading>Deadlock resolution</heading>
251 Formerly, arriving at a consensus was the only way to amend
252 Policy. That worked well when the Project was small,
253 however, we have apparently grown out of that phase, and even
254 the policy mailing list has grown more fractious than in the
255 days of yore. We now need a formal process of deadlock
256 resolution, and we need to recognize that on non-technical
257 issues a small minority should not always hold up deployment
260 If the issue raised is especially contentous, or is deemed
261 to be suitable for review by the full set of developers,
262 then four or more developers can call for a hold on the
263 proposal, and move to send the proposal to the larger
264 developer body as a General
265 Resolution. <strong>Note:</strong> The constitution may have
266 additional requirements for submitting a General Resolution,
267 for example, a minimum number of seconders, etc.
270 If a consensus is not reached, (or if someone submits a
271 formal objection to the proposal) and the end of the
272 discussion period is near, then one is faced with a dilemma.
275 <heading>Impasse on Technical Issues</heading>
277 On technical issues, popularity is a bad way of arriving
278 at conclusions. Hopefully, the policy group would arrive
279 at a consensus on their own. If that fails to happen, or
280 if there is a formal objection raised on the issue, and
281 the issue is a technical one, then the technical committee
282 may be consulted. This should be a rare occurrence. </p>
285 <heading>Non Technical and Subjective Disagreements</heading>
287 However, if the issue is non-technical and subjective,
288 then a vote of the developers may be taken (USENET voting
289 software should be available all over the place, right?);
290 and a super-majority of 75% is needed to carry the
291 amendment through. Failing the super-majority, the issue
292 should be shelved. It may be re-submitted as a a fresh
293 proposal after a suitable cooling off period (which should
294 be no less than a month, typically three months being
295 desirable, unless there are significant new
296 developments). (Demote bug, if the BTS is being used)</p>
300 <heading>Using the Bug Tracking System</heading>
302 A fascinating sub proposal has been that we use the bug
303 tracking system to track policy amendments in progress. If
304 this is used, we may initiate discussions in the policy
305 group by filing wish-list bugs (note: this should be open to
306 anyone at all) This simplies how me manage and track open
307 amendments and issues. I think both retitling and the
308 severity of the bugs can and should be used.</p>
311 <tag>Issue raised</tag>
314 wishlist bug opened in BTS, with a subject of
321 developers may second the issue by emailing "seconded"
322 to the BTS. (Issue: what if the so called seconder is
323 not a registered Debian developer?)
329 when a propsed issue becomes a formal amendment, the
330 bug severity is raised to "normal" and the bug is
331 retitled to "[AMENDMENT DD/MM/YYY] ...". Actually it
332 might be better to close the proposal and reopen so
333 the bug date reflects when the clock starts ticking on
334 the bug; in which case the bug could simply be
335 retitled "[AMENDMENT] ...".
341 if the amendment is accepted, the bug is marked
342 forwarded, until it is actually integrated into Policy
343 and uploaded and moved into the archive, at which time
344 the bug is retitled "[ACCEPTED]..." and closed.
350 if the amendment is closed, it is retitled as
351 "[REJECTED] ..." and marked as closed
354 <tag>Deadlocked</tag>
357 if the amendment is deadlocked, it is marked as
364 I think that the Policy is critical enough for the project
365 that any real flaws in the policy be automatically be deemed
366 important bugs, unless they affect release management.</p>