Hi Julian and everybody, Le Sat, Aug 10, 2013 at 10:21:45PM +0100, Julian Gilbey a écrit : > On Sat, Aug 10, 2013 at 01:55:38PM +0900, Charles Plessy wrote: > > [...] > > In terms of quantity of work, I think that the documentation of the > > triggers is > > almost done. This said, the loud silence makes me feel that this work is > > very > > controversal, so I am not sure if it will be part of the Policy anytime > > soon. > > Either controversial, or alternatively, that people are happy with > what has been said.
Happy or not, if nobody elses seconds the patch, it will never be part of the Policy... It is definitely a difficult decision to second or not, because text in the Policy has more impact than general recommendations in less formal documents. For instance, the documentation of the Built-Using field created some confusion which may not have happened if it had not been documented. On the other hand, the Policy only made this field more visible, and we did not change its semantic when documenting it. Similarly for the triggers, we may get rotten tomatoes if the text is misinterpreted, but on the other hand, most of it is already there in a similar or identical form in /usr/share/doc/dpkg-dev/triggers.txt.gz, and nothing prevents us to further clarify in a revision of the Policy, so in my opinion, Debian would be benefit that a third developer would "conclude that the patch reflects the consensus and addresses the original issue". (https://wiki.debian.org/PolicyChangesProcess) > > Here are possible directions for our future work. > > > > 1) Release 3.9.5 as it is, convert to XML as 4.0.0, and resume the > > normative work. > > 2) Same as 1) but skip 3.9.5. > > 3) Same as 1) or 2), but close a bunch of low-hanging fruits first. > > 4) Work on multi-arch first. > > > > What do you think ? Do you have other propositions ? > > I don't have a strong opinion, except to suggest doing whichever route > will cause you the least amount of effort makes the most sense. > > A number of years ago, when I was very active on the Policy Group, I > proposed transitioning the policy's use of the terms SHOULD, MUST, > etc. towards the meanings of those words as they are used in the IETF > RFCs. I also proposed restructuring the policy document, as it had > become somewhat disorganised over the years. I (sadly) do not have > time nowadays to be involved in this, but I offer it as a proposal to > reconsider now during such a major piece of work as this. The conversion to XML would indeed be a good opportunity to restructure the document and review each occurence of words such as should, must etc. Indeed, for chapter 5 (on the syntax of control data files), I made a proposal in 2011 (http://bugs.debian.org/647570). Similar works on the other chapters would be welcome. A full review would also allow us to use straight lowercase for the RFC 2119 keywords, which would help as the use of upper case was controversial. Back on disussion on the contents of the next update of the Policy, what I wrote at the beginning reminded me that we need ensure that it enhances the description of the Built-Using field (http://bugs.debian.org/688251). Any volunteer ? -- Charles Plessy Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-policy-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130817070859.gc15...@falafel.plessy.net
