On Thu, Sep 11, 2003 at 01:55:50PM -0500, Manoj Srivastava wrote: > You know, there is a reason that the severities document > specifically mentions the words "must" or "required"; because special > care is taken in inserting additions to policy where these words are > used, as oppsed to the "needs to", which may be a legacy wording. Then the bug severity definitions should be updated to emphasize that *ONLY* those clauses of policy which use specifically the words "must" or "required" can generate "serious" bugs. Other replies to my message indicate that I am not along in interpreting the Policy Manual as I have. Not everyone has as strong a background in reading RFCs as you have. > I understand that imprecision of speech is the norm in these > post modernistic times, Then presumably you are not averse to making the definition of "serious" even more clear. > but please do not extent such imprecision to this aspect of the > linkage between policy and the bug severity merely because you have > an axe to grind with another luminary (hi aj). That's not only unfair, it's backwards. Anthony's and my incompatible views are what led, in my assessment, to our conflict over the specific bug in question. I would not be so uncharitable to Anthony as to presume that he developed a theory of Debian quality assurance solely and retrospectively to justify his position on a single bug report. Likewise, I expect people, even you, to similarly refrain from leaping to such conclusions. My opinion has not changed. Bug severities should be coupled to their impact on our distribution, not to the stridency of language in the Policy Manual. It's the Release Manager's prerogrative to set technical release goals, which may include some degree of conformance with various requirements of the Policy Manual. It is neither the Policy Manual, nor the "serious" severity, that gives Mr. Towns the power to hold back the release of sarge until xutils gets an ELF object out of /etc. He is empowered by his office. (BTW, folks, the bug in question is tagged "upstream, help". Feel like helping?) -- G. Branden Robinson | I have a truly elegant proof of the Debian GNU/Linux | above, but it is too long to fit branden@debian.org | into this .signature file. http://people.debian.org/~branden/ |
Attachment:
pgpCmIV1h2oep.pgp
Description: PGP signature