[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Policy as rule of law, or whatever



Hamish Moffatt writes ("Re: Policy as rule of law, or whatever"):
> On Tue, May 19, 1998 at 01:11:31PM +0100, Ian Jackson wrote:
> > I think my main problem with the `pro-strong-policy' arguments that
> > I've been seeing here is that they seem to imply an assumption that
> > policy is by definition correct, and that any point where it wasn't
> > the relevant policy document maintainer would agree at once.
> 
> This seems reasonable. If adopted,

That's very strange.  That part of my posting wasn't something I was
proposing should be adopted.  Rather, this part, which you didn't
quote, was:

] ] How about the following: we define and use MUST and SHOULD in some
] ] appropriate way, and then say:
] ] 
] ]   If a package violates a policy MUST [or the conditions for a SHOULD,
] ]   if any] then this is either a bug in policy or in the package.  A
] ]   maintainer who notices this while creating such a package should
] ]   report the bug against what they feel is the appropriate package.
] ] 
] ] Then we can use the bug resolution procedure to fix the problem.

Did you mean to agree with that ?

>  can we define policy to be binding
> rather than just recommendations/guidelines? I have real trouble
> with the idea of a policy which is non-binding (for Debian at least).

Perhaps we have a different notion of the word `binding', but for me
this has exactly the problem I described in the part you quoted and
said you agreed with.  Namely, that it implies that policy is always
correct.

It implies that if policy and a package disagree that there is some
formal presumption that policy is right - as opposed to an informal
presumption.  For example, it seems to imply that getting the policy
changed should be harder than getting the package changed.

My wording in my proposal above was very careful to avoid notions like
`binding', `must', and implications of power.  Rather, it merely says
that packages and policy should be consistent, and that if they're not
then one of them should be fixed.

Ian.


--
To UNSUBSCRIBE, email to debian-policy-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


Reply to: