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

Re: Ballot update (was Re: usr/man vs usr/share/man?)


        This is seriously flawed. As I have mentioned several times
 here, there are two issues involved: the programs that read /usr/doc,
 and the humans who are used to having docs in one place and suddenly
 having to look at two places during the transition.

        The ballot is seriously flawed too, since Joey's proposal
 about fixed programs should be implemented in conjuction with the
 symlink propoosal to address both isses.

        Since The point obviously did not get across, here's another
        There are two issues involved:
 a) The principle of least surprise, which is the frustration involved
    in reading the docs in two different locations. The symlink
    solution addresses this 
  b) incremental upgrades to unstable packages from unstable, which
    makes documentation not be accessable with tools such as dwww,
    man, ect. Joey's stable upgrades solution addresses that.

        The stable-upgrades solution has no impact on the former
 problem, and the symlinks solution only addresses the latter in a non
 optimal fashion. The symlink solution does have the side effect of
 making (b) above less critical, though.

        I never made (b) a part of my proposal either before the
 policy list of before cebian-ctte, since I felt that a) was important
 enough to require a solution, and it made (b) less time critical, and
 we could deal with it by ensuring that all the packages in potato
 dealt with /usr/share/doc.

        The symlink solution would ensure that partial upgrades to
 potato would work with man/info/whatever packages that had not been
 upgraded, and since potato should have fixed packages, partial
 upgrades from potato to woody would work as well.

        Partial upgrades from slink to woody would have problems, but
 adding stuff to slink-updates (Joey's proposal) would fix that.

        With the symlink proposal adopted, Joey's proposal would only
 be required for partial slink-->> woody upgrades.

        All that you say below only affects the problem (b)
 above. That is not good enough.

>>"Raul" == Raul Miller <moth@debian.org> writes:

 Raul> Potato includes release-notes which

 Raul> (1) warn users about the directory change and the requirement that they
 Raul> fix any local procedures or programs which only work under fsstnd,

 Raul> (2) asks that users who wish to do a partial update from slink to
 Raul> potato first install an "fhs" package which conflicts with versions
 Raul> of programs like man-db, dhelp, etc. which don't support the double
 Raul> directory standard,

 Raul> (3) if there are any packages (such as boa) which can't be modified
 Raul> for the double directory standard we go into potato with "known bugs"
 Raul> filed against them,

 Raul> (4) we ask that the policy committee relax the policy manual
 Raul> slightly to tolerate this transition period (perhaps saying that
 Raul> http://localhost/doc/<package>/<filename> is optional as long as web
 Raul> servers which do not support this mention the issue in their package
 Raul> description).

 Raul> Personally, I consider such release-notes a sign that we did something
 Raul> wrong, but (in my opinion) we already know that we did something wrong,
 Raul> and the problem is to fix it...

 Raul>                              * * * * *

 We prefer to speak evil of ourselves rather than not speak of
 ourselves at all.
Manoj Srivastava   <srivasta@debian.org>  <http://www.debian.org/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05  CC 2D 27 12 1D F5 E8 6E

Reply to: