Re: Aptitude, ARs
Hi,
On Thu, Mar 13, 2003 at 02:58:30PM +0100, Javier Fernández-Sanguino Peña wrote:
> On Wed, Mar 12, 2003 at 05:16:17PM -0800, Osamu Aoki wrote:
> > Current official categorization ("Section:" entries in debian/control or
> > one in aj's override file) is too coarse. Daniel's new categorization
> > is OK but I think there are still rooms for improvement.
>
> IMHO bug #144046 should be fixed instead of trying to fix the
> package management tools. Someone said (Wichert?) way back when we
> implemented pools that the nice thing is that we could have arbitrary
> Sections: and a better approach could be taken in this respect. It has not
> been done, yet.
This bug should be fixed, I agree. But when and how?
Javi, I think you have right concept for the goal. We agree on this
goal. We need step-by-step plan to drive Debian toward this goal
without friction.
It is kind of funny that you always start saying "I have proposed this
problem with MY solution long time ago with bug #xxxx". Please think
that why this bug has not been fixed after 1 year. Tell me what stalled
this progress? (Do not blame on the unwillingness of others. Start
from the fact that you *failed* to *persuade* them. Sorry for my harsh
words but let's find a solution here. I know chances of success for my
attempt here is also not very high either. It is a real challenge.)
I think you are missing my intent of fixing these issues in aptitude. I
think first thing to do is to establish *proof-of-concept*
implementation of sorting category and keyword/attribute assignment.
That is what we are proposing here with aptitude.
(Here I really mean *proof-of-concept*. I know this can not go on this
way for ever in Debian.)
Many things in bug #144046 thread are realized or can be realized with
aptitude using local static data from user's perspective. I do not
think it went through scrutiny enough to say Daniel's implementation
(categorization) is the one to implement in new debian/rules entry as a
policy.
Really, once everyone reaches consensus on how these categorization or
key words are to be chosen and good examples over 90% of archives are
presented, it is just a matter of quietly introducing new "policy" with
new fields and some archive maintenance tools which provide "override"
capability in case package maintainer did not choose correct
categorization or key words. If we use new fields, we will be backward
compatible too.
I do not think we are there yet to put policy together. Hacking
aptitude's new Categorical Browser is a good initial step.
> PS: And we _could_ use a similar division as those used by other
> distributions which have tackled the issue already and do have
> Sections+Subsections...
Sure, but we need concrete examples or rule here. Can you point out
example? "like ..." is not enough. I need specific one for Debian.
I think it is easiest to communicate by implementing it in aptitude.
Regards,
Osamu
--
~\^o^/~~~ ~\^.^/~~~ ~\^*^/~~~ ~\^_^/~~~ ~\^+^/~~~ ~\^:^/~~~ ~\^v^/~~~ +++++
Osamu Aoki <osamu@debian.org> Cupertino CA USA, GPG-key: A8061F32
.''`. Debian Reference: post-installation user's guide for non-developers
: :' : http://qref.sf.net and http://people.debian.org/~osamu
`. `' "Our Priorities are Our Users and Free Software" --- Social Contract
Reply to: