On Sun, Feb 11, 2018 at 07:29:37PM -0500, Rich Freeman wrote:
> On Sun, Feb 11, 2018 at 7:12 PM, William Hubbs <williamh@gentoo.org> wrote:
> >
> >> Appeals sometimes reverse decisions because these lower groups are
> >> imperfect at enacting the policies set at the top, or they are
> >> operating in areas where no precedent exists.  These reversals
> >> shouldn't be seen as some kind of checks/balances system that adds
> >> value, but an inefficiency that wastes time deliberating matters more
> >> than once.  It is necessary only because it would be even more
> >> inefficient to slow everything down to a pace where one small group
> >> could deal with it all.
> >
> >  I agree that the higher body should not be involved in every case;
> >  However, I absolutely do not agree that appeals are not a
> >  checks/balances system. If someone appeals something it means that they
> > feel that the decision made by the lower body needs to be re-examined.
> > If the higher body then overrules the lower body, it isn't meant in a
> > shameful way, it is just guidance for the lower body in the future.
> 
> Checks and balances are when two bodies are allowed to be in
> opposition, with neither body being superior to the other.  In the US
> system the three federal branches operate in this way for the most
> part, with each branch able to block certain actions of the others.
>
> An appeal isn't a check and balance.  An appeal is a superior body
> having the opportunity to overrule the action of an inferior one.

Ok, this makes sense, but my point still holds.

If enough of the members of the inferior body are members and able to
vote on the appeal in the superior body, there is no reason for anyone
to appeal, and if we are going to do that, we should kill the ability
to appeal entirely.

> >
> >> So, if there were no QA or comrel, and there were just the council,
> >> and it handled everything and there were no appeals at all, this would
> >> not lower the quality of decisions, but it would actually raise them
> >> (since some incorrect decisions might not be appealed).  However, it
> >> would come at a cost of a lot less stuff getting done since you'd have
> >> reducing the pool of labor.
> >
> >  Rich, I don't follow this logic at all.
> 
> What is confusing about it?  Imagine that the Council dissolved both
> QA and Comrel, and directly handled both?  The main issue with this is
> that stuff would probably get neglected, but ultimately it is the same
> body that is making the final decisions.
 
 This still doesn't make sense.

Another thing to consider is, 
Comrel and QA members are already expected to recuse themselves from voting on
appeals from their projects at the council level. This means the council
that votes on appeals is different than the council that votes on other issues.
Also, council members are allowed to abstain from votes, and this
shrinks the voting pool further.

> > I know about the appellate courts, but there are other levels as well.
> > You would never find a district courte judge on an appellate court
> > simultaneously, and you would never find an appellate court judge or
> > district courte judge serving simultaneously as a justice on the Supreme Court.
> 
> As far as I am aware there is no provision in US law that prevents
> this.  It is just impractical, and would defeat the point of
> delegation.
 
 Do there have to be laws that prevent it? There are no laws that
 prevent it, but it doesn't happen. If someone did try this, I'm sure
 they would be shot down because of the perceived conflict.

> As I recall there have been complaints made on the lists that the
> leaders on the Council need to do more to fix problems actively vs
> just waiting for people to come to them for decisions.

This topic deserves a totally separate thread, but I will say here that it
depends on how you feel about how Gentoo should be lead. Some have said
that the council should be treated more like a dispute resolutions body
than a leadership body. I have heard a lot of talk about how innovation
comes from the developers and the council should stay out of the way
until a decision is requested from the community.

> I think this
> is the main reason why Council members ended up in lead roles on other
> projects.  Some project was considered to need help, and a Council
> member stepped into try to strengthen it.  I'd be careful about
> banning this sort of practice, because then the only thing the Council
> could do if Comrel or QA were inactive would be to whine about it on
> the lists until somebody else stepped up to fix things.

Don't even get me started. ;-)

> In any case, that's my opinion.  I suspect it might not be a majority
> opinion and that is OK.  The world won't end if a few more critical
> Gentoo projects go idle...

This is also a completely separate subject, but imo there are several
critical tlps that are idle.

William