[gentoo-council] Amending GLEP39

[gentoo-council] Amending GLEP39

[Denis Dupeyron]

Next topic on the (long) list: how do we go about amending GLEP39?
This is by the way a blocker for part of the previous discussion about
meeting format, i.e. getting rid of proxies and slacker mark.

I will keep my opinion for myself until later in order to not bias the
discussion. I will try and present the facts as I see them but as
neutrally as possible. This is the result of lots of discussions with
lots of yous. Feel free to add to this, I'm not going to pretend I've
heard it all.

There are various schools of thought here, and we can divide them
using mainly two axes.

One is about whether we can actually modify GLEP39 or not. Some think
that GLEP39 is so fundamental that it can't be changed, period. A less
extremist faction think that although it can't be changed we can
decide to write another and switch to it if enough of the right people
agree (TBD, see below). And then there's those who think that GLEP39
can be amended under the right conditions (see below again).

The other axis is about how we agree we can amend GLEP39 or switch to
an entirely different text. Some think that as long as GLEP39 was
voted by the whole dev population, in order to change it we need all
devs to vote the change. Some others think that the whole point of
electing a council is about delegating your vote and making the
maneuvering of this large ship that Gentoo is a lot easier than if we
had to resort to all-devs vote for everything. The required majority
for each alternative (all devs vote, or council members only) is a
detail right now, but will need to be discussed at some point.

Then there's a sub-axis. In the case we allow the council members to
vote on GLEP39 changes to make things easier, would it be the same if
we were to switch from GLEP39 to a completely different text or would
this require an all-devs vote? And even in the case we're only
changing GLEP39, how major can the change be before we have to go back
to the whole dev population? Because it's certainly possible to change
GLEP39 to such an extent that it becomes an entirely different thing.
Can we actually quantify how big a change is?

There's one last group: those who don't care but will find an angle in
the above to make personal attacks. Please don't.

Denis.

Re: [gentoo-council] Amending GLEP39

[Ferris McCormick]

[-- Attachment #1: Type: text/plain, Size: 4943 bytes --]

On Mon, 13 Jul 2009 17:06:52 -0600
Denis Dupeyron <calchan@gentoo.org> wrote:

> Next topic on the (long) list: how do we go about amending GLEP39?
> This is by the way a blocker for part of the previous discussion about
> meeting format, i.e. getting rid of proxies and slacker mark.
> 
> I will keep my opinion for myself until later in order to not bias the
> discussion. I will try and present the facts as I see them but as
> neutrally as possible. This is the result of lots of discussions with
> lots of yous. Feel free to add to this, I'm not going to pretend I've
> heard it all.
> 
> There are various schools of thought here, and we can divide them
> using mainly two axes.
> 
> One is about whether we can actually modify GLEP39 or not. Some think
> that GLEP39 is so fundamental that it can't be changed, period. A less
> extremist faction think that although it can't be changed we can
> decide to write another and switch to it if enough of the right people
> agree (TBD, see below). And then there's those who think that GLEP39
> can be amended under the right conditions (see below again).
> 

I think it is clear that we can change or replace GLEP39.  The question
is who can do it and how.

> The other axis is about how we agree we can amend GLEP39 or switch to
> an entirely different text. Some think that as long as GLEP39 was
> voted by the whole dev population, in order to change it we need all
> devs to vote the change. Some others think that the whole point of
> electing a council is about delegating your vote and making the
> maneuvering of this large ship that Gentoo is a lot easier than if we
> had to resort to all-devs vote for everything. The required majority
> for each alternative (all devs vote, or council members only) is a
> detail right now, but will need to be discussed at some point.
> 

Council as we know it was created by GLEP39, and GLEP39 is the one of
several alternatives which was chosen by a vote of the developer
community.  So in a sense, Council is serving at the pleasure of the
developers, and GLEP39 describes the rules the developers want for
the council (or wanted in 2005 at any rate).

So it is conceivable that if council were to replace GLEP39, they would
be working against the explicit wishes of the community.  This leads
into your next sub-axis, which is a harder problem.  For now here, I'll
just say that "significant" changes to GLEP39 probably require a vote
of the developer community.  And I think GLEP39 supports that, in this
way:  Glep39 says "This proposal has nothing to say about GLEPs.", and
in particular, it does not address how it itself can be changed.  I
*think* that suggests that significant changes would require the same
process that got us to GLEP39 in the first place, but that is really a
question for Grant (g2boojum) and Ciaran (ciaranm).  That would be
primarily Grant, I think, because I asked ciaranm something about
GLEP39 once, and as I recall, he told me that Grant was the primary
author.

> Then there's a sub-axis. In the case we allow the council members to
> vote on GLEP39 changes to make things easier, would it be the same if
> we were to switch from GLEP39 to a completely different text or would
> this require an all-devs vote? And even in the case we're only
> changing GLEP39, how major can the change be before we have to go back
> to the whole dev population? Because it's certainly possible to change
> GLEP39 to such an extent that it becomes an entirely different thing.
> Can we actually quantify how big a change is?
> 

I don't know.  But we can infer a couple things.  The main thrusts of
GLEP39 are (1) Council must hold at least one open meeting a month
(Council must not be moribund and council must not work in secret), and
(2) To meet, council must have a quorum, and (3) council members must
show up now and then or send proxies.  I don't think council has the
authority to change fundamentals like those without concurrence of the
developer community.

On the other hand, we've already decided that rather than hold an
election when a council member leaves council instead we go to the next
slot on the ballot from the election which chose council.

So we have implicitly accepted that some kinds of procedural changes to
GLEP39 are OK, but I think that just replacing it would not be OK.  I
don't know how to quantify the dividing line, though.
> There's one last group: those who don't care but will find an angle in
> the above to make personal attacks. Please don't.
> 
> Denis.
> 

I hope that some of what I am trying to say comes through.  I'm not
being as clear as I would like, and I think that I am not sure about
how big a change has to be to trigger a vote of the developers.

Regards,
Ferris
--
Ferris McCormick (P44646, MI) <fmccor@gentoo.org>
Developer, Gentoo Linux (Sparc, Userrel, Trustees)

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-council] Amending GLEP39

[Denis Dupeyron]

On Tue, Jul 14, 2009 at 5:33 PM, Ferris McCormick<fmccor@gentoo.org> wrote:
> So it is conceivable that if council were to replace GLEP39, they would
> be working against the explicit wishes of the community.

The council is elected by the developer community to serve the
community's best interest. The developers chose the members they
wanted and could reject those they did not want. The way I understand
that is that all those who finished above the _reopen_nominations
level are to be considered trusted by a majority of the community.
Thus, your assumption that the council could be working against the
wishes of the community is equivalent to not trusting them, and in my
opinion should not be thrown into the equation.

> that is really a
> question for Grant (g2boojum) and Ciaran (ciaranm).  That would be
> primarily Grant, I think, because I asked ciaranm something about
> GLEP39 once, and as I recall, he told me that Grant was the primary
> author.

With all due respect to both of them, what they had in mind 4 years
ago matters much less than what we want to do for Gentoo in the
future. What they had in mind was influenced by the then situation and
a lot of things have changed. Grant and Ciaran are welcome to
participate to this discussion but I would prefer if you all gave your
own opinion on the matter, not theirs.

Anyway, as promised here's mine.

1- Yes, we can modify GLEP39. Gentoo is our project and we can make it
what we want. The only unknown is who and how.

2- GLEP39 was initially voted by all developers and is significant
enough that changes to it shouldn't be treated as lightly as any other
council decision.

3- The council members should be trusted by default and their smaller
number (compared to the whole developer community) enables a smoother
and faster decision process.

4- There is no way we will agree on how significant every change will
be, so we have to consider them all the same.

So what I would propose is that a unanimous decision from all 7
council members on each change warrants them to amend GLEP39. My
reasoning is that if all council members agree then it very likely
represents the opinion of the majority of developers who elected them,
and there's no point to resorting to an all-devs vote. In the case
where one or more member(s) would disagree then we have the natural
fallback to the process used for any other council decision: somebody
proposes that all developers vote on a change to GLEP39, and, after
discussion, if a majority of council members agree (either in a live
meeting or on the list) then we start the voting process.

This way we can maintain the smooth process for changes which seem
obvious enough, and we involve the whole developer community for less
obvious or more important decisions. And we don't have to decide in
advance how major or obvious a change is, the fact that we reach a
unanimous decision or not will speak for itself.

Denis.

Re: [gentoo-council] Amending GLEP39

[Andrew D Kirch]

While I am not a council member or a gentoo dev, I am a student of
history, and to that extent, I
would like to offer the following thoughts.

Denis Dupeyron wrote:
> On Tue, Jul 14, 2009 at 5:33 PM, Ferris McCormick<fmccor@gentoo.org> wrote:
>   
>> So it is conceivable that if council were to replace GLEP39, they would
>> be working against the explicit wishes of the community.
>>     
>
> The council is elected by the developer community to serve the
> community's best interest. The developers chose the members they
> wanted and could reject those they did not want. The way I understand
> that is that all those who finished above the _reopen_nominations
> level are to be considered trusted by a majority of the community.
> Thus, your assumption that the council could be working against the
> wishes of the community is equivalent to not trusting them, and in my
> opinion should not be thrown into the equation.
>   

>> that is really a
>> question for Grant (g2boojum) and Ciaran (ciaranm).  That would be
>> primarily Grant, I think, because I asked ciaranm something about
>> GLEP39 once, and as I recall, he told me that Grant was the primary
>> author.
>>     
My strong dislike for Ciaran is well known, however I have to agree here
to some extent.
I believe GLEP-39 was poorly written for the most part, and has resulted
in, during the few
years since, several "constitutional crises" of sorts.

Guiding principles must be preserved from any foundational document
(Magna Carta,
Federalist Papers, Declaration of Independence, Constitution, Bill of
Rights, and more
modern, the Communist Manifesto, Mein Kempf (sp?), the UN Charter, the
EU Charter, and
other such documents.  GLEP 39, whether we like it or not is a
foundation document for the council.
> 1- Yes, we can modify GLEP39. Gentoo is our project and we can make it
> what we want. The only unknown is who and how.
>   
I agree here, insomuch as GLEP39 is a GLEP and falls now under the
purview of the council.
> 2- GLEP39 was initially voted by all developers and is significant
> enough that changes to it shouldn't be treated as lightly as any other
> council decision.
>   
I concur here in scope as noted above, but would emphasize that we tread
carefully.
> 3- The council members should be trusted by default and their smaller
> number (compared to the whole developer community) enables a smoother
> and faster decision process.
>   
No governing body should ever be trusted by default.  You are trusted
with a position on council,
for this term, and if you abuse that trust, you may be gone next term.
> 4- There is no way we will agree on how significant every change will
> be, so we have to consider them all the same.
>
> So what I would propose is that a unanimous decision from all 7
> council members on each change warrants them to amend GLEP39. 
While this sounds like a good idea, it has never ended well
historically.  The Romans
tried two triumvirates (which required the unanimous assent of 3
separate rulers),
both ended in civil war and resulted in dictatorship and the end of the
Roman Republic.
It was tried by the French after the French Civil War, and likewise failed.

I would suggest a 5/7 supermajority instead and require that all 7
council members
be present to meet quorum for such a vote.
The problem here however is that GLEP-39 is a GLEP and should therefore be
governed by the rules surrounding a GLEP.  This would therefore mean a
simple
majority should be able to override the document as with any other GLEP.

I think my final suggestion would be that rather than amending GLEP 39, it
should be replaced by a new foundation document which was not hastily
constructed, which, while not being GLEP 39 still contains the guiding
principals of the original document.

There is no present, nor future, merely the past repeating itself over
and over again

-- 
Andrew D Kirch
Funtoo.org

Re: [gentoo-council] Amending GLEP39

[Roy Bamford]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2009.07.15 00:33, Ferris McCormick wrote:
> On Mon, 13 Jul 2009 17:06:52 -0600
> Denis Dupeyron <calchan@gentoo.org> wrote:
> 
> > Next topic on the (long) list: how do we go about amending GLEP39?
> > This is by the way a blocker for part of the previous discussion
> > about meeting format, i.e. getting rid of proxies and slacker mark.
> > 
[snip]

> 
> I think it is clear that we can change or replace GLEP39.  The
> question is who can do it and how.
> 
> > 
[snip] 

> > ... Some others think that the whole point of
> > electing a council is about delegating your vote and making the
> > maneuvering of this large ship that Gentoo is a lot easier than if
> > had to resort to all-devs vote for everything.
[snip]

> 
> Council as we know it was created by GLEP39, and GLEP39 is the one of
> several alternatives which was chosen by a vote of the developer
> community.  So in a sense, Council is serving at the pleasure of the
> developers, and GLEP39 describes the rules the developers want for
> the council (or wanted in 2005 at any rate).
> 
> So it is conceivable that if council were to replace GLEP39, they
> would be working against the explicit wishes of the community.

[snip]

So what ?
Gentoo is not a real democracy. The elected council have just had their 
mandate created - they have not even met once yet. 
Let them use it.

Most developers won't care. Look at the turnout for the council 
election. 

Developers that do care can lobby the council prior to any vote then 
council can vote taking into account any lobbying. 

Lets not pretend that we can keep everyone happy - we are all wise 
enough to know thats not true, so lets not even try and lets not put 
any more effort into fixing the slackers and proxies parts of GLEP 39 
than we really need to.

The difference is between the mathematical precision of a referendum 
and being close enough for all practical purposes. The council were 
elected to look after the engineering side of Gentoo so being close 
enough for all practical purposes is exactly what is required.

> 
> Regards,
> Ferris
> --
> Ferris McCormick (P44646, MI) <fmccor@gentoo.org>
> Developer, Gentoo Linux (Sparc, Userrel, Trustees)
> 

- -- 
Regards,

Roy Bamford
(NeddySeagoon) a member of
gentoo-ops
forum-mods
treecleaners
trustees
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.11 (GNU/Linux)

iEYEARECAAYFAkphvlQACgkQTE4/y7nJvatyxgCfTfKsg6x3g0ul7zk7QhHS79bL
I9EAnRxaNElBl26auypD6T4nOb2Zfg9M
=o23J
-----END PGP SIGNATURE-----

Re: [gentoo-council] Amending GLEP39

[Denis Dupeyron]

Roy,

> Gentoo is not a real democracy. The elected council have just had their
> mandate created - they have not even met once yet.
> Let them use it.
>
> Most developers won't care. Look at the turnout for the council
> election.
>
> Developers that do care can lobby the council prior to any vote then
> council can vote taking into account any lobbying.
>
> Lets not pretend that we can keep everyone happy - we are all wise
> enough to know thats not true, so lets not even try and lets not put
> any more effort into fixing the slackers and proxies parts of GLEP 39
> than we really need to.
>
> The difference is between the mathematical precision of a referendum
> and being close enough for all practical purposes.

I totally agree with all of this. I proposed the unanimity rule
because I know some council members strongly think that any change to
GLEP39 requires a vote by all developers. And I figured they might be
willing to accept a unanimous council decision instead. Now, if they
can be convinced that a super- or simple majority vote is enough then
I'm all for it.

On Sat, Jul 18, 2009 at 6:21 AM, Roy Bamford<neddyseagoon@gentoo.org> wrote:
> The council were elected to look after the engineering side of Gentoo

Here's a common misconception I'd like to clear. Nothing in GLEP39
limits the council's responsibility to the engineering side of Gentoo.
On the contrary, GLEP39 deals mostly with project management and
organization, developer activity, discipline, etc... However, the
successive councils slowly shifted to be mostly technical, if not
only. I have a theory as to why this happened, but it really doesn't
matter. I would very much like to fix that during this term because
not only does GLEP39 imply that the council should care about
non-technical aspects, but I'm of the opinion that if it doesn't
happen we'll be slowly rotting. I will be sending more propositions
about this to this list in the coming days/weeks. The first step is
deciding how we can amend GLEP39 though.

Denis.