Re: GR: Declassifying debian-private: second call for votes

[Ian Jackson <ijackson@chiark.greenend.org.uk>]

Charles Plessy writes ("Re: GR: Declassifying debian-private: second call for votes"):
> out of context, it is hard to chose between the options that each of you are
> presenting in this GR.
> 
> Could you briefly rebut each other's options ?  I think that it would help a 
> lot.

Of course.

I support both Option 2 ("Acknowledge difficulty", my proposal) and
Option 3 ("Remain private", Iain's proposal).  I firmly oppose
Option 1 ("Repeal previous GR", Gunnar's).

I think Option 1 is quite bad.  I will rank option 1 below the FD (ie
the status quo).  I recommend everyone else do so.


So, firstly, the big problem with Option 1:

If Option 1 wins, no-one can be sure what that collective decision
actually means.  Some people have said that they think that without
the 2005 GR, decisions about list management simply revert to
listmaster.  Other people have said that after a repeal of the 2005
GR, listmaster have no authority to invent a new declassification
process.

I think this would be worse than the current situation.

There have been a few concrete proposals (to -project and elsewhere),
for ways to improve things.  There have been some vigorous defences of
transparency (and vigorous criticisms of the status quo).  There have
been proposals (for example) to publish historical traffic data (which
is currently secret).

When this GR is out of the way, I think some of those contributors who
care a lot about improving our transparency will want to revisit this
issue.  They will say that in the absence of the specific directions
from the 2005 GR, listmaster has the power to regulate debian-private
(subject to the usual principles of consultation, in the
Constitution).  They will argue that the Social Contract requires
listmaster to do so.

I think all of these are quite reasonable points of view; and without
a clear statement from the GR about where the majority of the
projects' opinion lies, who is to say that these contributors are
wrong ?

I doubt that listmaster will be pleased to enter this fray.  Our
listmasters are sensible people who will not want to act in such a
controversial area, when their authority is doubtful.  So this
question will drag on with occasional rumblings, perhaps for years.
The dispute might finally be ended only by a second GR.

This will be stressful and threatening for those who value privacy;
particularly, it will be stressful for those who are concerned about
messages they have already posted, in the expectation that  the GR; it will be frustrating and
unproductive for those who want more transparancy; it will put an
unreasonable burden on listmaster; and it will risk wasting our
emotional energy on disputes about procedure and authority.

Please vote Option 1 below Further Discussion, or at least below
both Option 2 and Option 3.


Then, on to the choice between Option 2 and Option 3:

Both Options 2 and 3 explicitly honour the privacy expectations
established by the 2005 GR.  Both Options 2 and 3 retract the broken
promise of declassification.

Both Options 2 and 3 prohibit the unwanted declassification of
messages: that is, people are not, in general, at risk of their
messages being published against their expressly stated wishes.

The main difference is that Option 3 would make it impossible to
invent, or experiment with, new ways of handling -private in the
future.

That would be a shame.  There are some threads on -private which I
think the participants would be quite happy to see declassified at an
appropriate time (for example ones discussing security
vulnerabilities).  Several people have suggested forms of subject-line
tagging, for example, which might make that possible, while still
allowing people to post messages which will never be disclosed.

I don't know whether any such schemes are feasible, or would command
consensus support, or would be used in practice if they were
implemented.  But I would like it to be possible to consider them, try
out any that find favour, and either keep and improve them, or retire
them.

If Option 3 wins, such ideas are almost entirely blocked.  They would
require a new GR.  Developing a scheme in the context of a GR
Discussion is a bad idea.  Even if the scheme could be developed
separately and ratified by a GR, it would then be set in stone.


There is another more minor difference.  Option 2 leaves open the
possibility that something like the 2005 GR process could be used to
declassify particular messages or threads of historical interest, if
someone has the effort to do do so.


So I will be voting:

  2 > 3 > FD > 1

Ie, ranking the options in this order

  Option 2 "Acknowledge difficulty"
  Option 3 "Remain private"
  Further Discussion (aka Status Quo)
  Option 1 "Repeal previous GR"


I think there is a legimate possible justification for a vote of
  3 > 2 > FD > 1

If you feel that benefits of possible improvements to the transparency
of -private are negligible, or that they are outweighed by the risk of
madness on the part of listmaster, or even by the necessary
discussions (arguments) about the shape of such a scheme, then you
should rank 3 ahead of 2.

For you, then, Option 1 is very bad.  If you don't have confidence in
our current and future listmasters, not do do something bad, then
leaving listamster with a wide but disputed authority is precisely the
risk you would want to avoid.


I think people who are very keen on transparency should vote
along with me,
  2 > 3 > FD > 1

It is true that some people do think Option 1 transfers control back
to listmaster, and in theory would permit even retrospective
declassification (which some of the more vigorous transparency
campaigners seem to want).

However, in practice, even if you hold this view, you should be able
to see that it is unlikely that listmaster would engage in a
retrospective declassification exercise without clear authority from a
further GR.  Even if listmaster did want to, such a proposal would
need to be consulted on and would inevitably trigger a GR.

Indeed, Option 1 leaves listmaster with no clear authority even to
make new rules for the future use of -private.  Again, in practice, as
I say, listmaster is going to stay clear of this political minefield.

So if you want to improve the transparancy of -private, your best hope
is Option 2.  That could be followed by the development of detailed
proposals for future improvement.  Option 2 gives up the chance of a
more practical scheme for retrospective declassification, but opens
the door to developing (in a consensual way) more practical and less
secretive workflows for the future.

I think it is quite conceivable that some such proposals might gain
widespread support, at least so long as they retained the option for a
poster to -private to avoid future declassification.


Thanks for your attention.

Ian.


-- 
Ian Jackson <ijackson@chiark.greenend.org.uk>   These opinions are my own.

If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.