Index ← 3336 CFJ 3337 3338 → text
==============================  CFJ 3337  ==============================

    scshunt is an eligible voter on the decision to adopt Proposal 7436.

========================================================================

Caller:                                 omd
Barred:                                 Murphy

Judge:                                  G.
Judgement:                              FALSE


Judge:                                  G.
Judgement:                              TRUE

========================================================================

History:

Called by omd:                          10 Jun 2013 03:33:22 GMT
Assigned to G.:                         17 Jun 2013 06:06:53 GMT
Judged FALSE by G.:                     19 Jun 2013 15:48:14 GMT
Reconsideration requested by G.:        19 Jun 2013 19:24:05 GMT
Assigned to G.:                         19 Jun 2013 19:24:05 GMT
Judged TRUE by G.:                      19 Jun 2013 19:24:05 GMT

========================================================================

Caller's Arguments:

This hinges on whether the gamestate implicitly contains a record of
history which rules that refer to the past state of the game use in
lieu of the actual past.  If so, Rule 1551's "the gamestate is
modified" also modified the historical record so that scshunt was a
player at the time of distribution of Proposal 7436, so e is in the
set of "those entities that were active first-class players at the
start of its voting period" (1950).  If not, e is not: if the decision
were purportedly resolved before ratification, the act of ratification
would itself cause the proposal to take effect (since, under the
simulated gamestate in which "the gamestate had been minimally
modified" (1551), the proposal took effect, so "the gamestate is
modified to what it would be" if it took effect).

My view is that "were" should mean "were": although modifying the past
is convenient in some cases and prevents some wonkiness in this case,
it contradicts the plain meaning of the rules, and adds unnecessary
moving parts, when usually there is not enough delay to require us to
decide on what the past was this far into the future.  To quote my
notes on Proposal 7419:

> [See
http://www.mail-archive.com/agora-discussion@agoranomic.org/msg26334.html.
>  There seems to be some confusion over whether a certain past
> judgement implies that the gamestate currently includes history,
> though that may be my fault, but I think the issue ought to be
> legislated.
>
> Not including history in the gamestate has the potential to cause
> weirdness when incorrect documents are ratified, because the true,
> ratified-away value can still affect the game if rules refer to the
> past, but the only case I can think of where this has actually
> mattered is ratifying victories (no longer an issue as you now only
> win the game after a victory announcement has ratified, and even if
> the gamestate includes history, "our record of history is modified
> such that you won the game" is a dubious substitute for an actual
> win), and it has the benefit that it's much easier to work out what
> happens when rules and proposals try to do weird things with the
> gamestate as a whole.  As I mentioned in that thread, whether the
> previous, confusingly-worded version of ratification was effective at
> doing anything might depend on whether this implicitly defined archive
> is "stored" as a list of instantaneous states or a list of changes;
> that could be legislated and I might be overblowing the ambiguity, but
> I think it's unnecessary mind screw overall.

Note that the current (2011-) version of ratification is effective
regardless of this format.

========================================================================

Gratuitous Arguments by scshunt:

The issue is more complicated than omd suggests. There are complex
elements of game state that are relevant, including three elements in
particular:

- The set of eligible voters
- The set of valid ballots
- The outcome

I submit that these things are rather clearly elements of the game
state, and hence are undoubtedly within the reaches of a ratification.
Some may, in certain circumstances, have a value that is not
immediately discernible solely by taking the rules and some other
subset of game state, and performing a calculation. This indicates to
me that they are properly considered as game state and not
definitions. In particular:
 - The set of valid ballots may change by virtue of some effects such
as veto which, while not present in the rules, have generally been
considered functional
 - The outcome may change by announcement in the case of a tied vote

All of these are defined by rules of power 3, but theoretically can be
overridden by the power-3.1 Rule 1551. Moreover, at the time that the
ratification is performed, it must be the case that Rule 1551
overrides them, as otherwise the gamestate would not be set to what
Rule 1551 specifies.

The sole question then, is whether the power-3 rules "reassert"
themselves and reinstate their versions of the game state over top of
the papering performed by Rule 1551, or not. However, we have a
precedent here. In CFJ 3289, it was determined that an instantaneous
effect (setting the power of a rule) taking precedence over a
continuously-defined effect (the minimum and maximum power of a rule)
does not allow for the continuously-defined effect to reassert itself
over top of the instantaneous effect. While this issue was not
directly addressed in the judgment, it implies that a value defined in
the manner of being a continuous effect should in fact be construed as
both:

1) An instantaneous effect setting a value
2) A continuous prohibition against changing that value.

========================================================================

Gratuitous Arguments by omd:

This is similar to CFJs 2277-8, where there was a debate
over what happens when one rule provides a definition of a value in
terms of something else, and another purports to allow changing it;
due to a trivial outcome for the CFJ, I'm not sure that was ever
resolved.  However, I believe, as I did then, that that constitutes a
redefinition of the value, in that case as N higher than it would
otherwise be, in this case as what the vote collector chose in case of
a tie, and whether the redefinition is in effect and its parameters
are in the gamestate.  This does not apply to rubberstamp (not veto,
at least recently) changing the set of valid ballots, however, since
its wording is quite clearly a redefinition:

      among the otherwise-valid votes on a decision that has been
      Rubberstamped, only the first AGAINST vote per voter is valid.

========================================================================

Gratuitous Arguments by omd:

The reports I used for those dates were Registrar and
IADoP's reports: the Registrar's for obvious reasons, but the IADoP's
because it listed scshunt as belonging to one or more offices, and
this is self-ratifying:

      The portion of a public message purporting to be an IADoP's
      report that lists the holder of each office is self-ratifying.

Rule 1006 (Offices) defines the holder of an office as necessarily a player:

      An office is a role defined as such by the rules.  Each office
      is either vacant (default) or filled (held) by exactly one
      player.

and ratification "cannot add inconsistencies between the gamestate and
the rules", so scshunt and I concluded that the minimal modification
required to make the IADoP's reports "as true and accurate as
possible" would be making scshunt a player.

========================================================================

Judge G.'s Arguments:

First, I accept that the Registrar's reports for a while ratified
scshunt as a non-player, and the IDAoP reports ratified scshunt as a
player, as this represents a minimal set of changes in the game state to
conform with the reports.  I have a bit of disagreement here in fact,
but I'm likely in the wrong and so will let it pass and deal with the
possible ambiguity legislatively (legislatively, because ratification of
the Registrar's report inarguably can induct someone against R101, which
demands attention).

Also, took a good look at what records would be affected or not affected
by scshunt "blinking in and out" via ratification.  Asset reports have
ratified so no LF&D reclamation is evident.  However, the following
things have NOT ratified:
 1.  Status as an elder;
 2.  Office Assumption status (that is NOT part of "the holder of
     each office" that is self-ratified by R1551);
 4.  Continuity of officeholding for the purpose of Yaks.
(there may be others)

I agree with scshunt that history CAN be subject to ratification, but
*not* that it is subject to self-ratification.  That is, the self-
ratification of a report fact ("as of this date, scshunt was a player")
does not ratify the state of play the instant before that report was
made.  In particular, R2139 explicitly only governs the set of players
at a particular instant (the instant of the report), and R2138 only
governs the older of each office at a particular instant.  The strict
wording of *exactly* what is self-ratified (state) precludes any
ratification of when or how that state came to be, historically.

An explicit ratification process could do so, by (e.g. without
objection) ratifying that the statement in question was true at an
earlier date from the report.

I'm really not sure how this whole thing pertains to "continuous"
versus "instantaneous" values that scshunt and omd discuss in the
arguments, so I'll leave it to them to motion it back to me if I've
missed something esoterically vital.

In any case, R1950 only cares about player status at the time of the
start of the voting period, for Prop 7436, the ratified state of the
Registrar's report was that e was NOT a player.  FALSE.

========================================================================

Request for reconsideration by :

[Re]-judgement of CFJ 3337.

========================================================================

Judge G.'s Arguments:

BREAKING NEW EVIDENCE!!

Above, my arguments assert that "self-ratification" of the specific
Registrar's and IDAoP reports affect state but not history, but
ratification in general CAN affect history.  New evidence shows that such
ratification without objection took place here, before this CFJ was called:
http://www.agoranomic.org/cgi-bin/mailman/private/agora-business/2013-June/031
344.html
and so scshunt was in fact an eligible voter.  I judge TRUE.

-G.

========================================================================