==============================  CFJ 3021  ==============================

    Tanner Swett's most recent ENDORSE vote caused him to vote AGAINST.

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

Caller:                                 Turiski

Judge:                                  G.
Judgement:                              FALSE


Judge:                                  G.
Judgement:                              FALSE

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

History:

Called by Turiski:                      05 May 2011 14:47:51 GMT
Assigned to G.:                         15 May 2011 23:31:35 GMT
Judged FALSE by G.:                     16 May 2011 03:49:03 GMT
Reconsideration requested by scshunt:   16 May 2011 05:54:14 GMT
Reconsideration requested by omd:       16 May 2011 17:41:33 GMT
Reconsideration requested by Roujo:     16 May 2011 17:54:23 GMT
Assigned to G.:                         16 May 2011 17:54:23 GMT
Judged FALSE by G.:                     17 May 2011 04:19:17 GMT

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

Caller's Evidence:

Tanner wrote:
>> 7051 2   Murphy        Janitor's Local #574
> ENDORSE the voter that most recently published the word "algebraic",
> not including me, as of the time I cast this vote


Turiski wrote:
>> 7051 2   Murphy        Janitor's Local #574
> AGAINST
>
>
> 4+sqrt(2) is an algebraic number.

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

Caller's Arguments:

The real question here is whether the wording "most recently published"
applies to the future as well as the past. Clearly, as nobody except Tanner
and me have posted the word algebraic in the last 2 days, I published it more
recently than any person who might have published it in the past; however, it
is unclear whether eir wording applies to my text, or even if it was eir
intention.

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

Gratuitous Arguments by ehird:

> as of the time I cast this vote

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

Gratuitous Arguments by scshunt:

UNDETERMINED, as we don't ever evaluate conditional votes
until the end of the voting period. While we know the outcome, that does
not change the fact that we don't perform the evaluation yet.

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

Judge G.'s Arguments:

Response to Disputant scshunt:
Your suggestion that a conditional vote is "undetermined" until the end
of the voting period would require a bit of hermeneutic dancing to get
around R683.  If a conditional vote is interpreted as undetermined at
the time the ballot is submitted,  it is not a valid ballot by R683(c),
as undetermined is not an "available" value.  For example (if a second,
unconditional, vote is later cast) it raises questions on the identity
of the "first" N valid ballots.  A better interpretation is that a
conditional ballot can be evaluated continuously throughout the voting
period, therefore remaining valid.  For example, here I would opine on
the value of the conditional at the time of the CFJ.  If the information
isn't available to evaluate the ballot at a given point during the
voting period (e.g. it is dependent on events following the ballot
submission) then the ballot value defaults to "not clearly identified -
PRESENT" by R2127 until the information is available.  This allows a
conditional vote to remain a valid ballot continuously during the voting
period.

Response to Caller Turiski:
It's pretty clear that the condition in question, by using "as of the
time", fixed the value of the conditional as being dependent on events
occuring prior to the ballot being cast.  Your later messages had no
effect on its value.

Response to Voter Tanner Swett:
As found by CFJ 1460 and reaffirmed since (most recently in CFJ 2998),
it's generally unreasonable to expect players to expend massive effort
in interpreteing an action.  In this case, a requirement to mount a
large-scale search of archives to interpret a vote does not qualify
under the "reasonably available" requirement of R2127.  Even if one
person (e.g. the Assessor) was willing to perform the search, the
requirement for public voting means that "reasonably available" must be
available to everyone[*], or in other words "available after what would
be reasonable effort for the typical Agoran player" (though if a person
performing heroic effort had published the results before the end of the
voting period, that also would have made it available to everyone).

In this context, I find that it's unreasonable to expect someone to
search more than the last several weeks of archives for an essentially
random term with no particular Agoran history or well-known context;
certainly expecting more than the current and previous months is
unreasonable.  I've searched the May-2011, April-2011, and (to be safe)
March-2011 archives and not found the term in question prior to your
ballot being cast, so this information was not reasonably available to
players during the voting period.   So by R2127 the vote cast (at the
time of the CFJ) was PRESENT - no further information came to light
during the voting period so this remained the value at the end.

FALSE.

[*] I'm surprised nobody has tested this by attempting to submit a
"secret" ballot using a conditional such as "I vote FOR if I sent
private message Y to the assessor, AGAINST otherwise".  Anyway, this
judgement firmly nixes that possibility.

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

Request for reconsideration by <function player at 0xb6d4d56c>:

I intend, with 2 support, to file a motion to reconsider this judgment
as the judge failed to consider:

Rule 2127:

       If a vote on an Agoran decision is submitted conditionally (e.g.
       "FOR if <X> is true, otherwise AGAINST"), then the selected
       option is evaluated based on the value of the condition(s) at
       the end of the voting period,

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

Request for reconsideration by <function player at 0xb6d4d56c>:

I support.

Also, I think the judge overestimated the difficulty of searching the
archives for a particular word.  I was able to do so by going to
agoranomic.org, clicking "public archives" next to agora-business and
agora-official, and on each page typing "algebraic" into the search
box and pressing search.  (If I had found no results, then the
requisite searching of earlier archives, some of which are
unavailable, would indeed be unreasonable; but it was published by at
least Murphy and Wooble in 2007 due to the proposal "algebraic
scores".)

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

Judge G.'s Arguments:

Response to Disputant scshunt:
Your suggestion that a conditional vote is "undetermined" until the end
of the voting period would require a bit of hermeneutic dancing to get
around R683.  If a conditional vote is interpreted as undetermined at
the time the ballot is submitted,  it is not a valid ballot by R683(c),
as undetermined is not an "available" value.  For example (if a second,
unconditional, vote is later cast) it raises questions on the identity
of the "first" N valid ballots.  A better interpretation is that a
conditional ballot can be evaluated continuously throughout the voting
period, therefore remaining valid.  For example, here I would opine on
the value of the conditional at the time of the CFJ.  If the information
isn't available to evaluate the ballot at a given point during the
voting period (e.g. it is dependent on events following the ballot
submission) then the ballot value defaults to "not clearly identified -
PRESENT" by R2127 until the information is available.  This allows a
conditional vote to remain a valid ballot continuously during the voting
period.

Further Followup:
It's important to note that this clause of R2127:
   "...the selected option is evaluated based on the value of the
    condition(s) at the end of the voting period,"
does not in fact prevent the conditional from having a determinable
value at other times prior to that; rather, it fixes the "final" value
of the conditional as it exists at the end of the voting period (as
opposed to fixing it for the time the vote is cast), with the purpose
of explicitly allowing events after the casting but before the end of
the voting period to affect it.  However, fixing this moment for the
final legal result does not prevent the vote in any way from having a
value at other times leading up to that fixing.

Response to Caller Turiski:
It's pretty clear that the condition in question, by using "as of the
time", fixed the value of the conditional as being dependent on events
occuring prior to the ballot being cast.  Your later messages had no
effect on its value.

Response to Voter Tanner Swett:
As found by CFJ 1460 and reaffirmed since (most recently in CFJ 2998),
it's generally unreasonable to expect players to expend massive effort
in interpreteing an action.  In this case, a requirement to mount a
large-scale search of archives to interpret a vote does not qualify
under the "reasonably available" requirement of R2127.  Even if one
person (e.g. the Assessor) was willing to perform the search, the
requirement for public voting means that "reasonably available" must be
available to everyone[*], or in other words "available after what would
be reasonable effort for the typical Agoran player" (though if a person
performing heroic effort had published the results before the end of the
voting period, that also would have made it available to everyone).

In this context, I find that it's unreasonable to expect someone to
search more than the last several weeks of archives for an essentially
random term with no particular Agoran history or well-known context;
certainly expecting more than the current and previous months is
unreasonable.  I've searched the May-2011, April-2011, and (to be safe)
March-2011 archives and not found the term in question prior to your
ballot being cast, so this information was not reasonably available to
players during the voting period.   So by R2127 the vote cast (at the
time of the CFJ) was PRESENT - no further information came to light
during the voting period so this remained the value at the end.

[*] I'm surprised nobody has tested this by attempting to submit a
"secret" ballot using a conditional such as "I vote FOR if I sent
private message Y to the assessor, AGAINST otherwise".  Anyway, this
judgement firmly nixes that possibility.

Further Followup on Reasonableness:
CFJ 1460 and the CFJs which follow from it generally invalidate all
actions that would require overly heroic (read: near-impossible) acts of
research for recordkeepors.  It is true that, for a recordkeepor who
offers to keep a particular record, a "simple" single-term search of
available archives going back multiple years is neither particularly
difficult nor heroic, and may not be considered unreasonable for a
recordkeepor given that e signed up for the task.

However, for conditional votes specifically, a higher standard is
explicitly contained in the Rules for reasonableness of effort.

Firstly, the R2127 clause in question enshrines the need for
reasonableness of effort by the double use of the term, and more
importantly by applying the term to "available", which is defined as
"present or ready for immediate use".  While the archives might be
immediately present, the search terms themselves are not immediate
unless the scope is narrowed to within a smaller number of locations
than the whole current archives moving backwards.  (I shouldn't imply
here that absolute time matters; if the vote said "I endorse the first
person who said X in the first three months of 2005" that would narrow
the field sufficiently).

Secondly, the availability must be to everyone, not just the
recordkeepor who signed up for extra work.  I won't claim this means
that the conditional must be so simple as to be interpretable by lazy
hurried morons, but it must not be available only to the most active.

Finally, and perhaps most importantly, R2127's mechanism itself is a bit
of a sneak.  It allows a power-1 Rule to affect votes by modifying the
term "clearly specified/identified" in R683(c).  Before R2127 explicitly
did so modification, a conditional vote wouldn't have clearly identified
an option.  This was done (by me, natch) on purpose to add conditional
votes at power-1, and it was controversial.  And it was implemented
before R754 was amended so that lower-powered redefinitions of higher-
powered terms were non-binding, precisely to prevent this sort of thing.
This re-definition probably wouldn't work today; in effect, functional
conditional votes at power-1 were grandfathered in, and as such, that
now non-binding re-definition of "clearly" can't be stretched too far.

What this all means is, quite simply, there is a considerable extra
requirement for the terms of "reasonably available" and "reasonably
interpreted" in R2127's conditional voting to mean "pretty darn
available and straightforward to figure out by most people who might put
a little bit of extra effort in."  Requiring the searching of "a few"
months of archives is reasonable (that took me 15 minutes including
checking backup fora and base-64 encodings).  Requiring the searching of
all of them back to 2007, by those specific standards of reasonableness,
isn't.

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

Judge G.'s Evidence:

Rule 2127/6 (Power=1)
Conditional Votes

       If a vote on an Agoran decision is submitted conditionally (e.g.
       "FOR if <X> is true, otherwise AGAINST"), then the selected
       option is evaluated based on the value of the condition(s) at
       the end of the voting period, and is clearly specified if and
       only if the value of the condition(s) can be reasonably
       determined (without circularity or paradox) from information
       reasonably available during the voting period.  If the option
       cannot be clearly identified, a vote of PRESENT is cast.

       Casting a vote endorsing another voter is equivalent to
       conditionally casting a vote whose value is the same as the most
       common value (if any) among that voter's valid votes on that
       decision.

       Casting a vote denouncing another voter is equivalent to
       conditionally casting a vote whose value is opposite to the most
       common value (if any) among that voter's valid votes on that
       decision.  FOR and AGAINST are opposites.

History:
Created by Proposal 4875 (Goethe), 1 December 2006

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