On 1/17/2020 9:08 AM, Aris Merchant via agora-discussion wrote:
> On Fri, Jan 17, 2020 at 8:57 AM Gaelan Steele via agora-discussion
> <[email protected]> wrote:
>>
>>
>>
>> Sent from my iPhone
>>
>>>> On Jan 17, 2020, at 7:08 AM, Kerim Aydin via agora-discussion
>>>> <[email protected]> wrote:
>>>
>>>
>>>> On 1/16/2020 7:45 PM, Aris Merchant via agora-discussion wrote:
>>>> Create the following power 3.0 rule entitled "The Reset Button":
>>>>
>>>> An officer CAN, without 3 objections, pursuant to a memorandum finding it
>>>> in the best interests of the game, issue an adjustment, which shall be
>>>> a public document. When the adjustment is issued, all changes that are
>>>> included in the adjustment take effect.
>>>
>>> One thing that had occurred to me was to tie this to the convergence concept
>>> in some way - I'd put the two concepts together in a single rule, the
>>> convergence text is kind of tacked on to R2143 and would read better merged
>>> with this (perhaps with something like "a convergence is a type of
>>> adjustment..."?)
>>
>> Be careful with convergence rules—while amending them, make sure you don’t
>> destroy existing convergences. If you do, you’ll force the Rulekeepor to go
>> back and figure out what happened in some historical messes.
>
> I'd really like to get rid of them; they're a mess and they'll never
> be used again. I'm not sure how to do that without breaking everything
> though.
So first: what history actually self-ratifies in a report? The current
state self-ratifies, but not the history - or am I missing something?
If the only issue is protecting an Officer from breaking a SHALL (in
particular, "the Rulekeepor SHALL record a historical annotation")?
I think the fix is pretty simple: Add the notion that historical records
must be "reasonably" accurate and don't self-ratify (for the rulekeepor
text, just adding "reasonably accurate" before "historical annotation" may
do the trick). We can either define that a little more in the rules, or use
precedent to define "reasonably" accurate such that the recordkeepor can
just note "uncertain - see CFJ XXXX" or "uncertain due to ratification" in a
historical record and not worry about it.
I'm fairly sure that, "recording the past as a convergence, even though
convergence concept has been repealed" would fit under "reasonably accurate"
(or we could add some slightly more explicit text as a safety).
Also fun to note: the current Rulekeepor rule seems a little broken if
you're going to be super-picky. Here:
Whenever a rule is changed in any way, the Rulekeepor SHALL record
a historical annotation to the rule indicating:
"Recording" is a single act. It implies that the Rulekeepor makes the
record ONLY at the time ("when") the rule is changed. If it later turns out
that the record was in error, it's no longer the "whenever" time, so a
perfectly logical outcome would be to say "the Rulekeepor didn't record the
right thing when the rule is changed, so e broke the SHALL, but there's
nothing to say e can change the record to be correct, now that e's recorded
it." (this has a real-world analogue: if I'm keeping a "signed" lab
notebook and record data incorrectly, I'm not allowed after signing to go
back and erase the data and put in the new values - I have to make a new
entry to say the previous entry was wrong).
The fix for that (if it needs fixing) is to replace "SHALL record" (as a
single act) with "SHALL maintain" (as an ongoing thing).
-G.
