If anyone sees any holes in this, it would be nice if they would point them
out soon. I'm going to try to get some version of this into tomrorows
distribution.
-Aris
On Sat, Feb 24, 2018 at 7:01 PM Aris Merchant <
thoughtsoflifeandligh...@gmail.com> wrote:
> Alright, this is a bit terrifying, but it's reasonably clean, and
> should catch all of the problems brought up so far, as well as any
> others of the same class. It involves a large amount of dark magic
> unfortunately, including a nested hypothetical (hence the title), but
> it's better than the alternatives that I can think of:
>
> - manually patching each error (and possibly missing some);
> - worrying about having an inconsistent gamestate for an
> infinitesimal period, as we would if we just retroactively redid the
> proposal "as if" it worked; or
> - undoing everything and starting over, potentially invalidating ~1
> week's gameplay.
>
> This is v1. I expect to make revisions.
>
> Fix procedure (to be passed by proposal): {{
>
> Amend Rule 105 by removing the sentence "Rule changes always occur
> sequentially, never simultaneously." [Generally, an important safety
> measure. In this case, it gets in the way of my fix. I could adopt
> this as a rule above the power of that rule, but amending it out for a
> little bit is simpler.]
>
> For the purposes of the rest of this proposal:
>
> - The "gamestate", includes the ruleset and all other parts of game itself;
> - An "effect" refers only to an effect directly prescribed by the rules;
> and
> - For something to "successfully" take effect is for it to have the
> result implied by its text, taking into account the context of the
> rules, but ignoring any rule that would directly block its action.
>
> No effect of the rest of this proposal causes it to fail or become
> non-existent or otherwise prevents it from successfully taking effect.
>
> Set the gamestate (including the ruleset) to whatever it would now be
> had every effect of Proposal 8014 been canceled and never occurred.
>
> [The above vastly simplifies record-keeping. No need to reason about
> what succeeded and what failed, because everything failed.]
>
> Set the gamestate to whatever it would now be if, in the infinitesimal
> period between the resolution of Proposal 8014 and the resolution of
> Proposal 8015, the gamestate had been set to whatever it would have
> been had Proposal 8014 successfully taken effect.
>
> [This portion is based on the observation that all inconsistencies of
> the proposal were internal to it, i.e. they did not persist after it's
> adoption. We can figure out what the gamestate should be at the end,
> presuming that they did indeed take effect. The problems are about
> errors like a rule existing momentarily without text, or it being
> unclear how many actions were being taken to get the rules into a
> certain state. All problems were in the transition, which this
> eliminates by setting everything all at once.]
>
> Amend Rule 105 to be in the same state it was before the passage of
> this proposal.
> }}
>
> -Aris
>
