On Wed, 2010-01-13 at 19:58 -0800, Kerim Aydin wrote:
> But if this can be done over/outside precedence claims, wouldn't it mean that
> any (power-1) rule could do anything at all by saying "It doesn't actually
> do X, but causes the effects to to occur as if X happened."
In general, yes. That's why we secure things. If a high-power rule says
"Change Y cannot happen via mechanism X unless Z", then a low-power rule
can simulate the occurrence of mechanism X even if Z is false, by
applying change Y "by hand". If a high-power rule says "Change Y is
secured", though, there's no way round it; any change to the gamestate
that would get around it /is/ that change.
In the case at hand, "Change the rules as indicated by proposal N" is
possible at power 1 (so long as none of the rules changed are above
power 1, due to Power Controls Mutability). "Cause proposal N to pass",
however, is not possible at power 1, even though it does pretty much
exactly the same thing. However, neither "Change the gamestate such that
G. is not a player" nor "cause G. to deregister" is possible at power 1,
because changes to citizenship are secured at power 2.
In other words, you need to make a distinction between securing
mechanisms and securing effects. A power-3 rule can change arbitrary
gamestate (because no gamestate is secured at a power greater than 3).
As an example, assume there was a rather weirdly worded hypothetical
rule 9000/1 (power 3) that said the following:
{{{
If a proposal is submitted on 26 December, then if the Promotor does not
distribute that proposal on 27 December, the gamestate is changed
minimally such that the Promotor violated a requirement to distribute
that proposal on time.
}}}
Compare to the following, more sensibly worded rule:
{{{
If a proposal is submitted on 26 December, the Promotor SHALL distribute
that proposal on 27 December.
}}}
The second hypothetical rule has no effect due to 1769 (which has the
same power, a lower number, and which takes precedence over it); the
first does, because rule 1769 prevents the action of requiring a player
to perform an action during a Holiday, rather than the gamestate change
of a player failing to fulfil a requirement to perform an action during
a Holiday. However, the vast majority of high-power rules do not have
this problem, because they regulate gamestate directly; for instance,
rule 1030 prevents any change to the ruleset which would prevent it
being the main arbiter of precedence, rather than preventing a specific
mechanism of changing the rules so as to make it take precedence.
Likewise, rule 105 explicitly bans all mechanisms of changing the rules
other than its own; this prevents changes to the rules by direct
gamestate mechanism, as that would be a different mechanism to do so.
So to answer the original question: A rule banning doing X does not ban
other methods to achieve the same effect. However, nearly all security
rules do explicitly ban other methods to achieve the same effect, so it
isn't a problem.
--
ais523
