On Tue, Sep 5, 2017 at 12:36 AM, Alex Smith <ais...@alumni.bham.ac.uk> wrote:
>> Lines beginning with hashmarks ("#") and comments in square brackets ("[]")
>> have no effect on the behavior of this proposal. They are not part of any
>> rules
>> created or amended herein, and may be considered for all game purposes to
>> have been removed before its resolution.
>
> It might be worth getting some confirmation on whether this actually
> works. (The last paragraph of rule 106 would suggest yes, but I think
> this is untested territory in Agora. B Nomic and BlogNomic both
> explored this space quite thoroughly, though, so it certainly can work
> in a successful nomic.)
I'll CFJ this.
> What about pledges? (Note that in the past, a pledge was just a
> contract with one member.)
If I'm going to follow your suggestion about paying for contracts, it
makes sense to keep pledges separate.
>> Repeal rule 2457 ("Lockout").
>
> This isn't technically tied to Organizations, but I agree with
> repealing it anyway just because it's likely to become a dead-end
> mechanic without them.
It's not doing much as is, either.
>> Create a new power 2.5 rule, entitled "Contracts", with the following text:
>
> Any reason behind the choice in power here? I vaguely remembered
> Contracts working at power 1.7, although looking through the actual old
> rules, they were at power 2.0. If we're going to have a power split
> between 2.0 and 2.5, it'd be helpful to have guidelines and reasoning
> for which goes there.
>
> Incidentally, my feelings on the matter are that contract rules should
> definitely be below the power of rule 2166 (but that's power 3 anyway),
> probably below the security limit of voting strength (which is 2.0; see
> rule 2422), and arguably below the security limit of Cards (1.7). I
> can't see any reason to take the power below 1.5, and as such, I'd
> consider 1.6 to be a reasonable Power for most contract actions.
>
> Note that "escalator mousetraps", where a player is forced into a
> contract via a scam-created rule, probably aren't worth worrying about
> much as it'd be easier to place the restrictions on the player
> directly. The exception is related to the use of contracts as Agencies,
> which is very powerful (and probably needs to be power 3 to function);
> that might need to be in a separate rule that specifically requires the
> player to have explicitly consented to the contract and any changes to
> it (or in a form of "constitutional law").
I have thought about it a little bit. Basically, I want contracts to
be bellow the power ~3 rules. I also want the core structural rules to
be above the other contract rules, but just by a little bit.
"Contracts as Agreements" needs to be more powerful than the highest
powered card rule, which is 2.0. Hence I decided to camp out the
relatively unused ~2.5 range. Any way I could satisfy you by tweaking
the power of other rules?
>> A contract is a document, and the ruleset described entity embodied
>> therein.
>
> Need something like "A contract is a document that has been made into a
> contract via a process described in the Rules", to stop random
> documents becoming contracts by default when the proposal is enacted.
> "Document" also has a rules-defined meaning, which may not be what we
> want; and the ability to amend a contract's text can easily interfere
> with several possible wordings (I'm finding it hard to make a good one
> where the contract /is/ its text). Perhaps better to make a contract
> /have/ text, rather than /be/ text.
Will this do: "A contract is a textual entity, and the ruleset
described entity embodied
therein. A document can only become a contract through the appropriate ruleset
defined procedures. Changes to the contracts text by rule defined mechanisms
do not change the identity of the contract."? I really like the "a
contract is just special text that can do stuff" abstraction.
>> Create a new power 2.5 rule, entitled "Parties to Contracts", with the
>> following
>> text:
>>
>> Contracts have parties, who are persons. The person(s) who create(s) a
>> contract is/are automatically a party/parties. Other persons CAN become
>> parties by announcement if the contract permits them do so. Parties can
>> leave
>> a contract by announcement, ceasing being parties, if the contract permits
>> the to do so. A contract CAN expel a party or group of parties, causing em
>> to cease being parties.
>
> "A contract CAN" is very hard to define; by what mechanism does it do
> so? Is it pragmatic or platonic? I'd recommend pragmatising this, so
> that some player (/ contract party?) has to explicitly expel the
> players by announcement.
I've added a distinction between "a contract can by announcement" and
"the text of a contract CAN". The former is pragmatic, the later
isn't. This case would fall into the former category.
>> It is IMPOSSIBLE, by any means, for a person to become a party to a
>> contract,
>> or for an contract to be created with a person as a party, without that
>> person's clear, willful consent. This rule takes precedence over any rule
>> that might make such a change possible.
>
> There's a potential loophole here: making an existing non-contract
> entity that has parties into a contract. In general, the terminology is
> unclear as to whether a contract is a "legal fiction" type of entity
> that the rules deem to have been created out of nothing from scratch,
> or whether a contract is some other sort of entity that's been made
> into a contract. (Proposals have similar issues but they are at least
> immutable, which makes things simpler. I don't want to repeat the
> issues here, though.)
How is this: "It is IMPOSSIBLE, by any means, for a person to become a
party to a contract,
for an contract to be created with a person as a party, or for an entity to
become a contract with a person as a party, without that person's clear,
willful consent. This rule takes precedence over any rule that might make
such a change possible."
>> Create a power 2.5 rule entitled "Birth and Death of Contracts", with the
>> following text:
>>
>> A person CAN create a contract by an announcement, specifying text. A
>> person
>> SHALL NOT create more than 3 contracts per week by this method, and the
>> Notary CAN destroy any excess (i.e. beyond the 3 permitted) contracts
>> by announcement in a timely fashion.
>
> CAN with a time limit? That's pretty unusual, although I guess it
> works. Also note that this rate limit is useless against players who
> are willing to break the rules if they think they can weather the
> resulting punishment; although I consider intentional rulebreaking to
> be something you don't do in games, we've had such players in the past.
> Spamming second-class players has traditionally been a very easy way to
> scam, and contracts have a tendency to turn into players in Agora more
> often than you'd think.
Oh, they're going to be second class players pretty soon. I've already
added most of the rules needed for that to happen. We can deal with
that problem when we come to it though; right now spamming contracts
is pretty useless.
>> Any public textual agreement or set of inseparably linked public textual
>> agreements between a group of two or persons, made with the intention that
>> the
>> agreement(s) be binding and governed by the rules, is a contract.
>
> Hmm, this seems like yet another definition of what a contract is.
> Also, when written like this, it arguably resurrects all the past
> Agoran conrtracts (which were made this way). Given that this is meant
> to be public anyway, it'd make things a lot simpler if creating a
> contract were done by announcement, rather than implicitly.
I borrowed that phrasing, which is why it's so similar (see principle
3). You're right though. I'm thinking about whether to yank this
entirely or rework it somehow, and leaning towards the former.
>> The person or persons who create a contract CAN and SHOULD also specify a
>> name for the contract; if e/they do/does not do so, the Notary CAN and
>> SHALL assign a name in a timely fashion.
>>
>> A contract CAN amend, destroy, or retitle itself if its text permits it to
>> do
>> so. A player CAN amend, destroy, or retitle a contract without objection,
>> even
>> if the text denies em the ability to do so. Players SHOULD only use this
>> mechanism to recover from situations where the Charter is underspecified or
>> has unintended effects.
>
> It's probably best not to allow contracts to perform actions
> platonically (i.e. without someone sending a message causing the action
> to happen). There's already been a couple of Wins by Paradox off that,
> and they were triggered in a much more dubious way than this very clear
> one. In general, "paradox escalation" is a real issue with contract-
> like rulesets, in which the contract creates an internally undefined or
> paradoxical gamestate, and then promotes its internal gamestate into
> something that the rules as a whole can see (such as contract text).
>
> I was excessively paranoid about letting this happen with Organizations
> and ended up making them unworkably clunky, but requiring players to
> trigger this sort of automatic contract action explicitly makes sense.
> Perhaps we could create a new mechanism via which contracts could take
> actions, requiring a party to the contract to announce that the
> contract takes the action /and/ that the contract's text says that it
> will take that action, and wording rules like "A contract CAN, by party
> announcement, …".
I'm just using "by announcement". The person doesn't even need to be a
party: "If a rule says that a contract CAN do something by
announcement, it is equivalent to saying that that any person CAN take
that action by announcement if the contract permits em to do so. A
person SHALL NOT cause a contract to violate a rule using this
method."
>> If a contract has fulfilled its purpose, does not specify any gamestate
>> affecting statements, or otherwise seems unlikely to be used, the Notary
>> CAN and SHOULD destroy it Without 2 Objections or with Agoran Consent. Any
>> player may destroy a contract with 2 Agoran Consent.
>
> I'm not convinced these are the right dependent actions. Imagine a
> player has a contract where they loan out assets and get repaid later
> (with all the loanees joining so that they can be forced to pay them
> back). Then the loanees just get out of the contract without 2
> objections, as only the lender will have an incentive to object.
>
> The standard should probably be "without objection" or "without party
> objection" (the latter might need defining). Incidentally, "without
> party objection" would mean that a 1-person contract could easily be
> dissolved; a "pledge" used to simply mean a contract that couldn't be
> amended or repealed by unanimous agreement of parties, and was used
> when a player wanted to make a promise that they couldn't back out
> from.
I'm already letting people get rid of contracts entirely by proposal
(it's unsecured), as a safety feature, so there's not much point in
making it harder than this. This procedure was mostly written for a
contract by a small group of players that somehow makes itself a
public nuisance. Given all that, would a SHOULD be satisfactory?
>> As an exception to the provisions of the previous paragraph and the
>> circumstances in which cards would ordinarily be appropriate, a person
>> awarding a card under this rule MAY consider the equitable interests of
>> justice and interests of the game, including the importance of the
>> observation
>> of contracts, as a mitigating or aggravating circumstances when awarding a
>> card. Such a person MAY and SHOULD also consider the instructions of the
>> contract or contracts in question when issuing a card.
>
> I think we'll want to bring back equity CFJs eventually (and overhaul
> SHALLs in other ways), but this will do as a starting point.
I was waiting for someone to come do that (I made some rather pointed
suggestions), but no one did, and I don't have the patience for
judicial systems. This is my, admittedly somewhat clumsy, solution.
>> Create a new power 2.4 rule, entitled "Acting on Behalf via Contract", with
>> the following text:
>>
>> A person CAN, by announcement, cause a contract perform any action the
>> rules
>> authorize that contract to perform, if the contract states that it is
>> possible for em to so.
>
> Ah, so this is how contracts take actions. It's typoed (missing a "to"
> after "contract"), and isn't clear enough IMO; it really needs to
> define "A contract CAN" not only so that it's clear how you can make a
> contract do something, but also so that it's clear that a contract
> can't perform such actions by, say, stating in its text that it
> performs such actions.
Added it, see above.
> Contracts would also have more autonomy if we allowed the action to be
> performed by a public message stating that the contract performed the
> action, even if the action wasn't performed by a person. (Last time we
> had a contract act as Promotor, we had the occasional proposal
> distribution sent from a server connected with the contract, and
> triggered by Googlebot.)
That's a very interesting idea, but I think I'm going to save it for
the second class persons proposal.
>> If a rule specifies that contract SHALL or SHALL NOT do something, it is
>> equivalent to saying that the parties to the contract are jointly
>> and severally liable to see to it that that thing is done or not done.
>> [Thoughts? Should I save this for when I actually write a partnership
>> proposal?]
>
> We can word this without needing to import existing legal language: "If
> a rule specifies that a contract SHALL or SHALL NOT do something, each
> party to the contract SHALL ensure that the contract respectively does
> or does not do that thing."
Done, thanks for the phrasing.
>> The text of a contract CAN permit persons to act on behalf of a party or
>> group of parties. It must specify:
>>
>> a. Which of it's parties can be acted on behalf of;
>> b. What actions can be taken;
>> c. Who can take the actions; and
>> d. Any conditions or limitations upon the actions. Such limitations
>> and conditions CANNOT be inextricable, and if they are,
>> the actions CANNOT be used.
>
> "it's" is a typo. I'm shocked that this is possible at power 2.4
> (although under the current ruleset, it's possible even at power 1!).
> Rule 2466 could definitely do with being secured at at least power 2,
> or there's a pretty trivial 1→3 escalator right there.
Typo fixed. I may secure that.
>> # 2.3 Contract Interpretation and Maintenance
>>
>> Create a new power 2.5 rule, entitled "Interpreting Contracts",
>> with the following text:
>
> If we're having a Power split in the contract rules, I'd prefer this
> one to be highest.
Done, now at 2.6.
>> The following are protected actions:
>>
>> 1. Deregistering;
>> 2. Submitting, pending, distributing, voting freely on, and assessing
>> a proposal, where the sole effect the proposal would have if adopted
>> is to create, modify, or destroy a contract or group of contracts,
>> or to cause an entity or group of entities to become or cease to be
>> a contract or group of contracts; and
>
> This is a little hard to read. "where" should probably be "but only if"
> to make it clearer how the parts of the sentence relate.
Done.
>> 3. Intending to destroy a contract, and supporting, objecting to, or
>> resolving
>> such an intent.
>>
>> [Can anyone think of any additions to this list?]
>
> It's worth considering adding registering (it's quite common for
> players who aren't very connected with Agora to make a
> contract/agency/promise/whatever that lets other players reregister
> them if they deregister, which seems problematic; and a contract
> forbidding registration seems broken).
Done.
> More importantly, a contract that forbids its parties from accurately
> describing the contract could be a huge problem, especially if the
> Notary ends up as a party. I'd recommend banning a contract from
> requiring false reports. (Sorry, G.) Along similar lines, I'd recommend
> banning contracts from requiring false judgements. (Sorry, BobTHJ.)
Done.
>> Rules to the contrary notwithstanding, a contract CANNOT compel, forbid,
>> or in any way alter, tamper with, or modify the performance of a protected
>> action. A contract CANNOT reward or punish a player for performing or
>> failing
>> protected action, or for doing so in a particular manner. A contract
>> also CANNOT allow a person to do any of the things prohibited to the
>> contract
>> by this paragraph. Insofar as a contract or a provision or clause of a
>> contract contravenes the letter or spirit of this paragraph, it is void
>> and without effect.
>
> "CANNOT allow" here seems broken; arguably that implies that a contract
> is void if it doesn't explicitly forbid performing all that stuff. I'm
> not sure what you're trying to block here, but when you figure it out,
> it needs rewording.
Changed to "enable".
>> Create a new power 3.0 rule, entitled "The Notary", with the following text:
>>
>> The Notary is an office, and the recordkeepor of contracts. The Notary
>> tracks
>> contracts, including their name, text, and parties. The Notary also tracks
>> the list of private classes of asset.
>>
>> The Notary's monthly report inludes all information which e tracks as a
>> part
>> of eir offical duties. The Notary's weekly report includes all changes to
>> this
>> information. [Should the monthly report be self-ratifying?]
>>
>> Make <someone> Notary. [Any volunteers? Maybe our current Secretary or
>> Superintendent?]
>
> Why power 3? I can't see a reason why this rule would need to be above
> 1, other than theme. (Also, we don't have a Superintendent right now;
> the previous holder resigned.)
Done, and o volunteered, which resolves that. (Thank's again o!)
> Self-ratifying contracts is probably a bad idea, incidentally.
> Contracts can come up with all sorts of ridiculous internal gamestate
> by themselves, it's probably worth reducing the stakes for reporting it
> inaccurately; and on the other hand, a misidentified contract is rarely
> a huge problem as it normally only promotes up into SHALLs. (Again,
> Agency-like behaviour could be a problem here.)
Agreed.
>> # 3.0 Asset Changes
>>
>> Amend Rule 2166, "Assets", by changing it to read in full:
>>
>> A contract's text CAN specify whether or not that contract is
>> willing receive assets or a class of assets. Generally, a contract CANNOT
>> be given assets it is unwilling to receive. If the contract is silent on
>> the
>> matter the procedure to determine its willingness is as follows:
>
> Should be "can", not "CAN". This could also do with anti-paradox-
> escalation technology (either in a separate proposal or this one).
Thinking about how to word that.
>> The recordkeepor of a class of assets is the entity (if any)
>> defined as such by, and bound by, its backing document. That
>> entity's report includes a list of all instances of that class
>> and their owners. This portion of that entity's report is
>> self-ratifying.
>
> We need some clarification as to what happens if a contract defines a
> class of assets without defining a recordkeepor for them. I know there
> were previous attempts in former Contracts rulesets to create gamestate
> purely so that the default-recordkeepor had to report on it.
Added: "Rules to the contrary notwithstanding, a contract CANNOT
oblige a person who isn't a member to record its internal state."
And also, in 2139, "The
Registrar is also responsible for tracking any switches, defined in a rule
or regulation, that would otherwise lack an officer to track them,
unless the switch is defined as untracked."
>> An asset generally CAN be destroyed by its owner by
>> announcement, subject to modification by its backing document. An
>> indestructible asset is one defined as such by it backing document, and
>> CANNOT
>> be destroyed except by a rule, other than this one, specifically addressing
>> the destruction of indestructible assets or that asset in particular; any
>> other asset is destructible. In circumstances where another asset would be
>> destroyed, an indestructible asset is generally transferred to Agora,
>> subject
>> to modification by its backing document and the intervention of other
>> rules.
>>
>> To "lose" an asset is to have it destroyed from one's
>> possession; to "revoke" an asset from an entity is to destroy it
>> from that entity's possession.
>>
>> An asset generally CAN be transferred (syn. payed, given) by its owner to
>> another entity by announcement, subject to modification by its backing
>> document. A fixed asset is one defined as such by its backing
>> document, and CANNOT be transferred; any other asset is liquid.
>
> To cut off CuddleBeaminess, "by announcement by its owner", surely?
> This is more important with contracts, because "A contract CAN" will
> probably end up defined, so how does "A contract CAN by announcement"
> fit into that?
Done.
>> To spend an asset is to pay it for the purpose of doing some other action
>> or
>> fulfilling an obligation; if the action would not be completed, the
>> obligation
>> would not be at least partially fulfilled, or more of the asset would be
>> spent
>> than is needed to perform the action/fulfill the obligation, then the
>> attempt
>> to spend fails. Generally, when an asset is spent, it is paid to Agora.
>
> The last sentence isn't precise enough IMO. Something like "If a rule
> specifies that an asset should be spent, but not the entity to which
> the asset is paid, then the asset is paid to Agora."
Done, though I added a generally in front, in case it disagrees with
common sense.
-Aris
