On Sun, Jan 12, 2020 at 5:45 PM Alexis Hunt via agora-discussion <agora-discussion@agoranomic.org> wrote: > > On Sun, 12 Jan 2020 at 19:05, Aris Merchant via agora-business < > agora-busin...@agoranomic.org> wrote: > > > On Sun, Jan 12, 2020 at 1:28 PM Kerim Aydin via agora-official > > <agora-offic...@agoranomic.org> wrote: > > > > > > The below CFJ is 3792. I assign it to G. > > > > > > status: https://faculty.washington.edu/kerim/nomic/cases/#3792 > > > > > > =============================== CFJ 3792 > > =============================== > > > > > > The above-quoted message contained a valid Promotor's weekly > > > report. > > > > > > > > ========================================================================== > > > > Gratuitous: > > > > Responding Arguments: > > > > > I'm pretty sure we've allowed this in the past by game custom. It was, > > I believe, part of the intent of Proposal 8245 to allow the Promotor > > to report on only the things at the beginning of the week. I also > > believe I've seen other officer reports with effective dates other > > than the time of their publication. At any rate, I've certainly done > > something like this on at least one prior occasion, and no one > > complained. [1] > > > > I do not believe that we can read too strongly into history here, unless > the possibility that this was a violation had actually been raised and > considered. Furthermore, I would argue that Proposal 8245 actually > indicates that the previous interpretation of that rule was that the > Promotor was bound to, at some time during the week, to distribute all > Proposals then in the Proposal Pool. It took a proposal to change the > interpretation to refer to a single point in time.
Sure, but that wasn't covered by the weekly report stuff. It's a different case entirely. > Additionally, I think Alexis's reading of the rule isn't mandated by > > the text. It is true that the Promotor's report includes "a list of > > all proposals in the Proposal Pool, along with their text and > > attributes". Certainly, I'll grant that the most common sense > > interpretation of the that is that it means the list of all proposals > > *currently* in the pool, but the word "currently" appears nowhere in > > the text, and must be supplied by the reader. I propose an alternate > > interpretation: the report being talked about is the Promotor's weekly > > report, and so it must be current as of some point *in the week* in > > which it is published. This interpretation has numerous advantages: It > > gives officers time in which to make sure their reports are accurate, > > by publishing and receiving comments on drafts, and thus is in the > > interest of the game. Furthermore, it is in accordance with game > > custom on the matter. I argue that these concerns are sufficient that > > we must set aside our first reading of the text and read it in the > > manner I proposed. > > > > The Promotor must, at some point in the week, publish the the abstract > information "a list of all Proposals in the Proposal Pool, along with their > text and attributes". There is no need for the word "currently" simply > because it is implied. English grammar treats this as an elided a > conjunction and copula here: "a list of all Proposals [that are] in the > Proposal Pool". A list of Proposals that were in the Proposal Pool > yesterday is not a list of Proposals in the Proposal Pool today. We would > require a clear contextual basis for a different time reference to be > inferred, like "a list of all Proposals [then] in the Proposal Pool". Really? There's an elided copula? You appear to be suggesting that the list of UK monarchs and the list of persons that are UK monarch are the same list. The British have a lot more current monarchs than they think they do. > Moreover, other officers' reporting requirements do contain clear > references to the present. Switches, for instance, include "the value of > each instance of that switch whose value is not its default value" in > officers' reports. There is no reasonable way to interpret this as "for > some point in time in the week, the value of each instance of that switch > whose value was not, at that point, its default value". > > Aris's argument can thus only be grounded in Rule 2143, and apply equally > to all reports of officers. But again, this is a massive stretch. "If any > information is defined by the rules as part of that person's weekly report, > then e SHALL maintain all such information, and the publication of all such > information is part of eir weekly duties." Aris's interpretation would mean > that an officer could, at the end of a month, publish the state of eir > tracked information as of the beginning of the month. This would imply that > they did not have an up to date copy of the information, in violation of > eir duties to maintain it. But the duty is clearly to publish the > information which they maintain, so Aris's interpretation would require an > officer to maintain up-to-date information but to publish outdated > information. Alternatively, Aris's argument is basically that a document > can be considered to be a Promotor's report at some point during the week, > and thus publishing it at another point meets the requirement. But Rule > 2143 makes no reference to a document being a report. A report, for the > purpose of Rule 2143, is merely a specification of the information which > must be published. The end of the month thing only works for monthly reports, which are generally of information that is extremely slow to change anyway. > Finally, Rule 217 does not permit us to consider alternative > interpretations on the same footing as the text. The text can only be > supplanted where it is "silent, inconsistent, or unclear". The rules are > none of these, and thus Aris's arguments about what is a convenient > interpretation are irrelevant. It is silent or unclear as to when the list must be as of. > In the alternative, if the text is indeed silent, inconsistent, or unclear, > Aris's interpretation is not in the best interests of the game, in keeping > with custom, or following common sense. > > Aris's interpretation would allow a situation where an officer reaches the > end of the month and realizes e still has to publish eir monthly report. E > has not been keeping eir report updated over time. E knows that e owes > Agora an updated accounting on the office, but e does not want to put in > the full effort. So instead of bringing the report fully up to date, e > brings it up to date only to the start of the month, and publishes that > instead, fulfilling eir obligations. > > Aris also cites numerous advantages, presumably under the basis that they > are in the best interests of the game. But eir arguments are weak. The > supposed advantage of allowing an officer time to make sure eir report is > accurate and circulating drafts is incredibly weak. Under my > interpretation, nothing stops an officer from circulating drafts and > requiring feedback, where required. E is free to do so. E also is allowed > to take as much time as e would like to ensure that eir report is accurate. > The only difference is the expectation that the officer, before publishing > the report, verify that no messages have been sent that would alter the > contents of the report. I do not believe that this is an unreasonable > expectation on an officer. Agora does not have the volume of messages that > would make this into a physically difficult limitation. So the only reason > an officer might realistically be concerned with this would be if e felt e > needed to recirculate another draft. Do you know how fast the proposals go when the game is fast? There might be 5+ new proposals between a draft and the time I can update it. You suggest that one might need to recirculate a draft like it's a theoretical possibility; it happens decently often (not a weekly occurrence by any means, but during the speedy surges of gameplay). > But the point of being an officer is to be the primary person responsible > for tracking some aspect of gamestate. We have an explicit process in place > to allow corrections of errors after the fact, and it works well. Of > course, there is a place for drafts, particularly when there is a large > amount of game state to catch up on. But this is hardly the norm, nor the > expectation placed on an officer. If an officer cannot confidently bring > eir report up to date for publication on a regular basis, e is not > fundamentally fulfilling eir duty. Furthermore, if every report is to be > drafted publicly before being published, then that requires every player to > review two copies of each report, even when the draft is correct. By > contrast, after-the-act corrections require only one report if they are > correct. Encouraging excessive drafting and excessively cautious officers > is not in the best interests in the game, nor in keeping with game custom. I disagree with you; officers make errors all the time. Trying to keep your official reports reliable through extensive drafting isn't excessively cautious, it's a legitimate policy position, and IMO a responsible one. > As a side note, I recognize that in the last few years I've played, draft > reports of proposal distributions have become more common. This is > explained, in part, because the consequences to the game are higher when > the Promotor makes an error and has to correct a distribution. It makes the > Assessor's job particularly difficult, and frequently requires all of Agora > to re-cast their votes. But distributions are not reports. In fact, while > custom is that the Promotor report in the same message that e distributes > proposals, the Promotor could even choose to use eir report as, in effect, > a draft distribution. > > There is another strong argument in favour of requiring reports to be > current as of their publication: time stamp consistency. Official reports > are an extremely important tool for players trying to understand the game > state. Getting an accurate picture of game state can be a rather trying > task at times. But, with rare exception, one can do this working purely > chronologically through the archives. The point of self-ratification is > that, when doing this, you can assume that a report is true as written, and > thus minimize the amount of gamestate recalculation to a single week in the > event of an error. Typically, when an erroroneous report self-ratifies, > gameplay that occurred after the ratification date will likely have taken > place under the assumption that the report was correct, meaning that > recalculation will not be meaningfully required. It's an argument, but not IMO a strong one. CoEs and CFJs make trusting official reports complicated; you might have to consult both a-o and a-b to figure out if a report ever actually self ratified. This is why trying to increase the accuracy of one's reports through public comment in advance is such a prudent policy choice: it means that the reader can have more confidence in the accuracy of a report, at the small cost of needing to take note of the effective date. Indeed, having an effective date before the actual date should be viewed as a stamp of quality and a sign of the extraordinary diligence of the officer. > But if a report could self-ratify at a past effective date, then we could > have a monthly report self-ratify nearly 38 days after its effective date. > This could result in significant gamestate recalculation. Even worse, > because the strict chronological sequence of self-ratifying reports is > broken, such a recalculation might involve intervening self-ratifying > reports, the effects of whose ratifications would need to additionally be > re-evaluated. We have had explicitly conflicting self-ratifying reports > before (CFJ 3337). The consequences of trying to ratify reports which > change the ratification of other reports could be rather serious. > Consequently, ratification with an effective date predating the ratifying > document should be reserved for rare circumstances. Not buying it; if the report self-ratified, no one CoEd it, which means people thought it was the truth anyway. Also, what's this business about reports changing the ratification of other reports? > Finally, Alexis's argument about ratification is grossly misleading. > > Any document purporting to be a report is self-ratifying, regardless > > of whether it actually is the report. No one questions that the > > document I published purports to be the Promotor's weekly report; it > > is accordingly self-ratifying. Many officers include the current day > > as the effective date of their report, and nothing would stop them > > from inserting a different day, which would allow them to perpetrate > > the same scam. All they would forfeit would be the penalties for > > failing to report and for publishing a false report, as well as their > > weekly wages. It is true that someone could catch the deliberate > > error, but that would be no harder than catching a scam would be if my > > report remained valid, as I claim. Thus, my interpretation does not > > make the game more vulnerable to scams. > > > Consider, for instance, a typical switch-trackor's report which contains > the state of some switches, as well as some history. At the heading of the > report, the officer writes "Effective date: yesterday". Then the entire > report is not self-ratifying, but only the portion which (implicitly) > claims to be the part listing each instance of the switch which is not at > default. If the "Effective date: yesterday" portion of the report is > clearly outside the portion purporting to be the part listing each instance > of the switch which isn't at default, then it is not included in > ratification. By Rule 1551, the ratification is thus effective as of > publication of the report, because the claim as to when the document would > have been accurate is outside the scope of the document to be ratified. The time isn't part of the information being ratified, it's determines when the minimal modifications for the ratification take place, a concept explicitly allowed by Rule 1551: "however, if the document explicitly specifies a different past time as being the time the document was true, the specified time is used to determine the minimal modifications". (Not all of my arguments here are very good; however, I do honestly believe that the rules are better if they allow for drafts without having to recirculate, and if I can convince the judiciary to interpret them that way, it would be lovely.) -Aris -Aris
