omd wrote: > I don't remember the judgement as well as I should have, I guess... in > that case, doesn't the argument hinge on something as inconsequential
Not inconsequential if it, well, has substantive consequences: > as the format in which states are saved in the gamestate? ais523 > seems to be making the argument that each state is stored separately > (but at the current time, the gamestate gets continually cloned onto > the next one), so you can modify some past archive to say that the > document was accurate without any effect on the infinite other > archives or the present, and clearly this is the minimal modification; Yes, and thus e parsed R1551/(13? 14?) as claiming "the present archive is modified such that the past's archive is modified" which is clearly impossible, rather than "the past archive is retroactively modified" or "the present archive is modified to include the legal fiction that the past's archive was modified" or (R1551/15) "the present archive is modified to what it would be if the past's archive had been modified in the past" (presumably including the legal fiction that the past's archive had been so modified) This used to be further complicated by "this does not retroactively change the possibility or legality of past actions", but that was amended away at some point. (As usual, issues of legality can be addressed by adjusting sentence as the judge sees fit.) > but it makes more sense to me to represent it as a list of changes, > like RCS, where the minimal modification is exactly one change, not a > pair that cancel each other out. Maybe; the usual conceptual tradeoffs apply (we use periodic snapshots in practice to minimize computation). ais523's argument can still be expressed within this metaphor, though: "a change is made now such that a change was made in the past" rather than (etc.) > I'm pretty tired right now, but is there already a CFJ directly about > this gamestate issue? If not, anyone mind if I call one and get this > properly decided? Not that I know of. No.
