A good stab.

“does not define legal actions or variables” might be too strong, though - we still want a piece of flavour text to be a “variable”, don’t we? They just have to be editable strings of characters rather than ruleset-recognised words or sentence fragments.

I’m not sure “direct and unambiguous” quite cuts it. If we had a careless stub rule of “a player who has 7 gems wins” (where gems were set up as a narrative but currently defined and had no other meaning) and someone named their flavour-text dog “7 gems”, should the victory rule be recognised as “unambiguous” and victory handed to the dog owner?

I think part of the problem is we’ve started using “flavour text” for two different things. On one hand, flavour text can be purely ornamental, eg an item description that has no effect. On the other hand, it can act as a label, eg the item/rule name. The similarity is that the actual string is devoid of intrinsic meaning, but one retains a function as an arbitrary signifier for something else. Might be best to separate these into two categories.

@Kevan that example sounds pretty textbook ambiguous to me! it could be interpreted multiple different ways, and a non-obvious meaning is being leveraged instead of the more obvious, probably intended interpretation.

@redtara huh, that’s a v good & insightful point there!! i would prolly vote for a separation of those two– maybe “null text” and “flavour text”, respectively?

Agreed that this is a good start, but I’d like to see something along the lines of what redtara suggests here. “Label text” or just “label” seems like a good candidate for the appendix.

It’s a good first step but I don’t think it’s quite right

re: “good start”/“first step”, its not a “first step” at all if it gets shot down! if we reject improvements bc they dont do everything they want, we’re not gonna get anywhere w/ this rule. do y’all think this is worse than what we have right now?

[redtara] Flavour text was initially created to address the exploit where a rule allows players to name objects at will, and they do so in a way that makes the ruleset react to something having a name we didn’t expect. Their brief has drifted a bit since, expanding to replace an earlier phrase about “names of rules are not themselves rule text and have no effect other than being rule names”, and then to cover wiki pages to (and thanks to Lemon for adding historical context in her vote) address an ambiguous rule in one dynasty.

I’m not sure it ever needs to be applied to anything other than freely-creatable text, really.

[Lemon] If it’s a stub rule where we haven’t otherwise defined gems, and all we have are dogs, then there’s really only one thing that “7 gems” can refer to - a dog with that name.

I agree with lemon that this is strictly better than the current provision.

Saying that flavour text does not define a variable seems wrong: a piece of flavour text is tracked gamestate that can be modified by proposals and rules. Perhaps people have different ideas of what a variable is.

@Kevan “defines variables” and “is a variable” are two separate things, to me.