comex wrote: > Within the rules, terms related to permissibility such as 'may', > 'must', 'can', 'should',
Eek, you're going to have "can" canonically refer to permissibility rather than possibility? And "should" for permissibility instead of recommendation? What are your unambiguous terms going to be? Or did I misunderstand what you intended by "permissibility"? > have their ordinary language > meaning except that: So explicitly ambiguous, except in a couple of edge cases. I like precise definitions, which is what MMI is all about. We've been using the vague ordinary language meanings so far, and we've found them insufficiently expressive. Are you trying to move to precise meanings by gradually changing the definitions of these words in that direction? If so, I think it's an unworkable approach. I think it can only be done by the existing MMI approach: dual sets of terms (one precise and one loose), gradually switching the usages over to the precise terms, and then when everything is switched over the loose terms can be abolished (and the precise ones respelled). The basic problem here is that when switching to more precise terminology the rules that use the terms often need some rewriting. > * The terms 'must' and 'shall' assert possibility, and > additionally that failure to satisfy the condition is a > violation of the rule in question. Yuck, violation of orthogonality. And as discussed some weeks ago, not nearly as useful as you'd expect. Consider "the toastmaster shall initiate a toast to Agora as soon as possible": if relying on your implicitude, the toastmaster cannot initiate the toast if e is late in doing so. > Within the rules, logical terms such as 'if', 'or', etc. have > their mathematical meaning. This seems less troublesome than changing the definition of "shall" et al. I'm still more comfortable with separate terminology and an explicit switch, though. -zefram
