Hi Ralph, Ralph Corderoy wrote on Wed, Mar 05, 2014 at 12:44:04AM +0000: > esr wrote:
>> I had a simpler design in mind. My premise is that once hygienic mode >> has been set we in general don't want or need to unset it, and that's >> going to be the very last thing a macro package does. > What about the case of one main macro package and several helper ones > that do a small thing well but aren't specific to ms, mm, mom, etc? > Is the invoker to know which padlocks the minibar and that it must be > specified last? Hum, thinking again, maybe i was too pessimistic in my previous posting and maybe there *is* a simple solution. How about having each (semantic) macro package declare, via .hygienic, which requests it considers fit for use by documents, but *without* calling .hygiene at its end, leaving that to the document? That way, it's completely trivial to use more than one macro set, but at the same time, documents having something like .hygienic dirty .hygiene at their beginning, or completely lacking .hygiene in their first two lines, are instantly suspicious. Yet, you are not driving authors towards designing contorted workarounds. You simply tell them to declare the amount of cheating they intend up front. That's also changing the smell of the whole thing a bit, away from "technical hindrance" more towards "help communication about document properties". Still not sure it is really needed, though; maybe mere validators are sufficient, after all. Yours, Ingo
