Re: a vote in favour of ditching PE's

Although I am probably in the 'PEs for 0 or more declarations only'
minimalist camp, honesty compels me to say that David is right, that
implementing PEs without restriction on their occurance, except that
the RESULT has to be valid, is indeed easy in a lex/yacc environment.
I know, because that's how PEs are implemented in LT XML.  The only
thing that's still tedious is the complexities of not barfing on
references to non-defined PEs if they are inside IGNORE sections, but
that won't go away on the minimalist PE stance, either.

ht