
So, you have failed to address Mr. Newby's core question, which is "how do we determine the quality of any arbitrary PG edition?" I'm sure that any concrete proposal you would care to make would be genuinely appreciated.
You could allow critics of a particular work to identify real errors in PG works, and if the count exceeds some threshold number, then the work ought to be labeled "crappo." A "real error" would be some location where the work doesn't match a current coding requirement of PG, doesn't match the actual referenced work, doesn't match current and/or historical typographical standards etc. Of course some of the early PG works don't appear to reference any actual work -- which is a real problem. And of course, the high priests and priestesses will claim that which they do isn't a real error, but that which others do *is* a real error, which again is part of the problem. Working with "real" programmers, when you show some a bug they say "oh crap" and run off and fix it. But others you show them a bug and they will deny deny deny that it even is a bug. Which in turn begs the question of how the whitewashers determine that which is permissible or not... ...and the answer depends on the day of the week and the phase of the moon. Again, this takes you back to the original question: How does PG/DP actually identify errors and fix them after a work has been posted? Answer: They don't.