Andy Bakun wrote:
what I'm suggesting for weighting mechanisms is _not_ just a
simple count of calls to gethostbyname() and shouldn't be
compared to each other solely on how many queries they spawn.
The main advantage of simple counts is that they are simple,
it's difficult to get it wrong in "setup wizards" etc.
Of course there are some cases, where you can't cache results,
and a smart metric could try to add more points for these
cases, e.g. in something like Radu's optimizer, or in future
FAQs, manuals, guides, and setup wizards.
But for the RfC I prefer a very clear and very simple MUST, no
excuses, no tricks, and shorter than in draft -00, or at least
not more text than in draft -00.
Radu's idea with 20 or 25 is shorter. Wayne's 3*10 limit with
a "minor" fix 15/10/10 results in a worst case 15 + 15*10, so
whatever Wayne does, it must be a simple single query counter.
But I suppose none of this is worthwhile if no one thinks
it's possible to change the way expensiveness of records are
calculated and still remain backward compatible
Not for a hard MUST, but for a RECOMMENDATION elsewhere it
could be a good idea.
Bye, Frank