--On 8 August 2006 05:13:37 -0500 Brad Knowles email@example.com wrote:
At 10:56 AM +0100 2006-08-08, Ian Eiloart wrote:
Right, but if we can't fix the problem of the multitude of broken MTAs out there, and the fact that most of them probably don't assign the appropriate extended response codes in accordance with the RFCs, then the likelihood is that we are going to be lead to make the wrong guesses based on the response we get.
We already do that. This is the problem that we're trying to solve, not a new problem introduced by the proposal!
No, that's precisely the problem -- the proposal does cause new problems that have to be dealt with.
Well, that's not true if the new default behaviour is the current broken behaviour. Would you accept that?
Because of all the broken MTAs out there, I believe that the probability is high that we will be unable to guess correctly what type of bounce we have for a statistically significant subsection of the population, and that the potential consequences of either a false negative or a false positive in this case are higher than taking the K.I.S.S. approach and not making any attempt to guess what type of bounce we're dealing with.
So, feel free to go ahead and make this change and to put this entire issue to rest, at least for the data you've collected from your site.