Al- > Yes, but your comments on this applied to section 3. > > We use the message number in the reordering singleton definition. We > also agreed to keep SrcTime as a mandatory parameter, and mention time > and bytes only as secondary means to determine order. When we > discussed this in Minneapolis last year, you came to the mic to agree > with this, even where we did not implement your comments verbatim (as > with the SrcTime parameter). A summary of comments and responses is > here: http://www3.ietf.org/proceedings/05mar/slides/ippm-1.pdf > > We're discussing a later section here (4.5), and you expressed no > objections about using time to quantify the extent of reordering, > AFAIK. Doh! This is what I get for not going back and looking at the draft. You're right that my objection was in terms of determining the out-of-orderness and not determining the extent or the characteristics this dynamic created. I do support the use of time for characterizing reordering events, just not for determining the event. But, somehow I misread the context of Bert's note. In any case, sorry for the spurious chud. I'll shut up now! :) allman
Attachment:
pgpnKezAUcncm.pgp
Description: PGP signature