On 26-aug-2007, at 18:21, John Heffner wrote: [...]
It looks like our approaches are slightly different. I like to include all the information that could be useful in the protocol; you seem to favor a more sparse approach.
My philosophy is "overdesign and underimplement": make the design as complete as possible and then implement only those parts that you really need. This works a lot better than the other way around. Partially for this reason, I'm going to keep all of the mechanisms (except the switch advertisement) around but make as much of it optional as possible. Also, it's easier to discuss the merits of a protocol mechanism when that mechanism is laid out in the draft. We can then see if there is more that can be removed later on.
_______________________________________________ Int-area mailing list [email protected] https://www1.ietf.org/mailman/listinfo/int-area
