Dave, I tried to phrase my comment carefully. I wrote:
1. You succeeded.
2. For my own part, I had no doubt that your model of the world had something
like the rules you listed.
3. My reason for citing your note was that the issue of expected usage
environment is, I think, exactly what drives many people's design models and
exactly what, I think, can be a strategic danger.
In other words, I view this issue as specifically what should be discussed as an
essentially architectural topic, looking for consensus, guidelines, or the like.
It's acceptable, to me, to make an *engineering tradeoff* about how a
protocol design is balanced, if there's a natural operational environment.
What's not acceptable is to design it so that it *only* runs in one
environment or the other.
Over time, my experience has caused me to believe that protocols designed for
the high vagaries of a WAN can usually be adapted to run well over a LAN. In
considerable contrast, protocols designed for the high bandwidth and low
latencies of a LAN are difficult or impossible to make work well over the
vagaries of a WAN.
So while I agree with your basic view of engineering tradeoffs, I think that the
topic is filled with dragons.
Worse, I believe that, as a community, we don't have much basis for considering
the tradeoffs well. That is, as a community, I think we tend to lock ourselves
into designing for only one expected environment and fail to incorporate your
listed rules properly.
Indeed, that is what prompted my starting this thread. I think concern for
protocol chatter is not high enough (ie, not as high as it used to be.)
Ietf mailing list