Re: Terminology [Re: Some Comments on ID/Loc Separation Proposals]

[jnc@mercury.lcs.mit.edu (Noel Chiappa)]

    > From: "marcelo bagnulo" <mbagnulo@ing.uc3m.es>

    > could you provide a definition of what do you mean by end system? would
    > it be similar to JNC's endpoint definiton? or NSRG stack definition?
    > I mean, are those three different names for the same concept or do we
    > have a conceptual discrepancy here? As far as i can understand it, we
    > only have one class of objects and we are just discussing about the
    > name to name it.

I think that's basically the case.

See, I don't recall seeing any cases of confusion caused by different
understandings of what is meant by use of the term "endpoint" or "stack" (or
whatever). So my sense is that although we don't necessarily have an
agreed-upon precise definition of what one is (and I don't have the energy to
explain now why "a fate-sharing region" is the way to go - there is a very
powerful argument as to why but I'm too groggy to reproduce it right now), we
do have a (roughly common) understanding of what it is were're talking about.

This is rather different from the case of the term "identifier", where
clearly there have been people with very different assumptions as to what
that term meant. So I would therefore suggest (based on prior bitter
experience with the word "address") that any documents we create use some
*new* term, and *not* "identifier", for whatever name we come up with that
identifies the end-things (without giving their location).

	Noel

PS: As to which name to prefer for end-things, I would suggest we not use
"stack", since that has been demonstrated to be a poential source of confusion
with "dual-stack" hosts.