[BXXPwg] terminology wrt "protocol binding" and "BEEP channel profiles"?

bxxpwg@lists.invisibleworlds.com bxxpwg@lists.invisibleworlds.com
Thu, 17 May 2001 15:08:12 -0700

The question, then some background, and then a proposal..

My question is: 

  "how does one, from 'outside' the BEEP context, talk about specifying 
   the 'binding' of some protocol onto BEEP?"


The term "protocol binding" is sometimes used to refer to the specification of 
the layering of one protocol onto, or "over", another.

The term "profile" is used (outside of the BEEP context) in the sense of 
making use of a specification (e.g. X.509) in the context of some other spec 
(e.g. PKIX is a "profile" of X.509; c.f. RFC2459).

The BEEP protocol spec (RFC3080) uses the unadorned term "profile" to name a 
collection of definitions of syntax and semantics that specify a particular 
channel behavior.


A proposal is to say something like (where "FOOP" is a short for "FOO 
protocol" (which is imaginary))..

  "In order to specify a protocol binding for FOOP over BEEP, one must write 
   and register a 'BEEP channel profile for FOOP'."

Other ways to say it..

  "FOOP's BEEP channel profile"

  "The FOOP BEEP channel profile"

  "The FOOP-over-BEEP protocol binding is specified via the 
   FOOP BEEP channel profile."

  "The FOOP-over-BEEP protocol binding is specified via the BEEP 
   channel profile for FOOP."

A sub-proposal, implicit in the above, is that when speaking/writing from a 
perspective "outside" of (or "external to") BEEP, one should use "BEEP channel 
profile" to refer unambiguously to what's called simply a "profile" in the 
BEEP context (i.e. RFC3080).

So, does the above make sense?

For completeness, I'll note that the term used to describe the so-called 
"binding" of BEEP itself onto TCP (RFC3081) is "mapping". Nominally, it seems 
to me the above still holds and makes sense even if the word "binding" 
everywhere above were replaced with "mapping".