RE: [newtrk] where to state that something is required (Comment: PIm Sparse Mode to Proposed (fwd))

[Larry Masinter <[e-mail]>]

> On Fri, 28 Oct 2005, Harald Tveit Alvestrand wrote:
> > I think Sam's note to the IETF list today pretty clearly
> > indicates the problem we have with saying "this is the mandatory
> > to implement mechanism for this protocol in this context", given
> > that neither the protocol or the mechanism should need to be
> > revised in order to make that statement.
> 

draft-ietf-newtrk-interop-reports-00.txt
section 2, 4th bullet: 

  o  A feature description should indicate whether implementation is
      REQUIRED or optional for the protocol.


So the protocol feature set could say, for each feature, whether
it is mandatory to implement.

5th bullet

   o  Protocols may define multiple roles (e.g., client/server/proxy).
      Protocol Feature Set can include sets of roles, and feature
      specifications can identify the roles for which the feature is
      appropriate.

I think the 'mandatory to implement' status should be per-role,
so that "in this context" would also be covered.

Larry
-- 
http://larry.masinter.net

.
newtrk resources:_____________________________________________________
web user interface: http://darkwing.uoregon.edu/~llynch/newtrk.html
mhonarc archive: http://darkwing.uoregon.edu/~llynch/newtrk/index.html