uploaded rev27 of multihop draft

Just uploaded 
rev27 of multihop draft, based on comments from Sander on Feb 
11:
 
- in 1.6.4: 
distinction Available routing input / Effective routing input , and related 
recommendation in 1.9 on managing RM sequences. ()see use of 
PMode.reliability.correlation)
 
- in 1.6.4: about 
the case where only the @mpc value is used for routing: requirement to not use 
this mpc in different endpoints.
 
- in 1.4.5: 
recommendation on how to interpret "reinsert" in the soap intermediary 
model.
 
- in 1.6.5: remove 
the requirments of which forwarding patterns MUST be supported by intermeidaries 
(TBD by conformance profiles)
 
- moved in 1.7.1.2 
the paragraph about wsa:Action that was in Appendix 
B:1.11.2.
 
 
-jacques