RE: Proposal for new last call issue: Some unprocessed headers should stay

Hi Graham,

While not unsympathetic to that point of view, I think all the scenarios
imply is that there are certain features for which the contract between
the sender and the receiving node cannot be extended downstream unless
the feature is understood by the relaying node extending the contract.
How this may be enforced is, I think, an orthogonal question so to
speak.

Henrik 

>I think this argues for treating role identification and header 
>removal/propagation as orthogonal matters.

Received on Thursday, 17 October 2002 14:08:54 UTC