Gen-ART LC Review of draft-ietf-tcpm-urgent-data-06

["Richard L. Barnes" <[e-mail]>]

I am the assigned Gen-ART reviewer for this draft. For background on  
Gen-ART, please see the FAQ at
<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.

Please resolve these comments along with any other Last Call comments  
you may receive.

Document: draft-ietf-tcpm-urgent-data-06
Reviewer: Richard Barnes
Review Date: 2010-09-14
IETF LC End Date: 2010-09-21

Summary: This draft repeals RFC 1122's update to the semantics of the  
"urgent pointer" field in TCP, given that all known implementations  
follow the original (RFC 793) semantics anyway. The draft also  
generally recommends against new usage of the urgent pointer, because  
of a lack of interoperability and interference by middleboxes.

The only issue I have identified is some confusion about references/ 
updates.  The document claims to update three documents:
-- RFC 793, the original TCP specification
-- RFC 1011, which contains a revision to the urgent pointer in a  
comment
-- RFC 1122, which codifies the revision suggested in 1011
The update that it makes essentially reverts the semantics of the  
urgent pointer to that specified in RFC 793, so it's not clear that  
the document as it stands needs to update RFC 793.  (It seems to me  
that the normative reference to RFC 1011 can safely be downgraded to  
an informative reference, since the actual normative change to TCP was  
only made in RFC 1122.)

However, the original reason for the changes in RFC 1011 and RFC 1122  
was that RFC 793 contains an algorithm that suggests the semantic that  
those documents specify for the urgent pointer field.  It seems like  
if this document is returning to the original specified semantic of  
RFC 793, it should correct that algorithm as well.

Other than that, this document should be ready to proceed.