Re: Re-revised idea for early cross-area review

["Steven M. Bellovin" <smb at cs.columbia.edu>]

In message <1AA39B75171A7144A73216AED1D7478D01869C35 at esebe100.NOE.Nokia.com>, j
ohn.loughney at nokia.com writes:

>
>Question - would this be for all documents, or (as I would suggest)
>non-Informational documents.
>

To me, it's most important for architecture documents or whatever other 
documents set up basic data flow.  I've seen WGs where the documents I 
disliked most were the detailed protocol specs, but that's rare.  To 
give a current example, if a document came up that used MD5 instead of 
SHA-256, it's an easy fix.  But it's much more of a problem if there's 
no data flow that allows negotiation of the hash function, since there 
will almost certainly be a new one in 4-5 years.  Adding a new data 
flow at or after Last Call is painful and will meet with a lot of 
resistance.

		--Steven M. Bellovin, http://www.cs.columbia.edu/~smb