[ih] Why a new etype of IPv6

[jnc at mercury.lcs.mit.edu (Noel Chiappa)]

    > From: John Kristoff <jtk at depaul.edu>

    > I don't recall seeing any of the official, or even unofficial for that
    > matter, reasoning for assigning a unique ethertype value .. for the use
    > of IPv6 over Ethernet / 802.3.
    > ...
    > Maybe some specific performance or data link-specific reasoning? 

I have this vague memory that there were some boxes that weren't checking the
'IP header version' in the IPv4 header, and IPv6 packets were confusing them.
But don't put too much weight on that memory, like I said it's vague.

	Noel