Re: mech-v2: processing of non-ipv6 packets [Re: mech-v2-05pre]

[Brian E Carpenter <brc@zurich.ibm.com>]

Fred Templin wrote:
> JINMEI Tatuya <jinmei@isl.rdc.toshiba.co.jp> wrote:
>
> > I think option 3 makes most sense.  I can live with option 1, but it
> > seems just redundant as Vlad pointed out.  Option 2 seems odd to me
> > because proto-41 should be specific to IPv6 and there seems no
> > reasonable reason to mention other versions than 6 in this context.
>
>
> If we can't reach agreement on new text with "MUSTs" then we should
> revert to saying nothing at all and trust that robust implementations will
> do the right thing w/o gratuitous specification, which I believe is also
> reinforced by Brian's citations:
>
> Brian E Carpenter <brc@zurich.ibm.com> wrote:
>
> > RFC 1958 points 3.5 and 3.9 (and even 3.10) strongly suggest silent
> > discard in such a case, i.e. solution 1.
>
>  
> I agree that these points strongly suggest the silent discard, but wouldn't
> that indicate solution 3? 

Well, I think it's a matter of taste, really. Philosophically, silent
discard is supposed to be the Internet default, so you can argue that
it isn't necessary to state it everywhere. I could live with solution 1
or 3.

   Brian