Re: [Softwires] map-t to proposed standard rather than experimental

[Rémi Després <[e-mail]>]

17 nov. 2014 14:45, Ted Lemon <[e-mail]> :

> On Nov 17, 2014, at 3:25 AM, Rémi Després <[e-mail]> wrote:
>> If this is true (which isn’t clear to me at all ) wouldn’t RFC4821 deprecation be the right action?
>> Without that, considering here, implicitly, that RFC4821 is negligible would be too confusing.
>> 
>> An alternative, to take your point, could be to add "in view of doubts expressed about RFC4821 practicability"  after "negligible".
> 
> If you want to start the process of deprecating RFC 4821, we could certainly do that.  I don't think it's a requirement for this, though.   We have lots of disused RFCs that have never been deprecated.

1.
- I cited deprecation as the right action only if it is true that " no one expects RFC 4821 discovery to work anyway", and made clear my serious doubt that it it is indeed true).
- BTW, keeping the quoted sentence in your comment would have been appreciated: it was what gave sense to what I said. 
- To be more precise, the reason why RFC4821 does makes sense to me, is that, for hosts that send numerous large UDP datagrams (e.g. for some networked games), it is the only ICMP-independent solution (i.e. practical solution) to avoid excessive network fragmentations. 

2.
I agree that "disused RFCs" are in general not worth spending energy on them.
But a standard-track RFC "no one expects to work"  would be different (IMHO).
It would raise a problem of IETF credibility .

Regards,
RD