13.01.2011 17:58, Julian Reschke wrote:
On 13.01.2011 16:51, Mykyta Yevstifeyev wrote:
...
That sounds like an editorial error to me.
"any ranges to be *reserved* for .... "Unassigned"..."
doesn't make any sense at all. They are not reserved.
Yes, that is a type of error, but the meaning is that unassigned and
reserved values MUST (yes, must, that is in RFC 5226; see citation
below) be mentioned.
I do not see a citation "below".
I meant in the previous message.
This should probably be raised as erratum.
So the document specifying the regsitry MUST mention what are
Unassigned. Moreover, IMO, it would be useful to assign one value for
Experimentation.
No. should != must.
See below.
Even further? :-)
The same.
There are tons of registries where this is not the case; namely all or
most of those where the values are strings, not numbers.
The strings registries are rather exceptions from the rule I cited
above.
Well, we have many of them. The rules should takes those into account.
That, IMO, was the mistake of authors of RFC 5226 that didn't take the
text registries into considerations. We have no way to correct that now.
Mykyta
Best regards, Julian
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf