Re: [sidr] [Errata Held for Document Update] RFC7115 (4973)

[Randy Bush <[e-mail]>]

> I am marking this report as "Held for Document Update" [1], which
> means that the author might consider its merits for a future update.
> If the use of the "666" octet was intentional, then a short note
> explaining might be appropriate to avoid further confusion.

problem is it's not a short note.

    In the current internet routing ecology, a /24 is the longest prefix
    which has a good chance at global propagation.  The prefix allocated
    by RFC 5737, with much verbosity, are three non-contiguous /24s.
    The result is that documents which want to discuss routable prefixes
    which involve subnetting can not use space from RFC 5737.  This is a
    general problem which someone (else) should fix.

    In RFC 7115, the subject of this erratum, routeable and subnettable
    examples were needed.  So the well-known private network 10/8, see
    RFC 1918, was used in the examples.  To indicate that it was not
    really intended to be used, an impossible octet, 666, was chosen.

this was explained a number of times as this rfc went through the
original sausage factory.

randy