[ietf-dkim] Re: Use of LWSP in ABNF -- consensus call

[Frank Ellermann]

Charles Lindsey wrote:

> we could follow the lead of RFC 2822 and remove the defiition of
> LWSP to a separate "obsolete" section, with the provision that
> its use MUST be accepted (by existing RFCs, which is at most the
> three above), but MUST NOT be generated (by new RFCs).

No obs-LWSP, please, it's already bad enough.  RFC 2822 was only a
"treat as immature" PS, 4234bis will be a STD, no new requirements.

And a MUST NOT is overkill for the LWSP.  My first attempt to get
rid of it in 4646bis promptly resulted in the same problem without
LWSP.  Authors need to check their syntax, mechanically replacing
LWSP by something else might be still wrong, and then it's also
not more obvious.

Frank