[mail-vet-discuss] tightening the ABNF

[Tony Hansen]

I disagree with this suggestion. I don't think it's necessary to move 
all semantics down into the syntactic level. The lexical layer doesn't 
need to be overloaded to this degree.

Just my two bits; YMMV.

	Tony Hansen
	tony at att.com

Michael Thomas wrote:
> I'd like to suggest that the current ABNF doesn't do a very good job at
> limiting silly states, in particular about the relationship of propspecs
> to methodspecs. What exactly does dkim=hardfail smtp.mailfrom mean? I'm
> pretty sure that it's meaningless, but neither does the ABNF limit this,
> nor does the current draft even discuss why you'd choose one of the
> ptypes at all.