Charles Lindsey wrote:
In <3E22BC35(_dot_)6020907(_at_)Sonietta(_dot_)blilly(_dot_)com> Bruce Lilly
<blilly(_at_)erols(_dot_)com> writes:
Charles Lindsey wrote:
No, I am talking primarily about gateways *from* news.
Yes, from news and to ... what? If not "to email", as I said, then why
are you discussing it here? If in fact we *are* still talking about
news->mail gateways, then please answer the question.
Gateways *from* news are a matter to be defined by Usefor. If those
gateways are *to* mail, then Usefor has to ensure they generate compliant
mail, or else make it clear that they do not. We are discussing it here to
see whether or not Usefor has so ensured, and in particular to look at
some awkward cases to see how they are best resolved.
You have once again evaded the question at hand, which is:
> Therefore, it is reasonable to proceed on the basis that existing gateways
> will either drop non-conforming input on the floor, or they will pass it
> on in a form that will be obviously "wrong" (i.e. will appear as
> gobbledegook) wherever it is viewed.
>
> The question is whether that situation is acceptable in a transitional
> phase,
You're talking about gateways to email -- what "transitional phase" -- what
are you assuming will change in the email protocols?
In addition, it has also been pointed out (by several people) that handing
illegal input (in this case non-ASCII header field content) to existing
implementations where that is illegal is neither conforming nor backwards
compatible. That is not a mere matter of *appearance*; is is a protocol
violation.
Perhaps it doesn't really matter that you are unwilling to answer the question;
your assertion that "it is reasonable" for "existing gateways" to "pass on"
"obviously 'wrong'" content in violation of protocols is incorrect. It is
neither reasonable nor acceptable (full stop).