At 2:23 PM -0800 3/20/02, Steve Dorner wrote:
If we didn't get many mailers to implement it before, changing it
makes me afraid we'll get even fewer.
That's a risk, and an argument in favor of plan (A) [amending the
spec to give advice on when to discard the space as well as the CRLF,
subject to the characters on either side of the SP CRLF being 7-bit
US-ASCII and the folding rules for the language/charset being
displayed.
However, plan (B) [deprecating f=f in favor of f=x] is much simple to
implement, and thus may be more likely to be supported over time.
and I think the software that supports it will support a new f=x.
Some will, some won't. And others will implement f=x only. And
those who want to use f={f,x} won't know what to send.
Any new or updated software should support f=x only, that's the point
of deprecating f=f; to make life simple. Support f=x on sending and
receiving. Support f=f on receiving is nice, but not required for
software that doesn't already support it. Everyone knows what to
send.