At 05:40 PM 10/24/00, Chuq Von Rospach wrote:
>At 3:45 PM -0400 10/24/00, Jim Trigg wrote:
>>That sounds nice, except that 50-75% of mail readers in common use cannot
>>filter based on the Sender: header, let alone any particular
>>header. They can filter on: To, From, Subject, Reply-To, or "any header".
>so we shouldn't use any technology everyone can't use? We shouldn't put
>the hooks into things to encourage other client programmers to accept and
>you have to move things forward or they stagnate or die. Especially in a
>case like this, where if they don't want to take advantage of it, they
>don't have to. But should we not do this for the people who do have the
>capability because some folks don't?
>(in case it's not obvious, I don't but taht approach. I don't add gizmos
>to my systems because they exist, but I don't hold back just because some
>subset can't take advantage of them, either)
No, that's not what I'm saying. I have no problem with adding features
that allow more advanced users to use them while less advanced users
can't. I have a problem with changing things so that the less advanced
users are no longer able to do what they once could.
>>Which is exactly the problem with using "List-ID" to identify that it's
>>not personal email and should be sent to a separate sorting area.
>but this is a case where the MLMs have to take a lead in promoting the
>RFC. the clients aren't going to support it until the MLMs do (why should
>they?), the MLMs have to put it out there so the client programmers will
>know it makes sense to support it. Same with the list-* RFCs.
That's true. The problem is that you seem to be advocating making the new
feature the *only* way to identify the list that the message came by.