** Sometime around 15:53 -0800 11/30/01, Chuq Von Rospach sent everyone:
>On 11/30/01 2:41 PM, "Vince Sabio" <vince@vjs.org> wrote:
>
> > However, for
> > those brane-ded MUAs that cannot filter on headers other than To:,
> > From:, etc., List-ID: (alone) will not be useful; a quasi-invariant
> > "standard header" would have to be used.
>
>Double-however, how much time and energy should we (as list owners, admins,
>list server authors, etc) put into supporting non-conforming, brain-dead,
>broken mail clients? It's not a trivial question, given the two top clients
>(OE and AOL) both can be classified in this category. But should there be
>work done to workarounds that allow them to continue to be brain-dead, or
>should the pressure be to encourage users of those tools to pressure the
>developers to come into the 18th century with the rest of us?
I look at it as a trade space, with "user popularity" traded off
against "difficulty of supporting" [the MUA]. Or, perhaps more
correctly, the axes are "user popularity" and "ease of supporting."
If the score on either axis is sufficiently [1] high, I do my best to
support it. Hence, we use three invariant headers: List-ID:, To:, and
Reply-To: (yes, we munge Reply-To:, yes, we've covered that
discussion in detail, yes, it is a holy war, yes, I have my reasons
for munging Reply-To:, and no, I'm not willing to change <g>).
On all of my lists, I add in an "X-Message-To: <subscriber_address>"
header, to help the Listmoms (thanks, Chuq) handle those cases where
e-mail addresses may have changed.
And on my largest list, I also added in an Errors-To: header, for
those MTAs that still use it.
- Vince
[1] #define sufficiently we're not likely to receive complaints from users
References:
|
|
