On Mon, 2015-03-16 at 14:07 +0000, Pete Biggs wrote:
Not that it particularly matters, my server-side filters already filter the Evolution list on To: and Cc: headers because I couldn't rely on the list headers. I have a choice you see - either I get things in my Inbox, which is full enough as it is, or I get non-list originated mail in my Evolution folder. Neither of which are ideal.
Well, if your preference is to only receive *one* copy (which I accept is a valid choice even if it's not *my* choice), then surely you have to choose one or the other? Sure, *I* agree that neither is ideal, but isn't that what you *wanted*? Ok, we discuss the "primary downside" of the missing RFC2639/RFC2919 List-* headers below, but other than that, what's the problem with directly-received list messages getting put into the list folder? Is it that you're concerned about the case where someone deliberately adds you to Cc as well as sending to the list... as distinct somehow from someone replying to you *and* the list, which you already said you want to land in the list folder only?
But the primary downside, which you have omitted, is that the received message does not have any of the list headers, because it didn't go via the list. So you can't "Reply To List" at all
OK, so you need a few extra keystrokes if you want to ignore my heartfelt pleas and still continue to deliberately cut people out of the discussion. I do concede I had omitted describing that as a 'downside'. :) Of course, you could add the headers yourself in your filter if you *really* wanted to, couldn't you? More sensibly perhaps, we already have per-folder options for replies. We have the 'Send Account Override' which controls which identity to use for replies within a specific folder. It shouldn't be that hard to add a *destination* override for the 'Group Reply' action in each folder too.
and you break everything to do with list handling.
You mean the 'Message' -> 'Mailing List' -> 'Unsubscribe' and similar options? Do you actually *use* those very often? They're cute, but they don't seem to be *common* actions. And again, if you *really* wanted to you could probably add the relevant headers at your own end. But still, the overall picture hasn't changed much. When making the choice of how to reply, we are comparing the fact that some people might be entirely cut out of the discussion if we reply-to-list, with what is still a fairly minor inconvenience for you if we reply-to-all.
My experience of the mailing lists I use is that cross-list posting is vanishingly small as is CC'ing external "experts". YMMV but that argument cuts very little ice with me.
My mileage definitely does vary. I find myself added to Cc on quite a lot of different discussions, on mailing lists I'm either not subscribed to, or which I *am* subscribed to but almost never look at their folders. On those lists which I rarely check, it happens quite often that when I *do* look into them, someone has *replied* to one of my messages and was asking for help, but didn't get it because they didn't actually send their message to *me*. Sometimes I take pity on them and follow up. Other times I just leave their message unanswered. As for list cross-posting, I took a look at a Linux kernel related mailing list which I run. Of the 2662 messages in there since I last archived my own mail store, 1633 were also copied to a list @vger.kernel.org. That's more than half. And many of the lists I'm on have a lot of cross-posting. It does also happen on *this* list, although obviously less than 50% of the time. Sometimes between this list and the evolution-hackers list, and other times between other GNOME lists. For example, https://wiki.gnome.org/MaintainersCorner says that release-team, gnome-doc-list *and* gnome-i18n must be be notified of new stable branches, and those messages are usually cross-posted. Likewise, string breaks in a stable release also IIRC require a message being sent to multiple lists. It *isn't* as uncommon as you make out, *even* here. It's kind of pointless to argue about the relatively frequencies of the different types of users, and the prevalence of cross-posting and adding people to Cc. Each person will have different experiences. All we can agree on is that they *do* all exist, and it *does* all happen. And even if it's considered rare, it's often the case that when people or other lists are added to Cc, that's done because they are the *most* useful people, and most likely to be able to solve the problem. They were added for a *reason*. We have to compare the down-side of cutting those people out of the discussion, with the down-side of a trivially minor annoyance to some other people. Even if there are *significantly* more people in the latter category, it still doesn't necessarily change the outcome. -- dwmw2
Attachment:
smime.p7s
Description: S/MIME cryptographic signature