I’ve been using a new way to read my email for a few months now, and it seems to be working quite well, so I’m going to write down how it works, since I suspect others want something similar, and I could have benefited from someone describing this scheme to me.
Here’s the basic idea:

• At least one copy of each incoming email ends up in a non-inbox archive folder
• All important email goes to my inbox (but with a backup copy elsewhere)
• I delete emails out of my inbox when I’m “done” with them. Either I’ve finished the task that is related to a message, or I’ve replied, or I’ve just read it and decided that there’s nothing further for me to do.
• I never delete emails out of my archive folders
• I use indexed search for everything else

My main mail system at work is Outlook, so here’s how I implement this there:

• I have a few rules that filter out automated messages into their own folders (like commit messages from our VCS)
• I have rule that says “for a specific set of lists, put a copy of the message to the inbox-archive folder” and stop.
• I have another rule that says “for a different set of lists, move the message to a list-archive folder” and stop.
• Pretty much every rule has the “and stop” action so that the outlook will perform the action for the first match. The last rule is a catch-all, that says essentially make a copy of the message in my inbox-archive folder. I do this since if a rule hasn’t matched already, then the message is going to end up in my inbox, so I need to make a copy of it

That’s basically it. I have a similar scheme for gmail where I use archiving instead of deleting, but the idea is pretty much the same.
Things I like about this scheme:

• If it’s not in my inbox, I don’t have to worry about it
• No big complex set of rules I need to maintain
• I can perform the basic operations from all mail clients, since they can all at least see the inbox and delete messages. (except for gmail, sigh, where I can’t do the “archive” operation from the built-in mail client)

For further optimization, I set up search folders that search through my archive folders if I want to keep tabs on particular mailing lists (and this works even for crossposted threads, some of which might end up in my inbox archive vs my list archive). I only ever read lists like this from one machine so that’s fine.
It does require good search to exist. But with the gmail that’s there, and with later versions of outlook and Win7, I can get that for work mail as well.
I was originally very concerned about preserving conversation threads in my inbox. If I delete messages after I read them, but then I get another message that’s part of a thread for which I’ve already deleted, I was worried that I wouldn’t be able to recover the context.
In practice, searching for messages with the matching subject line quickly recovers the thread. Usually that’s good enough to build up any context, and it’s faster than picking through a large inbox anyways (since Outlook’s threading still doesn’t quite work in all cases).