Planning for the Long Game

Our group often gets a case of gamer Attention Deficit Disorder after about six months.   I would really like to try going for the proverbial "Long Game".   I've heard people talk about campaigns that have lasted for years on end, and would be interested to see that happen for us.  The closest we came was the Caedes campaigns, where we played in the same game world from April 2000, until May of 2003.  That consisted of 3 different "campaigns"  (1 Basic/Expert D&D, and 2 3.0 D&D campaigns). 

As I've said before, D&D usually has lasted 20-30 sessions for us, while other games are 5-15 sessions more or less.  So, chosing ACKS for a campaign is a good first step, it's a D&D game, like 3.0 and AD&D and Castles and Crusades, so we're on the right path.   I'm just wondering what we can do to see if maybe we can make, say, 50 or 60 sessions for the campaign?

A few ideas:
1)  Frequent Chances to change characters.    On the one hand, I do not want to have mutliple characters being played in the same session by the same player, that's bad.  But on the other, I'm sure I'm with everyone in wanting to try several different types of characters.   Also, changing characters will help keep things fresh for everyone.  There are pitfalls, however and we need to take precautions about them.

We need to come up with a scheme or framework whereby we can have mutliple characters in operation without having the Death-Lump effect we had in Traveller and Castles and Crusades.  I was loose with running mutliple characters in both games, and it worked out contrary to my preferences.

2) The Smirkenburg Dungeon:   I've made several old-school multi-level dungeons in the past.  I've found that players seem to have a reluctance to actually go in them.   Part of the reason, I finally realized, was that I put them too far from home base.   The way the original Greyhawk dungeon, and the Castle Blackmoor dungeon worked, was that they were less than a mile or so from the main town, and even had entrances in the town itself.  You did not need to do any wilderness trecking at all to visit the dungeon.  These dungeons were meant for low to mid level characters, while higher level characters were more focused on the wilderness.

What I have in mind, is coming up with a vast, multi-level dungeon, right at either Portchester, or maybe the other main city on the map, Orchester (which we've rarely or never actually visited).  But, that dungeon is designed to be run totally without a GM.   Any session we wanted we could just go into the dungeon and look for treasure.  We'd have to come up with some serious ground rules, but it would take pressure off the players and GM, excuse me, Judge, especially in the dark winter months.  It may be a case that we create the detailed map of the place at the moment of exploring it.  If it's cleared out, the map will be gradually repopluated. 

3)  Spending for Experience:   I know that the spending for experience rules we used did a lot to create tanglible effects by the PC's on the "map" as it were.   ACKS assumes you are going to use your money for adventure purposes (especially wizards who need a ton of it).  But, does allow for "trivial spending" to be used for an XP "pool" for later or replacement characters.  Maybe we should institute a "tax" where characters must spend 10% of their loot on things like awesome clothes, parties, statues etc., and have the equivalent XP go into a pool for replacement characters.