Games get old and fade away.
This is current thinking and practice, but I don't think it need be the reality. Starcraft, Diablo 2 and CS should be enough proof that folks will play the same game for a long period of time after the dev's have moved on, if the price is right. Course, these games are free. I think that games like DAOC shoud be asking are, "At what cash-point does it become worthwhile to a player to keep an account open to play occasionally?" And also, "At what cash per account point does it pay for admin services, server fees, etc"? Would I play $5 a month to keep an account active? I might if they gave me a deal on a full year (say $40), allowing me to occasionally pop in and out.
I might if the game had a seaon (say three months), a conclusion (after three months a winner was declared and starting conditions reset) and season-long objectives (a total-war game that focused on land aquisition and build-the-army type game play sounds cool).
I'm not going to apologize, but I am ashamed of the book that is to follow
Build the ArmyExample: season starts with everyone at level 1 and no intra-realm access, all pvp occurs in frontiers/battlegrounds. Character play time is capped at 3 hours per day for the next two weeks, which should be sufficient to get a char to level 50.
At this point total war ensues and guilds pick their keeps (guilds with highest realm point totals pick first until all keeps are picked, biggest guild will most likely choose the capital). All players in guilds (and the unguilded) that don't have a keep are unguilded and auto-enrolled into guilds that have keeps (alogrythym should try to equally-staff guilds with people, and equally staff guilds with realm points).
When a keep is taken, that guild is despawned, half it's players going to the enemy realm, half being absorbed into the defending realm (players respawn from the new capital) - thus building the army of the conquering realm. The keep now belongs to the conquering guild (which makes it more resilient to being conquered).
Each realm has a certain number of npc defenders, initially these defenders are spread thinly throughout the keeps, but as more and more keeps are taken, npcs are assigned to the reaming keeps, making the remaining keeps more difficult to take over time.
Long and rambling to followI think the evolution idea is decent if, at the same time, they balance the entire equation. The entire equation being cost per account, server population and developer/designer support. Figure out how many players per server are required to provide a decent RvR experience, what the anticpated player retention is, what those servers cost, what the desired profit base is, and set billing to that number. If the player base shrinks, consolidate servers.
Accept the fact that more classes results in more work, possibly without real benefit (as others here have mentioned). Decide how many roles an RvR game needs, and create enough classes to fill those roles with a certain amount of overlap. Specialist classes, in my opinion, are the devil. Specialist classes restult in groups that need specialist classes, and hybrids that are always either too good, or not good enough. Just make hybrids.
What roles does DAOC need? Lets just say melee damage dealer, ranged damage dealer, healing/buffing, shield tank, travel/auras and intelligence (radar and debuffing). If every class is a hybrid that encompesses two class roles, it's possible that a group of three could have every class role represented. Let's try to reduce the amount of class tweaking and say that each realm has 5 classes. A group of 8 should have more then enough of everyone to provide a decent mix of roles.
I didn't include stealth in the required class roles because it's stupid in games like this. If DAOC had supply lines, or things to steal, or assasination, it might make sense... it doesn't, it's about group versus group warfare, please stop the madness.
DAOC currently doesn't have an intelligence class, but radar software is readily available and already fulfills this role. Lets just add it as an ability to the debuffing class role so we can stop the talk about who is and who isn't a radar-using cheaty whore. Additionally this class will figure out a defender's posture (melee/spell), and the type of auras that a bard is running so that the group can make tactical choices on how to proceed.
Shield tank picks a group mate to defend, and sets a posture (melee or spell defensive), posture has a recharge of 5 mins. Make the defend ability beefy enough that attacking the tank first is a decent tactic (unless the tank is in the wrong posture).
Travel should be effected by terrain. Hibernians are good in forest, Albion in plains, Midgard in mountains. Roads should give a travel boost, especially to those with travel powers, possibly hard to code, but allows for interesting flanking tactics. Aura's should be twistable. Good bards in EQ were pretty impressive, no reason not to reward people for being good players
A realm ability system that doesn't create haves/have nots, but gives flavor/choices and access to ablilities that are dual edged or access to abilities that are strategic.
Flavor example - a class that is a melee damage/tank could get access to a travel ability (not as fast, and no group component at lowest investment).
Convenience - access to a portal system
Dual edged example - tank chooses an ability that increases his defense versus smashing, but this decreases his defense against slashing (he might make this choice if he frequently groups with an aura-class that increased a slashing defense song).
Strategic - crafting, or higher end crafting. Knowledge/ability with seige weapons, etc.