Easier said than done. MMO's have to be great and realistic games, but have to be the same game for each player. That means that the questgivers will stand in the same spot 24/7 and the quests will have to be text-based, and there are lots of queues.

I actually implemented my random idea (not to the full degree I wanted) as a NWN mod that ran as a persistant server (like a small MMO) and the players enjoyed the hell out of it. Nothing has to be the same. And nothing was the same.

I scripted everything up to cater to the individual player instead of what MMOs do, which is have set things stand there 24/7 and make the player cater to the preset content.

In a standard MMO there is a static broken merchant caravan at the same spot. The NPC will stand there 24/7 and repeat the same crap about the same guys that ran off to the same place over and over again. The enemies drop the same crap over and over, and there is one thing you can do with it.

In my system, there was no cart, there was no quest. I gave every player an array of 3 quest slots as part of their custom data, and I generated stuff on the fly to fill those empty slots.

So a player enters a zone. This zone has part of a road that goes from town a to town b. So it's a trade route. So after checking that the player has an empty quest spot, and he passes a random check to see if I should generate something,

I generate a merchant cart somewhere. Not on a specific pre set spot. Anywhere in the zone. In the grass, up a tree. Upside down in some rocks. Who cares? Just put it somewhere random. If the player stumbles upon it, then his imagination can fill in the blanks. The thieves hid it, they ran him off the road, or the merchant was drunk.

Maybe I generate a merchant there with randomly generated story that some (creature/guild/thieves) attacked him and ran off with (any random item).

Maybe I don't generate a merchant, and I just put some random object there that acts as a clue that (random creature) was there and ran off to (random place).

Maybe the merchant is fine and YOU decide to act as the bandit, creating new content for another player to find.

So now the player has a unique event to react to or ignore.

When the player enters the destination that has the (random creatures), I see this in the script that checks everyone who enters a zone, I generated the content for them.

If the player destroys the (creature) and gets any randomly generated associated loot. He can do what he likes with them. There are variables attached that say what they are, and where they came from. So the game always knows that player has the loot item from a bandit raid on the road (should he choose not to return it), and the game can react to it constantly. Maybe you don't want to return it because you're evil, it's useful to you, or you can get a better deal for it).

On rare occasions, that loot could have been an item someone else was looking for, causing new random quests between the factions that were interested in it.

If a another player stumbled upon that merchant that was waiting for the loot return and killed him. I'd generate an evidence item. So you'd have the loot to find a new owner for, and a quest item to find out who who murdered that guy (if you wanted to follow up on it). (I never coded up the part where I wanted this to happen only with nearby players, and to require the player to interact with them as witnesses, then the guilty player can have generated authorities looking for them).

I had quests to save NPCs from randomly generated places or villains. You could return them. Rob them. Beat them. Return them by demanding a ransom. Sell them to slave traders. Sacrifice them to some evil deity, etc... It was all up to the player.

What happens if a large chunk of players are sacrificing or selling into slavery? The game keeps track of this. So it can react by spawning appropriate content.

The cult of the evil deity is growing stronger. So spawn a higher percentage of followers. The town NPCs are spawned with pitchforks, clubs, and torches. They generate quests and jobs that are weighted towards cracking down on this cult or freeing slaves.

Those merchants you bump into aren't getting robbed now. They are afraid of the cult and need your protection! They are afraid to walk alone, or you can't let their deliveries get into the hands of that cult. Are you in that cult? Then you just found your next sacrifice alone in the woods.

Everything was random and flexible and apart of the same system. Everything was aware of the state of the world and could react to it and change it. Everything that happened could change the state of the content for the other players and it was player driven to begin with.

I hacked this up in my basement when I was 22 and knew half of what I know now. I'm sure a bigger, real MMO could try something similar. You just need a little data structure that says who/what/where/why/when and then to make everything aware of it. Even NPC chatter can be gossip about player deeds.

[quote name='Daaark' timestamp='1354095484' post='5004900']
An MMO is just a graphical MUD with a bigger playerbase.

Most early MMOs has their roots in MUDs (Meridian59, Camlelot, EQ ?), what is quite interesting is, that MUDs had often world builders (limited to certain players/admins), a way to alter the world dynamically. This was almost completely lost from the transition from MUD to MMORPG.
[/quote]

Thats a very good point. Not many seem to remember that most (if not all) MUDs had a staff (wizards, creators) that worked on new features or expanding the game world and that it was done dynamically, without strict release cycle. Things were done and deployed on the fly (depends on a MUD type, but most drivers/mudlibs allowed this). Somehow from a dynamically expanded game world that introduced unknown changes that had to be discovered by players we went into same-for-all experience with strict releases (expansions) that are consumed by most active players in 1/100 of time that took to create it. A pity really.

Where are we and when are we and who are we?
How many people in how many places at how many times?

Thats a very good point. Not many seem to remember that most (if not all) MUDs had a staff (wizards, creators) that worked on new features or expanding the game world and that it was done dynamically, without strict release cycle. Things were done and deployed on the fly (depends on a MUD type, but most drivers/mudlibs allowed this). Somehow from a dynamically expanded game world that introduced unknown changes that had to be discovered by players we went into same-for-all experience with strict releases (expansions) that are consumed by most active players in 1/100 of time that took to create it. A pity really.

That stuff never really went away. We had BBS doors that did that (LORD 2), and full multiplayer graphical RPGs. Those features just weren't advertised as hard. There was a big MMO sub community in Neverwinter Nights, It was like having a whole bunch of free MMOs to play on. Find whatever one suited your preferences, or start your own. Just the mainstream MMO games turned into those treadmills to nowhere.

There must be some Second Life communities that have thought up some good designs?

I disagree with all four points in the original post.

• The amount of content needs to be enough for each player's whole playing experience from a trial account to a cancelled subscription; as locations can be either instanced or shared by many players, content can be stretched too thin in time, but not in space.
  If there's enough content for a few players, there is enough content for many players, with two exceptions: player-owned land etc. might require a qualitatively larger world (e.g. adding railroads or highways or stargates to keep travel times in check) to accomodate a larger population, and occasionally some places could be too crowded to operate normally (but it's perfectly normal and expected).
  
• If simple procedural generation modifies content in superficial ways, it's always better than exact repetition and bad procedural generation.
  Good game rules generate fresh tactical and strategical challenges from small changes of scenery, monster power, etc. (questions like "is it going to kill me?" and "where can I run?" never cease to be interesting), and procedural generation could be taken further.
  
• Unless the designers introduce railroading in what should be an open-ended experience, telling the player things the character is supposed to know (e.g. the location and schedule of important events) or things that he might not have learned yet (e.g. in which market locations he can sell what loot) is just useful guidance.
  
• Socializing is the opposite of solipsistic private copies of the world. The only advantage of a MMO over a single player RPG is the interaction with other people, and MMO designs will (or should...) continue to sacrifice everything else to make friendly and unfriendly interactions with other players more fun.
  

One aspect that I think is being emphasized too much in this conversation is the need for one huge world for many players to explore.

An MMO can still have a feeling of open-world without having a massive over-map that takes hours to traverse. From my experience (I don't play MMO's, but had friends who were very addicted to them) players don't like spending half their time going from one point to another. So from that perspective, lets say we get rid of that.

Nearly all MMO's have different servers and some form of level streaming to help maintain the amount of content thats viewed by players at any one time. If you break it down into a smaller level, it could be easier to maintain; not only in regards to content but also cost to the developer. So then every level that is played could be maintained content. Again, if we take that approach, we can get rid of over-populated quests.

If you create something like a hub in which players interact before forming teams and tackling quests, then you are basically creating a more stream-lined instances system like in WoW. If you create faster, more intense action in the game and take the grind away for leveling, leaving only the instances with smaller teams, you create an MMO system that takes less of a toll on servers and content creation, thus minimizing cost.

Then again, this method of instanced action stemming from a central hub could also be used to explain nearly all FPS multiplayer games.

I guess it's all in the way you perceive games. I see Call of Duty as an MMO just as much as World of Warcraft since they both have addictive game-play with leveling and as you progress, you get better gear. And it all really boils down to playing games with friends.

So yeah...pretty much if a game; any game, is online, has a decently large community, and some form of matchmaking where you could play with strangers, than isn't it an MMO?

Lord DarkShayde, MMO is a buzzword. MMO doesn't make a difference when you only ever meet a few players at a time in any given session, and it's all instanced anyways. If your party enters a dungeon and gets it's own private instance, what difference does the MMO part make? You might as well be playing a something like Icewind Dale at a Lan Party.

What difference does it make if there are 64 or 6400 if I never see these guys? They are in other zones, doing other quests, and when I come across them, they are just passing by on the road on their to or from town. They extra players really don't add anything to the experience.

It's like an amusement part. They are just other people waiting in line to ride the repeating quests. The level reqs are just a replacement for the 'you must be THIS tall' signs. ;)

I'm silently reading all the great contributions to this post, thanks guys!

And now I found this recent article. Quote: "Production Problems Made it Tough to Find the Fun" ... "We really underestimated how long things would take to get right, and how much would have to be invested in... the core loops and making it fun."