In the first part of the discussion I outlined what I believed to be the central areas where MMO's needed to advance in order to move beyond the first generation of games. These areas were the world, character development, and story interaction. The worlds must advance beyond the static, playground like, worlds we are accustomed to and offer more flexible, dynamic regions for play. Character development needs to create an experience that is rewarding without encouraging tedious, repetitious, activities and a system for advancement that is not arbitrarily stunted by artificial level caps. Players also need to be able to feel that they have an impact on the world, that their contribution to the story has meaning within the context of the game's fiction and they need to be given choices that have real consequences both for themselves and for the world in which they live. These were the core concepts that a second generation MMO would be built upon. In addition to these there are other aspects of the genre which could be improved upon which would make game worlds more immersive and enjoyable.
Creature placement and AI:
In all traditional worlds computer controlled opponents are placed throughout the world at static locations called spawn points. Creatures appear at their specified spawn point and generally stand around waiting to be killed. Periodically the server refreshes the spawn points replacing anything that has been killed with another identical copy. A player could be standing directly on a spawn point when the refresh happens and the system will blithely recreate the creature right on top of them. It doesn't matter if there is one person in the area or fifty, how many creatures have been killed or how many remain. When the timer runs out first generation worlds follow the same mechanical procedure with brutal predictability. The greatest advances in the way worlds are populated amounts to random variations in the refresh timer or minor alterations to the spawn list. Current respawn systems are still being built on primitive algorithms the equivalent in depth and complexity to grade school arithmetic.
The system is so crude it is nearly impossible not to improve upon it but developers never seem to make much effort. Some few have made minor improvements but those are overlooked within the industry. City of Heroes is the only game I am aware of where npc's will not be spawned in plain sight of the player while a few other games have made meager efforts to disguise or dress up the respawn that amount to little more than cheap theatrics. If a game wanted to set itself apart from the crowd the easiest area to improve upon would be this one. Instead of hard wiring the game's creature populations some care should be taken to allow for a more natural method of populating the world.
The first step would be to remove respawns from the player's view. This precludes the immersion breaking event of having things appear out of thin air directly in front of the player, particularly things that the player has just killed and may be looting even as the respawn occurs.
The second step would be to remove all named or 'unique' creatures from the game world. Little is more absurd than having players lining up to kill 'The One and Only Gutripper' watching him die over and over again as they wait their turn. At this point immersion is not even a remote possibility, players become frustrated, and the game is reduced to poor parody. By all means continue to populate worlds with rare and interesting creatures but kindly do not christen them with flowery names that make it appear as if each is a unique individual unless you can well justify that being's never ending cycle of reincarnation.
The third step is to discard the blind, time based refresh. Instead pay attention to critical factors such as the current number of players in a region , the number of creatures available for play, and the ratio between them. If there are fifty people in an area supposed to be infested with orcs you will need to create an appropriate number of opponents to sustain that illusion. If there is just one player then spawning 200 orcs may be taking things just a tad too far. Similarly if there are a group of players in an area there is no reason to respawn all the opponents if there are still many enemies remaining. If the play area is a private instance there can be little justification for respawning dead foes at all. Having new opponents appear, without explanation, in a closed area that has already been cleared is an immersion breaking event that disrupts play and should be eliminated.
The final step requires an improvement in monster AI. Creatures should not stand idle waiting for adventurers to appear and slit their throats. They should have agendas of their own and go about the business of fulfilling their goals. They should be clever enough to establish lairs or camps and defend them, to send out patrols to search for intruders or gather resources. Creatures should forage for food, build nests, hide from danger, socialize and otherwise conduct themselves in ways that seem natural to the players. This does not mean creating a realistic simulation of a working ecology. It only means providing enough variation in creature behavior patterns that they appear to be living things instead of just a passive pile of pixels.
In addition to those improvements creatures need to have a fairly realistic range of perception. One of the great flaws in current games is that not only are creatures dumber than a rock but they appear to be deaf and blind as well. If a player can see an enemy then that enemy should be able to perceive the player. I don't mean that every creature has to aggro on the player from extreme distances only that if their buddy is getting killed ten feet in front of them they wont just stand around waiting for their turn to be murdered.
Making creatures behave in a somewhat realistic manner must extend to combat as well. Developers too often treat every enemy as a generic opponent within the rules of conflict. Many times I have seen snakes parry or enemies without shields block attacks. If it is decided that a particular creature should have a particular amount of avoidance don't start giving them abilities that make no sense in order to achieve it. Place attributes in ways that make intuitive sense so that the experience feels natural and players aren't cursing/laughing at you rather than playing your game.
The way in which games distribute new items to players after they kill an opponent has hardly changed at all in the past decade. The player defeats the enemy then checks it for loot. This loot is randomly selected from some table in a database that dictates what items may be dropped and the chance for each to appear. This same system was used way back when Dungeons and Dragons was published more than 30 years ago. It was a fine system for allowing a single human being to quickly fill monster hoards with piles of loot but in the MMO world it has become a crutch used by developers who cant be bothered to consider something better.
The first failure of the purely random system is that rarely do developers give any thought to elements of realism. The archetypal example is that of the wasp who drops a greatsword. The player is left to wonder 1) Where was he hiding that? and 2) How did he move with that strapped to his ass?
Another common example: after defeating a supposedly intelligent opponent wielding the crudest of clubs the player discovers that it had in its possession a +5 broadsword of slaying yet chose to defend itself with a broken table leg. The final example of bad loot is when players are given the task of collecting some particular body part from a creature. It might be spider eyes, wolf tails, or cyclops brains, whatever the item, as soon as the player starts searching for one it becomes apparent that the majority of creatures they encounter are either mutants or have been horrifically maimed as they do not posses the sought after organ no matter how vital that organ may be to their ability to function as a living being.
Random loot is frequently a good thing. Players greatly enjoy pulling the virtual arm on the slot machine and seeing what pours out. That does not mean realism must be sacrificed in order to accommodate the thrill. In general players detest the collect x organ quests because they are all too familiar with the lie that is implied. When the laws of reality are broken in order to make the player spend more time at the task they feel as if the universe is cheating them and they are correct. If you want players to spend a certain amount of time pursuing the task don't ask them to collect 10 items that have a 1 in 10 chance of dropping. Just be honest up front and ask them to collect 100 items that have a 100% chance of dropping. Better yet, don't create this type of quest at all. Players immediately recognize it for the pointless time sink that it is and resent you for asking them to do it.
Although players enjoy not knowing exactly what they will get from an encounter one thing about random loot that they hate is when they make a tremendous effort to defeat an encounter only to be rewarded with items that are of no use to anyone who participated in the fight. In one group I played with this was known as Felnarion Syndrome.
In World of Warcraft the first raid instance was Molten Core and in Molten Core there were 8 types of armor that could be found. Among these were the warlock and druid armors called Felheart and Cenarion respectively. These armors could not be used by any other class and those classes were among the least played. Random determination of loot from the encounters in Molten Core dictated that, despite the relatively few characters of those classes in a raid, an equal amount of that armor compared to the others would be found. This was further exacerbated by an apparent imbalance in the loot tables which dictated that those types of armor should appear more frequently than others. It then became common for moans and groans to ring out when, after killing a boss, the resulting loot was 'Felnarion'. I'm sure the few warlocks and druids did not mind overmuch at first but even they became agitated by the overwhelming amount of useless loot we obtained over time. The results of our efforts often seemed more like a bad joke than our just rewards.
One way game makers have made half hearted efforts to alleviate this problem is to replace items with tokens. Each token can be exchanged by several character types for a corresponding item or selection of items. This doesn't truly solve the problem however. By providing tokens instead of items the amount of time it takes for probabilities to yield unwanted loot is delayed but not eliminated. Sooner or later, unless all characters can use the same token, some tokens will be in high demand and others will be found far too often. Further, no developer I know of has ever totally replaced all loot with tokens in any encounter in their game. Nor would it be desirable to do so. If all items were replaced with a single token the loot is no longer random and players know exactly what they will get before they win. This spoils the surprise factor much like knowing what gifts you are getting before they are opened.
What then should developers do? The obvious answer would be to continue using randomly generated loot but filter it so that the items distributed to players are actually needed. If the loot generator says a shaman helm should be dropped but there are no shaman in the group then generate something else. If it says a mage robe should be given out but all the mages in the group already have it then select a different item. This filter would not need to be used in all cases, only those where the generated loot was not tradeable to other players in the game. Of course if a game allows all items to be freely traded then this is of far less importance. If loot distribution is consistent with the distribution of classes players can work out among themselves an appropriate system for getting items into the hands of people who need them.
When generating loot a well designed game will ensure that the items dropped are appropriate for the type of creature that was killed, consistent with the laws of nature, physics and common sense, commensurate with the time and effort it took to win them, and most of all, desirable to the players who earned them.
The worlds of the MMO genre become more visually stunning with each new release. Some go to great lengths to provide realistic landscapes while others create more stylized, colorful environments. Whichever route is taken developers are content to leave things looking pretty but being functionally crippled. In single player games it is common that if the player moves through the water a wake is created and ripples move across the surface. In an MMO you are lucky if players even swim when crossing a body of water much less disturb the surface in any way. Many bits of scenery in an MMO can be moved through such as bushes, cobwebs, or tall grass. In single player games players have come to expect these objects to move and rustle as the they pass but in an MMO there is never any reaction at all.
Another common problem is obstacles. Long ago the single player RPG began allowing players to climb over things, pull themselves up onto ledges, and hop over low hurdles. In the MMO world players are left with jumping. If an object is taller than a player can jump it becomes an impassable barrier even if it is only waist high. Even the implementation of jumping is usually quite crude giving players little or no control over height or distance.
MMO's have enormous opportunity for improvement when implementing their environments. They need to begin implementing many of the features commonly found in single player games to really bring their worlds to life. One of my most disappointing moments occurred while playing Lord of the Rings Online. I was riding along the road through the Trollshaws when I came to the famous Ford of Bruinen where Frodo escaped the Nazgul. I rode on toward the river with visions of spray being thrown up as I splashed my way across in the footsteps of the famous figures that had passed this way before me. Instead of churning waters and thundering hooves I was left with the familiar quiet clip-clop of riding across any other type of terrain without even the slightest disturbance to the water below.
Many situations arise in games worthy of exceptional effects of both sight and sound yet they generally go unfulfilled. Events which could be thrilling and memorable are left as flat as watching an action movie in black and white with the sound turned off. People respond strongly to the way the world behaves. Even simple actions could resonate with players if only more attention were paid to adding that final bit of polish that rounds out the experience.
The ultimate achievement would be to create malleable environments that would be marred when violence was done to them and recover slowly over time. If a pitched battle were fought in a field clods of dirt may be thrown up by the passage of men on horseback, great gouges cut from the sod and overturned stones left stained with blood. Trees might be set ablaze by magic and fires spread through the forest. Weeks after the conflict the land might still be scarred by the events that took place reminding players of what happened there conveying a sense of history and depth.
I am not going to quibble over implementation details or bemoan potential technical limitations. It is enough to say that these things are desirable and will add great value to the game that overcomes the obstacles and integrates them into their world. There are those who will say a thing is too costly or too difficult and make no effort to achieve it and there are those who see the value of a thing and strive to accomplish it. It is the latter variety who will succeed while the others struggle to keep up.