One of the things that separates a really well-designed role playing game from a hack and slash through a single corridor is the concept of choice. Players need to be able to have choices, and those choices have to matter. A player should be able to pick and choose a course of action for his character from as wide a variety of possibilities as is feasable, and while some of these choices will be obviously stupid ones, there should not be only one option for becoming "heroic." Unfortunately, the nature of modern CRPG design seems to mandate that players kill stuff and rob it, due to the relative ease of focussing on combat only as a path of advancement, as well as the "monsteritis" syndrome that relegates all non-player characters to the role of "thing that sits around waiting for players to attack it."
The most elementary system for expanding the number of options open to a player is meaningful craft skills. This means artisan trades that players can explore that exist for some reason besides equipping "real" characters who go out to kill stuff. In a world where food is required by PC’s and NPC’s alike, agronomy and foraging could be important skills, as could hunting game. Indeed, a nomadic character who stays away from town would need these abilities, even if he supplements his rations by murdering other players for their salt pork and waybread. If food is not required, other skills would certainly be valuable, like leatherworking, ore refining, smithy, woodwork, bowyery… the classics, as it were. If the engine is sophisticated enough to track the construction of new buildings over time, carpentry and architecture take on new possibilities. Cartography, dowsing, herbalism, medicine, tinsmithy… trade skills can number in the hundreds easily, limited only by the complexities of code and the ability of developers to think outside of the norm when considering trade skills.
Some of the most rewarding aspects of playing an RPG for some players lies in the less quantifiable pursuits like diplomacy, the acquisition of a title, political influence, and inter-community trade. These are more difficult to simulate in a system relying on hard code, as by and large these are subjective skills, not measurable in terms of points. However, one can always start somewhere. The acquisition of titles like "Grand Master of the Four Flowers School of Swordsmanship" can be done through quests, say to prove one’s worth in a contest of skill at the school, assuming one has spent enough time there to qualify for the test in the first place. This sort of contest is nice, because it doesn’t confer anything but a title and bragging rights, but only one person (presumably) can be Grand Master of any one school. This provides an avenue of competition amongst players that doesn’t involve PvP, which is nice for those not inclined toward human conflict. Acquiring the title of Ambassador may require several successful missions to neighboring city-states (although the heuristics determining a successful negotion would be rough indeed), and may confer on that player some extra status in his hometown that could translate into legal flexibility, or even better prices at the market. To gain more standing in the Merchant’s Guild might require successful caravans full of needed supplies to dangerous zones, and might confer similar price benefits and a certain amount of credit, plus economic flexibility between regions if your monentary system is realistically diversified. A Master of Lore, accredited by the not-so-local Wizarding College, would have demonstrated a high degree of aptitude in several areas of arcane knowledge, and would maybe gain access to some interesting (though not utterly powerful) incantations, probably of an informational nature, and better availability of ingredients, plus access to restricted tomes and such.
Even the overused motive of "kill stuff" can take on new meaning if the system is flexible enough to support it. If your goal is (using an Asheron’s Call cliche for example) to drive the Tumeroks out of Dryreach, wouldn’t it be interesting if it could actually be done? You’d think after losing about 18 million troops to marauding humans, they would have given up on the idea of holding the town and move somewhere else. The town is freed, the conquering players are hailed and honored, and now there’s something else to do as the evicted monsters make other plans, or call for backup. If the engine allows for the dynamic construction of buildings over time, a player could discover a group of enemies secretly building a fort in the woods. Maybe the players can band together to ruin their plans. If they don’t, then attacks on the locals will be launched out of the fortified base, forcing players to either do something about the situation or lose the town. Adding meaning to combat-related goals requires the same thing that noncombat goals require: imagination to conceive a new way for things to be done, robust code to implement the ideas, and sufficient technology to support the execution of these ideas without too much strain on the server or the client’s bandwidth.