By "instancing' you mean no multiple server worlds, or no server 'bubble' scenarios, or no respawn of NPC laid out in a minefield across the world) ???

Depends what you are after - realism - massive cooperative interaction - minimizing interference of other players

Various limitation of an MMORPG could force you into some instancing (separate server zones with overlap areas is a flavor like instancing).

If you have a Free-for-All everywhere, that lessens some restrictions.

If you want some semi-realistic economy/ecosystem that is another aspect requiring restrictions

If you have finely crafted mini-scenarios which you dont want interfered with (or worse Griefer-ized) that is another problem.

Im looking at a game design for a closed-in world (interior spaces) where a previously unknown generic manhole cover can lead to someplace interesting (and part of the player's somewhat personalized journey). So INSTANCING there is a fundamental aspect of the game.

Huge worlds Ive seen have been largely "Pretty Deserts" with anything of real interest far and few between. (and then the interesting places often swarming a result of the cattle-ramp fixed mission progression).

Mass slaughter of generic opponents is easy. Lots of individual interesting player experiences are hard. It all depends what you are aiming for (and for any individual goodness, unfortunately Procedural Generation of 'instances' is a most practical tool.

If your content is not that fun, then the world should be smaller, or your game should have fewer levels or rounds, so that players can finish the game before they get bored.

i beg to differ. if game play is so lackluster as to negatively impact game scope in the hopes of improving things, then you really shouldn't build the game at all. You're building on a flawed core.

norlin:

sounds like the thing to do is design a scale-able world, so you can add content as needed. Start with "should be big enough" and no bigger. As you create more content, or more players require more content, you can add it. By adding content i mean adding additional sections to the world map.

If you have a Free-for-All everywhere, that lessens some restrictions

What do you mean "free-for-all"?

Like what you replied ... open to griefing/loot stealing/interference with missions all which can cause serious disruptions of any attempt to choreograph scenes and mini-plots and forcing you into very simple 'mission' goals and actions

If you have finely crafted mini-scenarios which you dont want interfered with (or worse Griefer-ized) that is another problem.

There are planned lot of different quests, but I don't see how it could be an issues in sens of instancing/no instancing.

the issue Is : that is one way way (using instances) where other non-designated players are prevented from interfering and constantly making a hash of any planned (choreographed/designed/coordinated) sequence of interaction with npcs/terrain for the Mission the player has been told to do. Otherwise even a small subset of players will constantly be allowed to 'have their fun' disrupting/stealing other players experience and achievements -- and thus much effort of yours to attempt to create interesting situations will be so damaged as to be largely worthless to many players

Huge worlds Ive seen have been largely "Pretty Deserts" with anything of real interest far and few between

Well it's a content issue, I know... Hope to solve it one way or another :)

which is the point that if you want more complex missions (where something unique can exist and it NOT just being the only place where monster X spawns) your 'staged' situations cant be open to be pre triggered/skewed/manipulated into confusion by players who LOOK for just that kind of thing to do to piss off other players.

Mass slaughter of generic opponents is easy. Lots of individual interesting player experiences are hard. It all depends what you are aiming for (and for any individual goodness, unfortunately Procedural Generation of 'instances' is a most practical tool.

Sorry, can't say I understood this sentence...

procedural generation offers generated combinatorics and coordinated creations. Its used to mass generate varying situations (it can be made VERY complex and have whole miniplots, goal sequences (and be player customized) , etc... be machine generated for the 100000 people who will ALL go through largely the same places and missions (which in most MMORPGs are wholey staticly defined.

scenario instances there are (can be) created 'on the fly' and potentially anywhere on your terrain map world

even just as bulk filler for the map -- the generation logic is hard to do right but still can be relegated to the 'filler' content around other more complex/ornate/handcrafted mission scenarios (adaquate semi-interesting filler is a major improvement over largely generic spawn minefields most MMORPGs have).

The best game I can make right now probably won't hold up for 60 hours of gameplay,

Nothing says a game has to have 60 hours of content. Sure, $1 per hour of content and a $60 price point for a new AAA game are pretty standard, thus the expectation of 60 hours+ gameplay. But that's not written in stone anywhere. And fewer and fewer titles these days deliver on that one hour of gameplay per dollar spent. Either they are small, and deliver less than that, or they are huge, and deliver hundreds or thousands of hours for $60.

and I can think of some games that have great gameplay, imho, that didn't hold up for 20 hours. Not because it was lackluster, but just simply time investment requires a certain amount of depth to plumb.

I don't follow... they were too shallow? IE great gameplay, but not enough of it? At least not enough to want to do it for more than 20 hours?