One server, the Eve model, and MMORPGs. Viable?

[Falconeer]

One of the things that bothers me the most about "virtual worlds", no matter how sandboxy, diku, theme park or whatever they are, is how difficult is for people who play in the same universe to actually meet each other, or to feel they are playing in the same environment. Forums and official boards are the only place where the whole community that plays a game can meet, and they are OUT of the game itself, out of the virtual world. Patch notes are, in a way, the only real cross-server event. I don't mind instances (I mean duplicates of a given zone that pop up automatically when a zone is overflowing), I understand they are a necessity, but being in the same (virtual) place with someone I know without actually being able to meet them because of a permanent choice we made when we started playing seems like a disappointing development of what we hoped was the first step into "cyberspaces" at some point in the 90s with the first few persistent games.

You could put it in a different way: one of the things I love the most about Eve is the fact that it all happens on one big universe (which is still technically a lot of servers, right?), as not only I feel like I share the same "space" with everybody else, but that one big universe is the same one we all care for. Hell, they can even print a magazine with stories and recaps of what happened "in game", and it all still matters to everybody cause it's one single shared universe.

A very far from complete list of pros and cons, which I invite you to contribute to:

Pros:
+ No need ever for server mergers.
+ Literally hundreds of thousands of players available at all times for dungeons and PvP battlegrounds/warzones.
+ One world that everyone cares for, events in the community and the meta-game affect everybody.
- ?

Cons:
- "Huge city syndrome" vs. "Small town comfort" (where everybody knows your name): you are just a number and it's harder to stand out. Communities, ironically, might have a harder time forming.
- Technical difficulties affecting performances.
- ?


-------------------------------------------------------------------------------------------
Now, I have a few questions for you all here, technical and conceptual:

1) If you were a developer with _unlimited_ resources, would you go for a unique server for your MMORPG? (Sure, we are talking about multiple instances of the same zone to accommodate everyone, unlike Eve). If not, why?

2) Would your choice be different if you were going to make a diku or a sandbox?

3) Is it technically feasible with today's technology? I mean, if you create multiple instances of an area, aren't they (or couldn't they be on) different servers to begin with?

4) So, what's the cockblock to having a million concurrent players in one universe considering that such "one universe" is actually a collection of multiple physical servers and instances (and copies of the same zone) anyway? What's the biggest technical hurdle to this? How can EVE manage this, without zone duplicates, while other games seem to cap their servers at very lower figures than EVEs record (which is not in the hundred thousands, but still)?

5) Could this be an emergent trend with the improving of technology, if nothing else as a way for companies to avoid the dreaded "empty servers" effect that hurts gamers' morale, popularity, and the subscriptions (money) to the service itself?

6) Not counting Anarchy Online (it started with one server, but now has two), Guild Wars (so instanced it barely had any shared space outside of glorified 3D lobbies), or games that have only one server because they can't even fill that one, how long do you think it'll take for the first "traditional" MMORPG to attempt something like this?

7) Wouldn't it be awesome?

[ajax34i]

Um, if there are "unlimited resources" (i.e. money), what would be the motivation to even develop an MMO?  Already have the moneyhats...

The technical hurdle for "one universe" is all players going to Jita Station 4-4 (to one spot) all at the same time.  Eve has huge distances and most of the time everyone you see in-game is a bracket (don't even have to render the ship model), so the graphics lag is minimized. 

I don't think that DIKU MMO's have an incentive to do this.  For PVE gameplay / progression, it's detrimental to have a dense crowd (this holds true in EVE PVE gameplay, too).

[Falconeer]

In the model I am implying there would be many instances of Jita, so when like 200 players are heading there it creates Jita2, Jita3 and so on. It's not Eve, but it's still better than different servers the way it is now. Of course you can't have 1000 players in the same zone for real, you need instances. But do you need differently named, uncrossable servers?

[Numtini]

Eve can do a single unsharded server because there are like a half dozen or dozen different zones in terms of artwork and everything else is just a few pieces of math about where objects/planets/stations/moons etc are. They could probably double the size of their universe overnight if they wanted. It would be impossible to create that much physical space in art assets unless you were going to copy/paste and in a non-space game, that's going to be very very obvious very quickly.

Every game released since WoW has faced the problem of empty servers. Even in a relatively successful game like Rift, the market is so transient that you are going to have to reduce servers at some point after launch because people just don't stay in the same game for years and years like they used to. It seems like making everything instanced makes a lot more sense to me. GW2 is doing this technologically, but they're retaining the concept of a "server" which I think will come back to bite them when or if they need to merge some. Despite the fact that everyone since WoW has needed to merge servers, players still take it as a sign of an impending apocalypse.

I appreciated the feeling of a server being a small town, but every other design decision games have taken in the last few years has gone against community and in favor of convenience. It seems silly to retain this one particular bit. Just give me a pull down in the top left to switch instances on the fly like Korean games do.

[Sir T]

In the model I am implying there would be many instances of Jita, so when like 200 players are heading there it creates Jita2, Jita3 and so on. It's not Eve, but it's still better than different servers the way it is now. Of course you can't have 1000 players in the same zone for real, you need instances. But do you need differently named, uncrossable servers?

That's the way STO does it. Everyone is one one "server" there too.

[IainC]

I know of at least one DIKU MMO in production that is going to be using a single shard architecture. The technical problems can largely be overcome through better hardware (both at the client end and the server), just expect to need an enthusiast level rig and a beefy connection to play MMOs in future. As the tech trickles down and becomes more commonplace I expect that this will become a viable solution to server population problems with existing games in future.

[Malakili]

In the model I am implying there would be many instances of Jita, so when like 200 players are heading there it creates Jita2, Jita3 and so on. It's not Eve, but it's still better than different servers the way it is now. Of course you can't have 1000 players in the same zone for real, you need instances. But do you need differently named, uncrossable servers?

This is how Cryptic did Champions Online and Star Trek Online.  I didn't play STO, but I played Champions for a bit.  One of the problems with Champions is that it was very hard to actually run into the same people unless you put them on your friends list.  You basically said this in your big city v. small town comfort point.  Another thing is that it REALLY kills the virtual world feeling.  Nothing takes you out of the game world more than hearing "The robot attack event is happening in MIllenium City!" Only to jet yourself over there and find out actually its happening in Millenium City 54 not Millenium City 32.

It is definitely nice that you CAN play with anyone/everyone, but heavily instancing zones that way really makes it feel less like a virtual world in a way that definitely turned me off some.  Although you are limited to the people on your server in something like WoW I at least felt like it was in part a place I was somewhat invested in and I knew that when I ran into people I had a shared experience with them.  In Champions Online I had a shared experience with no one, or at least, it felt that way.  Unfortunately, it seems like this is a result which is the exact opposite of what you are looking for, but I don't know how to avoid it with this model.

[Kageru]

Is there an advantage to doing so is a question that also needs to be considered. Eve's approach makes sense because the end-game is "global war" and that doesn't work if everyone is split into 100 person shards or independent servers.

If your games largest activity is a 5 person dungeon, a 20 person raid or battleground then having the ability for every player to be in one spot isn't a particularly useful attribute. And certainly not worth sacrificing appearance, performance or server community (weak as it is) for. I mean it's technically cool, but what gameplay does it give? And Eve, even with quite a minimal landscape, has had immense battles with lag. The current solution of reality running at 10% of normal speed is clever, but in any other game it's going to be hard to sell as an advantage.

I would say that GW2 has a solid answer. You have a server with a relatively fixed population, you are given a server sized reason to work as a community (server vs server PvP), and you can go play on another server if the people you want to game with aren't on your server or your server is unavailable.

[Lantyssa]

I prefer a single server, but technical limitations for anything more than a space game do cause a lot of difficulties.  GW2 is close to my ideal of having a "home instance".  The more worldly the game is, the more a single server is important.

Some of those limitations can be reduced by providing enough landmass to let players spread out.  However, that has problems of its own in that you have to produce those assets and you have to figure out ways to encourage players to spread out instead of congregate.  Almost every game ends up developing some kind of social and economic hub, which tends to run up against technical limitations.

[Numtini]

It will be interesting to see if GW2 merges servers if/when it needs to and how the easily hopped servers and transfers alter how players react to it. They could use WvW to spin it as a competition thing rather than a "the game has failed they're closing servers" thing.

[Hawkbit]

I agree with the concept overall and I think we're headed there, but flip this on its head for a minute.

Server culture is one major reason to retain separate servers.  My EQ/WoW experience was molded completely differently by being on the servers I chose.  On WoW Earthen Ring it was a heavy population of mostly good people that are always, always busy chatting.  If I log over on Spirestone, there's virtually no chat and every single person only does things for their guild.  The trade interactions are completely different experiences. 

Not that throwing everyone together doesn't have potential for good things to happen, but generally speaking its okay to allow people to create small separations socially.  Note that my first degree was in anthropology, so seeing how cultures form differently under the exact same framework 100s of times is a doctoral thesis in itself.

[Scold]

I want to see us moving in exactly the opposite direction.

"Radically small" servers, aiming for a concurrent population of 100 or so at most, or perhaps even 40-50, with the game then having thousands or tens of thousands of servers. And then design the game around that, making players able to seriously influence the world around them.

Your game would have thousands of servers, and the 'state of the world' would look radically different depending on which server you play.  On one server, a collaborative atmosphere rules and players are working together to try and explore the harsh outer reaches of the game's wilderness. On another, a tyrant king (an actual player) has everyone under their boot-heel, and rebels (also players) are looking for ways to undermine and depose him. On a third, a formerly cooperative arrangement has splintered into civil war between two (player) factions. Design the game such that there are many, many cool 'roles' in the world to take on. And then do radically small servers such that players can actually reasonably dream of taking on those roles, rather than just ending up as Grunt #328195.

If you build your game around this premise, and can run the thousands or tens-of-thousands of servers 'virtually', it's totally doable... it would never work if you slapped it on top of existing DIKU. The ability to take part in an intensely personal experience like that might not be for everyone, but it would get an intensely devoted community, I think. I imagine that in an environment where player actions could actually shape the world,  GoonSwarm-esque "zerg destroyer of worlds" groups would pop up, attempting to hop from one world to the next and undermine whatever they have going on for the lulz, so you'd have to account for that and design accordingly.

I actually imagine this being most feasible in a 2D MMO with "retro-style" or otherwise non-cutting-edge graphics, so art assets and laborious design of physical spaces isn't a huge limitation. Games like Realm of the Mad God (and MapleStory and RuneScape, for that matter) have proven that 2D lo-fi MMOs can go gangbusters if properly designed.

[Stormwaltz]

Another Con related to "huge city syndrome" is the corresponding increase in grief play. Dunbar's Number / Monkeysphere theory suggests our primate brain is capable of maintaining a finite number of social links. Anyone encountered beyond a player's circle of familiar faces is often considered no more real than an NPC. And you know how people abuse NPCs.

[palmer_eldritch]

I agree with the concept overall and I think we're headed there, but flip this on its head for a minute.

Server culture is one major reason to retain separate servers.  My EQ/WoW experience was molded completely differently by being on the servers I chose.  On WoW Earthen Ring it was a heavy population of mostly good people that are always, always busy chatting.  If I log over on Spirestone, there's virtually no chat and every single person only does things for their guild.  The trade interactions are completely different experiences. 

This is a fascinating topic to me. How does that happen? I don't know the history of the servers. Were they founded at different times and attracted different types of people for that reason? Are they basically the same, but the behaviour of the players evolved differently for some reason? What could have done that? I don't expect anyone to have answers necessarily but it would make an interesting thesis.

[Venkman]

Others have already covered why the Eve model works and is unreplicable by fantasy MMOs. You can do it in PvP because you are more content than the game itself. In PvE, you can't have nearly as much consumable content. So duplicating it across shards is the only way to launch something you can afford to build.

Instead of trying to design different servers, the industry needs to prioritize server merges before launch. Stop drinking the kool aid believing you'll be huge forever. Stop de-prioritizing account migration and cross-server communication tools as "nices to have". Assume you'll merge, so when you do it, it's easy and painless.

Related: I don't think server communities are important to retain. They're based on emergent behavior, impossible to predict, not worth trying to control even if you could, and will evolve with whatever conditions come about anyway, and will quit for reasons completely unrelated to your social engineering anyway. Ensure they have a density of other people around and let them figure it out, because you won't get rewarded for anything else.

Another Con related to "huge city syndrome" is the corresponding increase in grief play. Dunbar's Number / Monkeysphere theory suggests our primate brain is capable of maintaining a finite number of social links. Anyone encountered beyond a player's circle of familiar faces is often considered no more real than an NPC. And you know how people abuse NPCs.

You need a density of people beyond your social group to believe a world is a vibrant community. Otherwise, it's just a multiplayer game. That's fine, but it's not MMO.

* And yea, I consider this a genre now because "making an MMO" comes with as many rules for how it "should be done" as any other mature genre.

[Falconeer]

I didn't know Champions was like that. But yeah that's what I thought. They are already duplicating the content across many shards. Why can't those shards be all in one server, with no boundaries for players. What's the problem without making one server with hundreds of duplicated instances of the same zone as opposed to one or two instances of the same zone, multiplied for X server with unpassable boundaries? If Champions did it (and I didn't know it, but sounds exactly like what I was thinking) what's stopping other companies? And are there actual technical limitations, if it is already in a non-space game?

[Draegan]

I eagerly await what Blizzard is going to do with MOP and their shared cross-server zones.  Also, GW2 seems to be awesome in this aspect as well, but only partially.

[Numtini]

On WoW Earthen Ring it was a heavy population of mostly good people that are always, always busy chatting.  If I log over on Spirestone, there's virtually no chat and every single person only does things for their guild.

I agree about Earthen Ring. People were nice, it was easy to get pickup groups, I had open invites to a couple of guilds. But LFD put a stake through that heart long ago. That to me is the thing. There have been so many other things destroying community that retaining individual servers seems pointless.

[Draegan]

On WoW Earthen Ring it was a heavy population of mostly good people that are always, always busy chatting.  If I log over on Spirestone, there's virtually no chat and every single person only does things for their guild.

I agree about Earthen Ring. People were nice, it was easy to get pickup groups, I had open invites to a couple of guilds. But LFD put a stake through that heart long ago. That to me is the thing. There have been so many other things destroying community that retaining individual servers seems pointless.

You have to create tools to facilitate community building rather than creating /randomchatchannel

GW2 has an interest aspect where you can join multiple guilds and then select which one you want to be part of at any given point.  To expand on this, create a web device that you can join guilds/groups on the official website or create embeddable devices to put on individual webpages.  Use portable devices or web pages to then create social groups.

Then once in game you can switch between these groups/guilds at will depending on your mood.  You can differentiate between a guild and a group too.  Combine this with fluid servers, then you have the perfect system.

Now you can create your own 10 man guild with you and your friends.  Then you can join the "Guildwars2Junkies" group and chat with both at the same time.  Then you can join the SomethingAwful group and turn of GW2Junkies and turn on SA.  Then you can have communtities that catch on and grow from official forums for specific things.

Groups can be created that are public or private.

I think that is a much more elegant solution.

[MahrinSkel]

Eve can do a single unsharded server because there are like a half dozen or dozen different zones in terms of artwork and everything else is just a few pieces of math about where objects/planets/stations/moons etc are. They could probably double the size of their universe overnight if they wanted. It would be impossible to create that much physical space in art assets unless you were going to copy/paste and in a non-space game, that's going to be very very obvious very quickly.

Procedural generation (which is what accounts for 99.5%¹ of Eve's content) for a terrain world isn't actually that hard, AC did it back in 99, Horizons did it, Wish was going to do it on a really big scale.  But then travel times kick your ass people complain about not being able to meet up with their friends, and everyone so far has given in and put in teleportation systems that made the overwhelming majority of their carefully generated terrain completely useless, "flyover country".

Eve found the answer: Let them bitch.  On shortest-route autopilot it takes hours to cross Empire space, and 0.0 travel is easy but survival is hard.  We need a fantasy themed world that makes a *big* world, and makes the players just accept that getting there (halfway across the continent) from here is something you do only when you really, *really* need to.  Wish (in my version of the "Vision") was going to have the potential for roads and portal systems that could reduce travel times from tens of hours to tens of minutes, if somebody put in the effort to develop them *and* allowed you to use them (I was looking a bit at Eve with this part, although at that point I hadn't actually played Eve).

There is an engineering problem, because it's easy for Eve to separate systems between processes running on different boxes but a fantasy world would need either really annoying zones on arbitrary lines, or all the problems that come with trying to synchronize data across "seamless" transitions.  I have an architecture design that should be able to cope with that, but it would need 18-24 months of development before you could even start building a world on it.  Nobody has the appetite for that kind of thing.

--Dave

[Venkman]

Or, basically, back to Everquest.

Nothing wrong with that thinking. Just requires a brand new audience not trained by 12 years of devolving metagame thinking that derived from emergent behavior (maps, player ports, SoWs for sale, etc). And that audience isn't coming to new MMOs in droves anymore. It's not a coincidence that the average age of core gamers is inching up each year while the tweens and teens aren't showing up to this kind of thing like they used to.

At this stage, the only developers capable of even concepting along these lines are the indies. And unless they can shoestring it forever, whoever ends up publishing will force them to capitulate on some of the "hardcore" stuff.

The world aspect of these games devolved long ago unfortunately. And the size of the playerbase here isn't big nor unique enough for that kind of thinking. We might see some of that from shared persistent mobile experiences once smartphone games finish their inevitable evolution to some type of MMO, assuming the players similarly evolve their wants. But it won't be anywhere near what we were hoping to see in the mid-2000s era when all options were still relatively on the table. And it likely won't even be called "MMO".

[Kageru]

I think you have the flow backwards. The rise of social, mobile and console games all encouraged a more "instant action" (or at least instant activity) demand from games that want a mainstream audience. And MMO's, even strategy games, have tried to adapt to this demand even though it's the polar opposite from "investing" in a virtual world which has virtual "work" required to achieve your long term goals. Gaming while mobile even more so just from the environment.

And most players will happily tell you they want both. They want server communities but not to have to depend on anyone else. They want to grow their character forever but don't want grind. They want the satisfaction of achieving a long term goal but they also want all the fun right now. Pick a target for what the game will focus on, know the audience, budget the game on realistic expectations, get lucky and there's still profits to be made.

[Ingmar]

 We need a fantasy themed world that makes a *big* world, and makes the players just accept that getting there (halfway across the continent) from here is something you do only when you really, *really* need to.  

So, Vanguard?

[Venkman]

I think you have the flow backwards.

The trajectory of early PC games is the same as the modern social and mobile systems. It's just that the latter are being compared to what has become of PC games over a three decade arc.

Otherwise though: yes, players will tell you they want everything. So will people in Marketing, Sales and management 

It's the old "time to foozle" debate. Blizzard always gets that better than most. But to do that is to remove much of what the masses would consider as sucky busy work. Kithicor runs? Screw that, whose got time, just port me there! Waiting to group at Oasis Dock? Seriously, what, you want me to play /gems while I med?! Buying from a player vendor? You mean I might need to haggle or, worse, be social at all?

Anything perceived as "work" is removed because in the end, the growth comes from selling games, not virtual lifestyles.

There'll always be a contingent of escapism seekers who are eagerly full on holodeck. But the games industry has recognized the entire world is a potential market if you scale the game to dabblers.

[Furiously]

My understanding of guildwars 2 is they have planned for server merges. Your name is unique. If you use it on server x. No one can use it on any other servers.

[Malakili]

Anything perceived as "work" is removed because in the end, the growth comes from selling games, not virtual lifestyles.

I know you aren't making this argument and are speaking for the hypothetical audience you are talking about, so this isn't an attack on you.  I really hate this "it is work" argument.  Can people just start saying what they mean "I don't like playing this game."   The absolute MAIN problem with games lately is that they seem more and more designed for people who don't actually like playing games.  They like getting rewards, or they like chatting with their friends, or feeling powerful, or some combination thereof, but what they don't seem to actually like it playing games. 

So, to try and bring this back around to the discussion at hand - what can we realistically hope for?  A huge title with the production quality of WoW isn't going to come along with an EVE style world (besides EVE I guess).   But then, when titles do come along that actually take a shot at it (Darkfall) most of the people even here laugh it off the front page of the board.  For the record, I still hold that Darkfall was actually not a bad game.  Yeah, it had some goofy shit going on, mainly related to the way skilling up worked.  But the game big picture stuff, the game world, the clan system, the clan cities, sieges, etc actually worked pretty damned well.  If you wanted to travel, it was both scary and time consuming.  You didn't carry valuables.  Owning space mattered because it gave you relatively safe dungeons to explore and farm, resources to harvest, etc.

Maybe an even better hope is for a sort of non-MMO virtual world.  For example, I would absolutely KILL for something like the old Neverwinter Nights persistent worlds to come back, those were some of the best "MMO" experiences I ever had, in a large part because they were community made and run, which meant the worlds were often tailor made for the type of community that was playing there, admin could arbitrarily ban asshats, and so forth.  Of course, the monetization of such a game might be a problem, but I just throwing ideas out at this point.  Basically, I don't think were are going to see any virtual worlds or single server type things and the mechanics that must come with such a game coming out of the mainstream MMO market, and I'd frnakly be just as happy with a smaller (but sandboxy) world with a smaller (and controlled) playerbase.

[Venkman]

Kinda three separate points you bring up but I think they all relate. We want experimentation, production values, and fundamentally good games we can enjoy alongside or against other people.

But even beyond that, the more important part is that these games are not the life sucking vacuums that were early MMOs. Because games are mainstream now, a very large market, and most don't want to spend multiple hours a day playing them.

It's basically hard to focus on a small audience when you've got bills to pay :)

[Malakili]

But even beyond that, the more important part is that these games are not the life sucking vacuums that were early MMOs. Because games are mainstream now, a very large market, and most don't want to spend multiple hours a day playing them.

I think part of the problem is that good games ARE life sucking in general.  MMO or not.  Things worth doing usually take a long time to do.  It doesn't matter if it is an MMO and you are building an trade empire (EVE), or learning to play Starcraft real well.  Or, to remove it from video games for a second, learning to play Baseball or Basketball, or learning to play guitar or knit/sew, etc etc etc.    I'm sure as hell not going to sit down with a guitar and then complain after 5 minutes because I feel like I should've learned to play a song already.  But for whatever reason in video games that type of behavior is totally normal.

[Margalis]

Things worth doing usually take a long time to do.  It doesn't matter if it is an MMO and you are building an trade empire (EVE), or learning to play Starcraft real well.  Or, to remove it from video games for a second, learning to play Baseball or Basketball, or learning to play guitar or knit/sew, etc etc etc.    I'm sure as hell not going to sit down with a guitar and then complain after 5 minutes because I feel like I should've learned to play a song already.  But for whatever reason in video games that type of behavior is totally normal.

Worth doing or worth learning?

Eating an ice cream cone is worth doing, it takes 5 minutes. If I spend hours practicing guitar I expect to get something lasting out of it, not sure what you get lasting out of an MMO. Being awesome at pressing 1 when a cooldown is up?

"Radically small" servers, aiming for a concurrent population of 100 or so at most, or perhaps even 40-50, with the game then having thousands or tens of thousands of servers. And then design the game around that, making players able to seriously influence the world around them.

Makes a lot of sense for an asynch game, not sure how much that is an "MMO" experience though.

[Kageru]

Eating an ice cream cone is worth doing, it takes 5 minutes. If I spend hours practicing guitar I expect to get something lasting out of it, not sure what you get lasting out of an MMO. Being awesome at pressing 1 when a cooldown is up?

You get memories. People still fondly remember (and will recount endlessly given half a chance) victories, betrayals and achievements in old games. And a lot of those memories remain because they worked with people towards the goal. There's more than enough evidence of that just on this board.

A game that is flashy but short and easy, like the endless parade of sameish console shooters, probably won't generate any memories other than the fact you completed it.

It's not actually a problem that needs solving. Big budget games need the mass audience and will court the short attention spans of gamers trained by console titles. Anything that takes a lot of time to achieve, or has complex mechanics, will get pared back. A title more willing to aim at a niche market has a lot more freedom so you get weird stuff like Eve, Perpeptuum, Minecraft, Wurm (from what I've read). Of course since everyone dreams of money-hats they generally aim for the former anyway.

[Lantyssa]

A game that is flashy but short and easy, like the endless parade of sameish console shooters, probably won't generate any memories other than the fact you completed it.

GLaDOS would like to have a word with you.

[Dark_MadMax]

Procedural generation (which is what accounts for 99.5%¹ of Eve's content) for a terrain world isn't actually that hard, AC did it back in 99, Horizons did it, Wish was going to do it on a really big scale.  But then travel times kick your ass people complain about not being able to meet up with their friends, and everyone so far has given in and put in teleportation systems that made the overwhelming majority of their carefully generated terrain completely useless, "flyover country".

I think AC had a great system. It had fast travel in terms of bind points.  But binding to new points required crossing terrain on foot.  There were travel hubs which facilitated travel to most popular areas. Communities were formed around bind points- they were your home. On  pvp server (darktide) those bind points were your zone of influence . And at the same time AC world was huge and worth exploring

You could expand the land mass  relatively painlessly with this system and have huge world without huge travel  times (while still having travel itself not obsolete)

[Malakili]

Worth doing or worth learning?

Well both I guess, but perhaps I overstepped my point a bit, which your ice cream example illustrates.  But my point is, if a game is worth playing, it is probably worth playing a lot.  Furthermore, the people who are willing to put in the most time will always set the standard for a multiplayer (especially PvP) game.  Everyone will tell you that a persistent medieval world with castles and wars and sieges and stuff sounds awesome.  But when that opposing guild is willing to play 24 hours a day, they set the standard for what is needed.  If everyone sort of agreed to put in no more than 2 hours a day, you could have a "casual" game like that.  But the problem is it doesn't work that way, and when you enforce the 2 hour limit or otherwise arbitrarily limit the ability of those sieges etc to take place all the time in order to allow the game to be playable by a wider audience you actually fundamentally change the game so that it really isn't the same thing anymore, and suddenly (at least to me) not only does is not feel worth playing 5 hours a day, it often times doesn't feel worth playing at all.

[pxib]

"Radically small" servers, aiming for a concurrent population of 100 or so at most, or perhaps even 40-50... making players able to seriously influence the world around them. [...] Design the game such that there are many, many cool 'roles' in the world to take on. [...] The ability to take part in an intensely personal experience like that might not be for everyone, but it would get an intensely devoted community, I think.

With only a thousand players or so on the whole server, a world of any size is going to seem very dead indeed most of the time. If people want to roleplay inside of (for example) a single castle, they absolutely can... but where are the "rebels" then? How do the factions "war"? How does the "king" rule when he's only online a few hours a day?

In a single large server, it's easy to guarantee that something interesting is happening somewhere. With enough players, it's not hard to find something happening somewhere nearby. Heck, it only required a few hundred players to allow Planetside's "instant action" button to have profitable results. But that's a lot more than 40 people, and it still felt like an awfully small world.

[koro]

On WoW Earthen Ring it was a heavy population of mostly good people that are always, always busy chatting.  If I log over on Spirestone, there's virtually no chat and every single person only does things for their guild.

I agree about Earthen Ring. People were nice, it was easy to get pickup groups, I had open invites to a couple of guilds. But LFD put a stake through that heart long ago. That to me is the thing. There have been so many other things destroying community that retaining individual servers seems pointless.

You have to create tools to facilitate community building rather than creating /randomchatchannel

GW2 has an interest aspect where you can join multiple guilds and then select which one you want to be part of at any given point.  To expand on this, create a web device that you can join guilds/groups on the official website or create embeddable devices to put on individual webpages.  Use portable devices or web pages to then create social groups.

Then once in game you can switch between these groups/guilds at will depending on your mood.  You can differentiate between a guild and a group too.  Combine this with fluid servers, then you have the perfect system.

Now you can create your own 10 man guild with you and your friends.  Then you can join the "Guildwars2Junkies" group and chat with both at the same time.  Then you can join the SomethingAwful group and turn of GW2Junkies and turn on SA.  Then you can have communtities that catch on and grow from official forums for specific things.

Groups can be created that are public or private.

I think that is a much more elegant solution.

So... it's the FFXI Linkshell system, minus the inventory constraints and with multiple chat channels.