So I don’t think we’re actually that far apart either… I wasn’t advocating doing everything in those articles — after all, I spent a chunk of time pointing out everything that went wrong.

Yeah, I think so. I still believe that “mainstream” target is going to be very directed, accessible, simple content. As the market grows its IQ and patience drop. But I don’t think this means only static content and I think a developer is stupid not to consider any automated ways to generate the data in which to seat their tailored/directed/accessible/simple content.

Nature has self-correcting mechanisms. Why not this virtual ecology system?

Actually I think that nature is very complex and our methods for modelling it and understanding it are still very naive/limited. Nature does have self-correcting mechanisms, which developed over millions of years, genetic manipulation acting as super-designer working in parallel in trillions of critters at once. Taking climate modelling as an example, we have made a lot of progress but (if you live in the northwest anyway) you know that weather prediction is still an incredibly inexact science. Also, in an MMORPG you are dealing with large numbers of players who actively want to screw up your system if they can. I still believe that these are very, very hard problems that won’t be solved anytime soon.

Guess what? There are multiple markets for interactive entertainment, multiple segments of each of the multiple markets, and there is no such thing as a single unified MMO market into which all MMO experiences must reach to attract and retain subscribers.

I agree. But I think that more casual play is still going to be the mainstream mode of play. More information from Raph only convinces me further of this. The casual market is full of players who want to play very simple games for an hour or so after work (or during work ). The players who are playing less and less with each MMO are looking to play games that require less investment and are burning out on onerous experiences that require them to invest huge amounts of time and thought and penalize them if they aren’t better than other players or don’t have more time than other players. I think that we are seeing a strong movement towards “alone together” play with other outliers taking shape. I consider myself not mainstream (which really isn’t that much of a shock, rarely in my whole life have I been “mainstream”) and am not interested in WoW or “casual” MMO play (although being out of college I certainly want to be able to play less hours per week and still feel like I can compete). I want the ecologies and such. I just think that, to be realistic, one must realize that as the market grows this stuff will be increasingly a specialized taste.

niche creation is good for business.

I also agree. However I think a key factor in succeeding in the creation of niches will be to identify that you are indeed targetting a niche. I am simply trying to bang that point across and get people like Raph to start talking about how to create cool niches instead of (as I read it) trying to infer that really this stuff will be mainstream.

I think 50 years from now we really will have that kind of stuff… because of this.

This is a good example of the sort of thing I am talking about with regards to confusing the speeds at which CPU speeds or manufacturing speeds advance versus the speed at which understanding of complex systems advances. How does having huge nanocomputers, for example, help you to solve the problem of understanding the brain? It gives you some nice shiny tools but it doesn’t create understanding in and of itself. Multiplying computer cycles by 10,000% is great if you want to do a matrix multiplication (order n^3 — i.e. it takes on the order of n^3 units of time per size of the matrix) but what if you want to solve a problem that is exponential or NP-complete (x^n, in other words the size of the problem increases the exponent of the time to solve the problem). Computers have become literally many millions of times more powerful since the 1950′s when AI began with quite a bit of optimism and we still haven’t made that much headway towards good AI or understanding of the brain.

Yah. I still don’t understand why people are convinced that if a player games the system, it becomes broken. Abuse it, and it dies. Or not.

Joe Ludwig said:

Pushing the boundaries is a great thing, but you have to be careful to pick your battles. Many of the failures in MMOs happen to teams that bit off more than they could chew. Sometimes that’s because they couldn’t handle making any MMO at all, but usually it’s because they tried to push too many boundaries.

I totally agree. Biting off more than you can chew can be fatal to your game. (Duke Nukem Forever, anyone?) On the other hand, making a bland, derivative EverQuest-alike won’t net you the kind of playerbase you need to be successful, these days. If you want to be successful, you’d better try and push SOME boundaries. Try and give players something compelling and new.

Raph said:

I’d love to see stats on “player fatigue” on all these different markets. I only know the virtual worlds well enough, and I can tell you that the typical player plays for two years or so, gets jaded, and gives up.

Anecdotally, I fit that description to-a-T. DAoC for about 8 months, SWG for 2, CoH for 2, a short hiatus and then WoW for about a year. At that point I tried EVE for 2 weeks and then gave up on the current generation of MMORPGs altogether. I’ve been everywhere, tried everything, killed 10 foozles more times than I can remember. Until I see something compelling and new, I’m sticking to free games like Continuum (which I’ve been playing on and off for almost 9 years now and still enjoy more than most MMORPGs).

StGabe said:

50 years from now is a long time. Maybe long enough. But basically, yes, I think that we won’t see a lot of that even in 50 years. Mostly because the knowledge that we will need to play VR games runs on a different clock than the knowledge we need to run these vastly complex ecologies or AI. VR technology, in comparison, is relatively straight-forward.

I think 50 years from now we really will have that kind of stuff… because of this. I know some pretty smart people who have looked at the predictions for MNT and they aren’t convinced. But I’m firmly convinced it will be everywhere within my lifetime. Not only will we have incredible (barely conceivable) amounts of processing power to throw at the problem, but we’ll also be able to build neural networks with the same sort of density and complexity as human brains, which will likely lead to true artificial intelligence (whether it would be feasible–or ethical!–to apply those things to gaming is not as obvious… but I bet somebody will sooner or later).

Raph said:

Handcrafted content isn’t just quests and stories. It’s everything — all the data you need to enter. Manually filling out tables of which ingot spawns can be used in which zillions of craftable knives is also handcrafted content. Having a starving wolf as opposed to a normal wolf is creating two separate monster entries, likely by different designers who may not know what the specs on the original wolf are. And so on. There’s countless issues like that. It all boils down to expense.

That is why we definitely need to figure out how to automate more content generation tasks. It would be nice to have more dynamic content, but it would also help a lot if we could create good static content more cheaply. Better tools can always help with this (improving workflow, catching/preventing more user errors, data validators, and automatic simulation cycles similar to the edit-compile-run cycle used by programmers are just a few things that come to mind). Of course the right architecture can also help significantly to avoid wasted time. Do you really want to enter all the attributes for your 10 kinds of wolf by hand? Or should there be some sort of “auto-templating” system where you define one generic wolf template, and then create the specific monster types from it by plugging in a level, a name, a color and then tweaking what it generates a little bit? The UO resources system also seems like the kind of thing that could speed things up, because there’s less stuff to hand-code to make it work. As a tools developer, one should be completely pragmatic about stuff like this. A feature that takes you two weeks to implement, but enables the artists to create 5-10% more stuff in a week, is probably a large win for the team.

Lastly: I don’t think having an audience for your creation has anything to do with whether you should pursue it or not. If you’re interested (and if employer constraints permit), pursue it! The world would be a dull place if everybody just made “mainstream” music, “mainstream” movies and wrote “mainstream” books (or games!).

My current hobby project is writing a code generator which generates C and assembly implementations of an emulator for certain CPUs–the ones in NES, SNES and Gameboy actually. There’s probably about 5 people in the world who think my project is interesting. I will be pushing a few boundaries when I get a little farther along with it. But basically I’m only doing it because its really challenging and I find it interesting.

Also, IMHO, the stuff you suggest in place of this, like the randomly attached hunger tag, the notion of animals accepting prefix tags — I’ve done stuff like that too, on Legend. It’s not really static content either.

So I don’t think we’re actually that far apart either… I wasn’t advocating doing everything in those articles — after all, I spent a chunk of time pointing out everything that went wrong.

It is at this point where, IMO, a crucial threshold is broken and we’ve entered into pipe dreams. Once you have your such tightly coupled and dynamic content for players to screw up they will screw it up.

Nature has self-correcting mechanisms. Why not this virtual ecology system?

I would be a lot more positive about what you had to say if you were indeed pitching it as niche instead of (as I see it) trying to convince yourself that this stuff is going to be mainstream. I think that to get a lot of this to a market that is interested you’re going to have state, flat out, that you are targetting not the mainstream but a derivative of the mainstream with very specialized interests.

You constantly claim that most players (i.e., the mainstream) don’t want to see such a virtual ecology realized. Where are you getting this information?

Implementing the virtual ecology as Raph suggested achieves a subtle subconscious experience for players that provides a significant level of immersion unsurpassed by any of the pseudo-interactive environments in today’s MMO experiences. This virtual ecology is certainly not something you would find typical players saying "I want one too" because this is a background system that drives the interactive world rather than a directly interactive system, such as combat and crafting. You wouldn’t find typical train passengers saying "I want pistons and valves" either. Typical train passengers want to travel to their destination. Typical players want a vehicle for immersion, for exploration, and for social integration. Properly implemented, the virtual ecology is effective for increasing the level of immersion, for providing more opportunities for exploration, and for enhancing social integration by enabling expectations of the unexpected that drive the establishment of communities.

You also seem to be of the opinion that niches are bad for business. We may just be using different definitions, but I’ll say this: niche creation is good for business. Read Strategy as Ecology. The article discusses the three critical measures of health for the business ecosystem: productivity, robustness, and… niche creation.

Also I think that you have to recognize that a lot of the “jaded” gamers were actually jaded with respect to games they considered to be too complex. WoW came out and they all breathed a collective sigh of relief. “You mean I don’t have to lose something when I die? I don’t have to work my ass off to create an actual story? I just log in, get handed a scripted quest and do it? Great!” They are giving up on all these pipe dreams of emergent narrative and are relieved to just play a game that fakes narrative in a convincing fashion.

Didn’t you say something about avoiding generalizations? You’re assuming that there’s a single market of gamers — that World of Warcraft has demonstrably captured more than half of that single market, and that this single market of gamers is only interested in the casual gameplay of the World of Warcraft total experience. Guess what? There are multiple markets for interactive entertainment, multiple segments of each of the multiple markets, and there is no such thing as a single unified MMO market into which all MMO experiences must reach to attract and retain subscribers.

Furthermore, simply because one MMO experience appeals to a different market and/or a different market segment does not automatically lessen the value and equity of the experience. The advanced virtual ecology system may not appeal to World of Warcraft‘s target market, but that doesn’t mean the virtual ecology-driven MMO experience will not attract and retain subscribers elsewhere. The fact of the matter is that direct competition with World of Warcraft is not necessary.

Perhaps I’m completely misunderstanding your message? Anyway, I’m responding to that which I understand you’re communicating.

I read the original posts on the resource systems again and what I was responding to was the second half of the second post. The bit where AI enters and we have wolves starving because they can’t find enough rabbits. It is at this point where, IMO, a crucial threshold is broken and we’ve entered into pipe dreams. Once you have your such tightly coupled and dynamic content for players to screw up they will screw it up. Up to there, the system didn’t seem too ambitious nor even that much more than has already been done in worlds like SWG and EVE. And I’m definitely in favor of using procedural methods to bootstrap your world. I never wanted to indicate otherwise.

Question– Is the result likely to be usable and worth the time invested?
Yes, for an ecological data layer and basic AI with the necessary tools to leveraged the platform with handcrafted content. Not yet worth the time invested for an fully integrated and dynamic ecological system due to the complexity of balancing the system (Raph stated the major issues). The perspective is to enrich the world platform on which games can be build on.

Question– Can we spare the time or do we need it for other things?
It’s a question of fixed sunken cost versus variable cost. Spending more on fixed cost can provide operating leverage and enables variable cost content to be create at a faster pace. However, a key consideration noted already is that if you don’t have enough time or budget to get the core set of features done right at launch, there will be no resources available to develop a broad and robust ecological platform. Some will see this platform as a core, other will not. Going to let the board room or the designer room decide?

Question– How much will this complicate the rest of the project?
An ecological data layer and basic AI should in theory simplify the development of other areas. However, as the industry has not gotten this done to pat, it may complicate the project. So, the most market efficient sweetspot is likely to be a middleware developer.

Hope this is a good summary of the common points.

Frank

Handcrafted content isn’t just quests and stories. It’s everything — all the data you need to enter. Manually filling out tables of which ingot spawns can be used in which zillions of craftable knives is also handcrafted content. Having a starving wolf as opposed to a normal wolf is creating two separate monster entries, likely by different designers who may not know what the specs on the original wolf are. And so on. There’s countless issues like that. It all boils down to expense.

I think you are oversimplifying what can be done with static content without your dynamic ecology. Procedurally generate terrain and scatter resources. Create a “hungry” tag that any “animal” can spawn with. You just have to then create a set of guidelines for your static designers to work with (i.e. all animal names have to accept a prefix adjective). There are ways to streamline this process that are far easier, and far more manageable, than creating an ecology that will almost certainly break the moment you add players to it and won’t make the entire project that much more unmanageable.

If dynamic narratives aren’t your goal then I really don’t see that you are gaining that much and I come back to most pressing of my three questions.

– Is the result likely to be usable and worth the time invested?

Probably not. The system will still be rather brittle and can be faked through much more direct, more manageable methods. If you can’t get at truly dynamic narratives then it’s not clear that the gains are worth the inevitable addition of complexity and confusion you are foisting on the player (unless you are willing to bite the bullet and declare it a title targetted at a niche of simulationist-interested players).

– How much will this complicate the rest of the project?

Greatly. Suppose there is a great imbalance in spawns after release (very likely). With procedurally generated datasets you merely write a script to massage the data. With an ecology you have to rewrite the code.

If this is alls speculative then I wonder why you keep mentioning doing away with handcrafted content as, to me, handcrafted content means people hand-populating quests and stories. Procedurally generating terrain and populations of critters is well and good. I’m not sure how an ecology helps with this though.

Hmm, I would have thought it was apparent from the articles why having that sort of data layer and basic AI would be helpful even in creating handcrafted content.

Handcrafted content isn’t just quests and stories. It’s everything — all the data you need to enter. Manually filling out tables of which ingot spawns can be used in which zillions of craftable knives is also handcrafted content. Having a starving wolf as opposed to a normal wolf is creating two separate monster entries, likely by different designers who may not know what the specs on the original wolf are. And so on. There’s countless issues like that. It all boils down to expense.

Other developers I talked to on the project indicated that things like, “it’s too hard to get routine stuff done” were very common on exit surveys. Later in the project, this was cited as a reason for killing a lot of the dynamic market and dependencies between players. These were, of course, the complexities of the game that still managed to keep it afloat with at least a niche audience.

I don’t know which developers you talked to, but I can tell you that the number of people assigned to static content versus the number of people assigned to dynamic content was literally an order to magnitude difference. Yes, many things were hard to accomplish given the toolset, but how much of that had to do with dynamic content per se… some, but not even the majority, IMHO.