VectorStorm Blog

So I have most of the object placement systems in place now;  you can now place objects within the world (although there aren’t many objects yet available), and they’re stored in a useful manner.  The final missing piece is needing the game to throw away the renderable geometry for structures once the camera goes too far away;  currently, all placed buildings keep their geometry in memory, instead of throwing it away and re-building it when the camera comes back close to them again.

Anyhow.  To get an idea of scale, and what placement of large, potentially city-sized objects was going to be like from such a low “human’s eye view” perspective, I set up the early-stages procedural geometry generator to give me simple cubes.  The one in the image above is a pretend-town that I just placed (okay, I’ll confess that the game thinks that it’s a graveyard, as I haven’t yet hooked up the unlocking of other building types in the UI).  The town shown here is actually quite small, compared to most MMORPG towns;  perhaps about the same size as an outpost like Ratchet in WoW.  It’s only 60 meters by 60 meters, and is 10 meters tall.  With such a (comparatively) small town being so massive on screen (and therefore difficult to place exactly where you want it), it looks like the player is definitely going to need a more aerial viewpoint for placing these things, and probably for landscape features as well.

Luckily, I was sort of expecting that, and already have plans.  Although I think I’m going to work on procedurally generating individual buildings and other structures first, before putting my cunning plans into effect.  ;)

In World of Warcraft and most other major MMORPGs, large cities are much larger than they ought to be, according to the scale of the world.  By this I mean that the inside of Ironforge is substantially larger than it appears from the outside, Stormwind is nearly as big as the whole region it’s ostensibly inside (probably now larger, since Stormwind has been expanded to include a port), etc.

One of the perks you could earn in the 1.1 development branch of MMORPG Tycoon was being able to create “capital cities”.  From a game mechanics point of view, “capital cities”, while very expensive to build, would provide all the services of each of the other building types (taverns, quest-givers, inns, auction houses, etc), and existed within their own server instances, which meant that they didn’t contribute toward the network limitations imposed on the regions they were in.

I’m planning in 2.0 to have similar rules about capital cities, but I’m interested in whether people would like their capital cities to be fully built out to proper MMORPG-style scale, or whether that would just get in the way, and the cities should be built to the same compressed scale as the surrounding world.  In practice, under the “full scale” approach, a capital city would be a type of region;  so instead of marking your region as wilderness that contains monsters of levels 1-5, you’d mark the region as “city”, and it would be filled up with roads and buildings, including the necessary “useful” feature buildings.  Once built, you’d be able to walk around inside the city (though not inside the buildings inside the city).  Under the “compressed scale” approach, the capital city would be a single (large) block of city geometry.  It would be much much smaller in expanse, and you would not be able to enter it or see what was happening inside.  On the other hand, it would take much less time to walk past, and wouldn’t be as expensive to draw, so would be better for people with old graphic cards.

Anybody have thoughts on which way they’d prefer?

With a lever and a place to stand, Archimedes could move the world. One wonders whether he, too, was trying to avoid z-buffer imprecision.

In a world that’s 20 kilometers on a side, OpenGL really doesn’t like looking at things which are very far from the middle;  all sorts of little glitches pop up as you move away from zero, mostly to do with penetrating objects, such as where water meets the land.  So I’m now moving the whole world around as the camera moves, to keep the camera and everything nearby within a few hundred meters of zero.  Oddly, the most difficult bit to make work with all this were the clouds and the water ripples.

I guess that water and water vapour don’t respond very well to the application of a lever.

You’ll all remember the zones from MMORPG Tycoon 1.0;  you set level ranges on them, to specify the monster levels to be found within them.  In 2.0, these areas of your MMORPG world are called “Regions”, and will have editable boundaries and configurable graphics settings.  There will be impassable boundaries between all regions (once finished, these impassable boundaries will usually be mountain ranges.  But here, I’m just using fifty-meter-high plateaux), so your subscribers will only be able to travel from one region into another via paths or roads that you create.  Here, we’re looking at a corner between four different regions.  For the sake of interest, each of these four regions encloses approximately 1.7 square kilometers of space, and there are just over 200 regions in the full MMORPG world.  Though about half of those are ocean, by default.

Now that I finally have these regions working properly (though unattractively) within the streaming world system, I’m ready to start placing buildings, walls, and roads within the world, as well as other terrain-changing features.

Early video games were typically single-player games.  Players had a limited number of lives, were attempting to achieve the highest score in an endless string of levels of slowly increasing difficulty.  Home console games in this era were usually designed around this approach as well.  The video games were tests of skill and reflexes, or else of pattern memory.

Later, some video games began to try to tell stories, or require logical thought from the player, and took some of the pressure off of the player.  These games usually had definite endings, instead of repeating infinitely the way the classic arcade games had done.  These would be games like Infocom’s text adventures and the early graphical adventures by LucasArts and Sierra.  These were still very hardcore games, where death was frequent and often unexpected, just like in the more classic skill/reflexes games which were still being made.

But then an interesting thing happened; somebody decided that “deaths” in an adventure game were a bad thing, and more than that, that one should never have to re-start the adventure game from scratch because they’d saved their game at a point where they’d done something which rendered the game non-winnable.  This approach was used for most of the later LucasArts adventures, and some of the Sierra adventures.

Now, this next bit is interesting.  While adventure games have largely died out from the modern video gaming world (with the exception of the ones now being produced by Telltale Games), the stories they told, the focus on reaching an “end” to a story rather than achieving a high score, and the practice of not aggressively killing off the player have all migrated out from adventure games and are now expected staples of most other game genres.

What’s more, as video gaming has been courting more and more mainstream audiences, it has been changing from tests of skill or brains toward being a more passive entertainment, where danger levels are far lower than they were in the past, and demands on player skill levels are far lower than before.  At the same time, though, control complexity has skyrocketed, requiring the players to master a much wider range of gameplay options and far more controls, but requiring far less precision in their application.  The health bar is just one example of this;  in early games, a single mistake would kill you outright, and if you died three times, your game was over;  now, there’s far more of a safety buffer;  you can take several hits before dying, and if you die, that’s  okay, you’ll respawn somewhere nearby and be able to continue, and you have infinite lives.  Recently, it’s even become pretty standard that your health bar even refills if you manage to stop taking damage for a short while.

Now we’re even starting to see games which are more of a creative tool than “games”, per se.  This would be titles like The Sims and Minecraft;  games which you can play for a short or a long time, but which provide no real game-designated “goals”, and no “end”.  The player never has to restart his game, never reaches a “no-win” situation, because there’s no “win”.  The player never dies, because the player doesn’t really have a presence in the game.  The player very much exists as an entity outside the game, looking in and influencing the game indirectly, from afar.

These games have certainly been around for a very long time, going all the way back to (arguably) Little Computer People, and later, SimCity..  but this type of game is still a vanishingly small minority of the games being produced at the moment, so there aren’t very many games to look at to figure out what keeps people’s interest, and what loses it in this sort of mostly-undirected play.  And this is undoubtedly being one of the most difficult things for me to get a handle on, as I’m designing MMORPGT2;  there are certainly days when I say “Why would anyone ever want to play this?  You don’t actually -do- anything except fiddle around with the rules that the simulated players obey?”