What a nice Tool
I leave the MMO Lessons to Ryan. He's got the web meme sown up at this point and they tend to be a good read. A very good thing to be writing down and mulling over whether you are planning your company's next project or if you are slogging away at the reality of the live support and ongoing design of an established MMO.
One thing he hasn't "lessoned" about yet is tools. Tools are it baby. Good tools are worth a gym locker crammed full of inspiring designers. A good tool is better than having John Carmak coding on your project every Sunday for a year for free! Good tools are better than getting asked "Hey dude want a pile of cash to make a Lego MMO?", having no idea why that would ever be compelling and in the end not caring cause it's just so damn cool (Hey MO!).
Tools, kids, are the key to life, the universe, and everything and these Crytek dudes have one impressive tool.
Labels: mmo design
17 Comments:
Yup. Tools are where it's at. Kind of makes you wish you were starting from scratch, eh? ;)
They've made one hell of an engine.
That sky is waaaay too blue. Noone will ever play that game.
Nice mpathy...nice.
I like the clouds. What can I do to buy those clouds Goph?
Easting
Holy shat.
O my gawd!, Thats amazing.
Good tools are such a huge time savings. And if you have someone who is really good with that tool, then you are talking about an individual that can produce some seious output.
Pretty dang impressive.
It's still oriented toward manual creation of intensely complex small areas, though. Thus it presents the same challenges as the present generation of tools for filling large areas with a coherent look.
Someday someone's going to come up with a tool that accepts arbitrary aerial photography plus DEM data as input, automatically interprets the photography to determine what objects are present, selects and processes a library of parametric terrain objects/vegetation types/buildings to fit the photography according to cultural and geographic appropriateness based on Earth location, and generates 3D-modeling of actual Earth locales in near real time.
*That* game will be highly saleable to military organizations as a tactical planning tool, and to a wide variety of governmental organizations and other businesses in general.
The bad news is that this engine only runs on Windows Vista... and it needs a last generation DX10 graphic card to be able to display those shades and cool looking clouds.
I'm worried about the future of OpenGL 3 (released, maybe in summer 2007)... and don't even know if OpenSceneGraph will be adapted so fast to the new hardware available.
In less than 2 or 3 years, the standard will be that hardware... and the visual range of Cryengine2 seems more than enought to think in a possible build and adaptation to use it in massive multiplayer games.
The graphics world it's crazy!.
DOC will have an easier life with only half of those tools for terrain building.
Cid250.
Has anybody shown an engine yet with similarly impressive tools for building large-area environments?
That's a rather sexy tool.
Ryan's dragon web fu is the master of my tiger link fu...
No J, not that I've seen, and I'm looking obviously. You are right though. We've got the satellite DEM (digital elevation map) part down and the imposing of culturals like roads and rivers isn't an insane challenge. The real trick as you not is the procedural generation of population centers in a way that make sense. SimCity is an excellent example. Getting all of those tools together is a chore, especially for a team as small as we are. As you say though, if you could do that all from google maps, and they are getting pretty good at that, you could really have a sim tool, though it would be modern.
There are some tools lke that in the Mil Sec but they aren't anything to write home about quite yet and as was recently commented about our sim "the graphics are not up to par with modern next gen games, however this still puts them about 10 years ahead of the best military simulators".
nice ... just wow ..
Dunno if it's competitively useful to note inadequacies in what initially seems an overwhelming package, but:
The hydrodynamics seem pretty good when all that's going on is noise and small linear stuff. Note however at 7:30 that when the demo briefly jacks up the wave magnitude into what should be a nonlinear range, the physics become obviously wrong.
I also wonder what the engine does about water-surface lighting under clouded-over conditions, and what it can do with wind-wave interactions i.e. spray and whitecaps. Nothing like that is demoed, and those would be pretty essential for naval modeling.
Everything we see is NOT being sent over NetCode to hundreds and thousands of real-time players.
As KFS1 points out: EVE is winning awards by making the game not care about weapon states and limiting environmental information. Everything in WWIIOL requires the information over-load that typical CQB presents to the average soldier.
Merging HalfLife 2's mastery of the shoebox and EVE's MMPOLG universe is the Holy Grail of 3rd generation MMPOLG.
WWIIOL will have to carefully decide what is possible, from what is desirable. GOPHUR - you are still doing the most amazing game with the smallest possible team.
The render engine with TE II is the solution published - and hopefully it will be as up to date as possible - but this video is not for the WoW millions of subscriber base type game.
I wouldn't judge the all-around capability of an engine based purely on a video of its editor. Who knows how flexible their networking code could be. There are several examples of MMO's doing a major network code rewrite of a shoebox engine to good effect, but it does typically takes years and millions to do it.
Would use of grayscale maps like the DEM info for terrain also be able to be used for something like vegitation, roads, rivers, etc make open areas less of a chore? In the case of this engine and say a project like WW2ol the major pop centers would need manual construction and such but something like presets for minor centers like villages or farms could be constructed for something like drag and drop additions. Ofcourse a varied library of these presets would need to be made to make it interesting. Just my 2cents:/
Actually our goal is a completely procedural built terrain from data sets like DEM, modified road and river maps and dbase built population centers. We can then go in and hand place the strat markers that designate capture points and links.
Post a Comment
Links to this post:
Create a Link
<< Home