Not to barge in on what seems to be a sparkling interplay between my peers... but... uh... did Pingu even bother posting here again?
Do most of these posts even have anything to do with his request?
I mean, let's be real here:
Really nice maps do not have to be huge to be pretty. If they're huge and pretty, they're going to be expensive. This is just a fact. Don't spaz, go on at length about how Map X which is mainly concrete and roads (good tiling surfaces, duh) is 3 MB whereas Map Y which is the same size but features beautifully lifelike terrain of an Earthlike forest during the fall is 45 MB... it's a silly comparision. Those of us willing to take the time to download something that big will pay the price to see the beauty. It's not like, say, playing Counter Strike on Steam, where everybody
must update their clients, and the updates include some really huge (if beautiful) maps. Spring is about choice- and if you don't like huge maps that take forever to download... then don't download them... and don't whine about it, unless you're ready to step in and make maps that are beautiful but small in filesize.
The above is pretty much a law of design in 3D game engines- while yes, we can see more optimization, there is an inherent tradeoff between beauty and performance. People whining about it... should just stop, because it's pointless, and it just shows that you need to upgrade your hardware. Don't like that? Tough... go back to OTA.
I can't afford the latest super-rigs out there (my specs are currently below the midrange) but I haven't had serious performance problems related to graphics on most maps. Most of the FPS load has been related to AI, pathing problems, and other things, which ultimately slow down my rig... when I have a few thousand things going on at once (shots flying, sounds playing, particles spawning, etc.).
But, if I was playing a mod that was smaller in overall scale (i.e., fewer units) I'd have zero problems. I will be building a mod fairly soon (after my current mini-project is completed to my satisfaction) that will feature a game design where less is more, and it will run very nicely on anything but the slowest systems. Also, keep in mind that, as uv-mapped models using DDS come into play, we should actually see some performance gains, if these models are done correctly, because the Spring engine won't be wasting GPU cycles on double-rendering polygons (and culling them) and other things that are serious perfomance drains. When Spring models are all closed-surface, tri-only and feature mipmaps (and, perhaps someday, LODs) we will see really nice performance compared to today's stuff... with the same hardware. So... what we're seeing here is baby steps towards pro-quality performance.
And if you're not a modder, AI designer or core programmer... be patient- whining does NOT HELP, and it effectively makes you sound like llamas. It's not like we game designer-types don't take performance into account, ok? We're smart enough to build the games you play- we're also smart enough to aim them at an audience's hardware. Current giant mods like AA and GA aren't necessarily good mods for people with low-end systems... but as we move forwards, this stuff will get taken into account. Right now, we're looking at kludge conversions... give it time.
It takes more than a couple of spare hours to make uvmapped models, script them, and incorporate them into a game design. And, for those folks (like me) who have zero interest in porting over OTA mods to Spring, it's just going to take awhile to get game designs clean and ready for you. Be patient- things are still evolving, and they will get better and better. A year from now, I suspect, people's ideas on what high-performance means will be very, very different. If you doubt me... go open up NanoBlobs... and see how effectively I've gutted XTA (which is what I started with) of all of the junk code and other things that weren't needed. When everybody's mods start to look this clean on the inside, then Spring games will really start to work better... that and model/texture optimization will be a big step towards better graphics (Pingu's original complaint) and better performance at less download cost for what we get (Gnome's point).
Will this mean that maps get much smaller for much greater beauty?
In some cases, yes! I have some very specific (and already workable) ideas on this topic, but I'm waiting to get a rough demo of this working first. But I will tell you right off the bat- I think that in the future, AI developers and modelers/scriptors will be playing a much greater role in terms of how we view map design. Right now, it's in the hands of Photoshop gurus, which means that we're getting pretty pictures. This is great, and I am glad we have some great talents here, but we can have so much more... in smaller packages.
In the future, I predict, it will be a much more mixed picture. I'm not getting into this too much yet, as many of you don't know me at all yet, and will assume I'm spouting bull... but I'm not, and you will see what I mean as time goes on.
So, to sum up... in a year, this "discussion" will seem pretty silly. There will be mods, for example, that are aimed at people with low-end machines. I will be building one, for sure, and maybe others will, too. They may not be XTA with zillions of units, but they'll be real alternatives that will run well even on lower-end machines.
There will be mods that are aimed at the latest-greatest uber-boxes... and there will be a lot've middle ground. 3DO will get replaced by S30 by modders who really care about performance, because it will ultimately look better... AND play better in Spring. And maps will be both more automated, more random, and much less dependent on the Photoshop guys to make them cool... which will mean, among other things, that smaller file sizes can still result in great beauty and variety.
If any of you want to help with all of this... help the community learn better techniques, help mod projects convert to S3O, and pay attention to the stuff that's going to be coming down the pike... Spring is going to rule when it all comes together
