I was just going to make this a comment, but I figured it got long enough for a post.
Looks like everyone's main concern is the lag. I'm not sure if that's actually from resource downloading. If it is the case though, I can remedy that a little bit by disallowing customization within game servers. You'd just have to do that on the "hub", which I think I'm going to make a little different from the last time. Crispy, I'd do it your way, except that pretty much all resources are generated at runtime. Sounds aren't, and they probably make up most of the file size. I could preload those seperately.
Dark Saber, I loved Bomberman 64, but BYOND is not well suited for that sort of 3d, at least not with the speed I'd want out of it.
As far as the reward system goes, I'm leaning towards a combination of the old and new systems. My beef with the old system is that players will do whatever they can to cheat it. Unlocks that are based on wins, I'd like to make unique, so if you lose your rank, you lose your unlock. That way people won't be so willing to allow others to kill them for unlocks.
Another idea I had was to tone down the potential speed of games in the interest of lag. When people spam bombs and run around with a shield on, it really kills the connection for other players. I'd kind of like to focus the gameplay more on strategically placing bombs and using interactive environments to your advantage.
In case anyone is wondering, I did start working on it this weekend. I just have the standard bomberman level layout, which is generated automatically. I'd like to make more procedurally generated maps, or at least the option for them, rather than all static maps loaded from files.
By the way, how does "Super Bomberman Arena" sound? ;)
ID:30567
May 14 2007, 3:32 am
|
|
May 14 2007, 1:59 pm
|
|
Sounds nifty D, what happened to your 3D Engine or Mavea?
|
Super Bomberman Arena, I like it!
Sounds like you've got things under control. I like your idea to make the game slower and more strategic in order to combat lag, that's good thinking. As for preloading - the sounds are a good start, but with some careful design I think you can go further. If you used overlays rather than icon procs for most of your customisation and pre-generated (at compile time, not runtime) the coloured versions of each icon segment (so pants and shirts would be different overlays), then you could probably cut down on the amount of resources needed... There were only a few colours available IIRC, so it wouldn't be that hard to do this. |
TNN: I've still got the 3d Engine on my flash drive, I move between projects alot. 4.0 almost makes me want to go back to BYOND for Maeva, but I can't do that now.
Crispy: I'm using IconMod for basically everything in the game. Each instance of an icon should only be loaded once though, since they are cached on the server, and are only really "generated" on the first instance. The players are made up of overlays, but those overlays are colored dynamically. If I pre-generated each possibility of colors, I'd have thousands of icons created that would probably never get used. Is a player connecting and contributing around 5 icons too much overhead? Also, does BYOND throw out those generated icons after everyone disconnects? |
Thousands? My memory is a little fuzzy, but I thought there were only about half a dozen colours available per body part.
It's hard to say really, but I would guess that 5 icons is probably bearable. Still, if you can avoid it... The impact on me (as a client) from another player connecting is not so much the direct impact of me having to download their icons; it's more the fact that the new player will be hogging all of the server's upstream bandwidth while they download all the game's resources; in this case, (number of players)*5 icons + environmental (non-player) icons + sounds + whatever else. Upstream bandwidth is normally quite limited on people's home connections anyway, and it doesn't help when you've got 2-3 people hogging all of it. So if it's too much trouble, just preloading all the non-dynamic resources (environmental icons, sounds, HTML help files, whatever else you can think of) from an external HTTP server will help a lot. Generated icons tend to stick around in the client's cache. Whether dynamic icons always get re-used from the cache correctly, I don't know - they should, in theory. |
Thousands? My memory is a little fuzzy, but I thought there were only about half a dozen colours available per body part. You're right, but they weren't actually available "per body part". Each icon has a potential for 8 different color regions. You might have noticed how the little ball on top of the helmet used the same color as the hands and feet. If I were to generate all the possibilities ahead of time, it gets into something more like 12 colors x about 5 regions x 5 potential icons x 10-20 variations of those icons, and I start running into trouble. Since lag has been the main concern of everyone, that'll be my main issue to tackle. The external resource downloading seems like it'll help a lot. Another thing I considered was locking games in progress, so that clients wouldn't join and start downloading resources that could be used for updates. Is there a way to tell if clients are in the process of downloading resources? |
OK - so why not make the ball on the helmet (for example) a "body part"?
I'm just saying, if you split up the icons enough then there must surely come a point where the possibilities start to "add" rather than multiply. Anyway, the external resource downloading will help the most and it's really easy to do, so check that out first. It may turn out to be enough. Is there a way to tell if clients are in the process of downloading resources? Not that I know of. |
make a second one coz i been waiting for a update or something to bomberman i mean...
the 360 has bomberman arena at 8 players allowed to play and on bomberman you can only have 5 and you can get a revenge revenge on the 360 but on this you cant |