[Ryan_Todd]

Originally Posted by gblake

Have an arena, make it out of cardboard boxes and chalk outlines. Have robots (RC cars, whatever). Have batteries. Have simulated user-activity bursts and droughts. Have real people be a simulated mix of clueless, knuckle-headed, mischievious, chatterbox, whiner, apathetic, know-it-all, complaining, trash-talking, trolling, non-English-speaking, etc. users. Have to seamlessly roll out new software into your robots and other computers during the week. Have members of your team all want to be elsewhere on two days to Christmas and Thanksgiving. Have one or more of your simulated users constantly posting negative comments on the Internet about every mistake you make, and making up negative comments just to be a jerk(s). Have one field gameplay and/or field-reset mechanism malfunction at random intervals.

This.


While I'm at it, here's a few suggestions from the mechanical side:

• can simulate the ability to look around by replacing all cameras with fisheye lenses, and processing the image accordingly in the software
  • client-side zoom and direction controls mean that multiple people can look in different directions from the same camera simultaneously
  • no additional hardware complexity, so no more things to break
  • this would work great with VR headsets, BTW
• add a "commander" bot
  • stationary turret
    • mechanical option: 1 or 2 servo motors --> can look around, but that's it
    • software option: see fisheye suggestion above
  • also able to pull up teammates' video feeds
  • a few possible special abilities:
    • EMP
    • air strike
    • supply drops
    • be able to pilot a "UAV" for a limited time to get a better view
      • could be quadcopter, but doesn't have to be
      • could be simulated with a gantry, skycam, or boom arm
  • special user interface: full console, multiple "screens"
• spectators can choose what they want to see
  • one or more stationary view(s) available
  • mobile overhead camera
    • gantry/skycam/boom arm is probably better than quadcopter for this application; no batteries to worry about!
    • movement is slow and graceful
      • possibly based on average input from all spectators that are watching this feed
        • with the fish-eye camera suggestion, both zoom and direction can be client-side (avoids "bad manners")
      • alternatively, the mobile camera could be controlled by the field master for that arena; that'll at least give them something to keep them busy
    • for simplicity, this same system could be temporarily hijacked every now and then to provide commanders with their "UAV" functionality
      • during this time, the commander would have complete control --> movement should change to be much more quick and responsive (within reason, of course)
  • direct feeds from each bot
• give each player a heads-up display
  • game stats (score, time left)
  • personal stats (score, health)
  • minimap (with icons for self and teammates)
• bots with omni drive would be useful for special effect purposes, even if the user doesn't have omnidirectional control
  • think of an "explosion" pushing you away, or of sliding out-of-control down a "slippery" surface
  • unlocking omnidirectional control would be a great incentive for competitive play: really cool to have, and helps a bit during gameplay once you get used to it, but doesn't automatically mean that you're going to always beat everyone

I would also love to see this game incorporate augmented reality. This would open up a whole host of possibilities, including some that *could* be done IRL but are more convenient, and also some that can *only* be done with the help of AR. Here's what I could come up with off the top of my head:

• see a floating flag icon above the bot that's carrying the flag
• watch a "laser beam" shoot across the arena, and see a puff of "smoke" emanating from the impact point
  • same goes for the "nuke" from the preview video (looks more like an EMP, based on what it did to the opponent)
• "sparks" arc over the surface of robots that are temporarily disabled
  • the affected player should get some kind of static effects overlaid on their camera feed as well
• project a virtual "barrier" that prevents bots from passing through
  • barriers could be server-controlled as part of specific game modes
    • for example, to block the view of the other side of the field and give each alliance enough time to plan and set up an ambush before the match starts
  • alternatively, smaller barriers can be given as power-ups
    • opponents can "ram" these barriers repeatedly (or hit them with some other kind of weapon) to damage them and bust through
• objectives appear as tall glowing beacons
  • commander-set objectives can be invisible to the opposing alliance
• deploy an "oil slick" that causes bots to "slide" down ramps and interferes with maneuvering
• deploy "glue" that dramatically slows down any wheels that are touching it
• "flood" part of the stage, so that bots under water move sluggishly (low top speed and limited acceleration), but are hard to spot from above the water
• permanent or temporary "safe zones" in certain game modes, surrounded by a translucent "barrier"
  • could be team-specific (blocks opposing bots)
  • could also be general (all bots can enter/exit, but weapon effects are blocked)
• "x-ray vision" that allows a player to "see" other bots behind walls
• new game types could easily be added in software without the need for new game pieces and additional field reset work
  • real soccer with real robots, but a virtual soccer ball? sure! get yourself a reasonable physics engine and figure out how to drive the bots' wheels to convincingly simulate pinning the ball against a surface, and you're golden.

Thoughts about implementation (with respect to the AR idea):

• not at all easy, but soooo worth it
• requires the server to have accurate real-time knowledge of each robot's position in the real world, as well as a detailed understanding of the layout of the arena
• two options for rendering AR resources:
  • relay all pertinent info to the user client (including all players' positions & orientations), and rely on the end users' computers to render and clip all SFX appropriately
    • hardware-dependent: some users wouldn't be able to enjoy the game as much, or at all
    • more data transferred --> more bandwidth required, both on your side and the client side
    • all changes require client updates --> version control is important
  • beef up those servers, and do all the rendering in-house for each camera feed
    • upfront co$t for hardware
    • this game is uniquely well suited for this approach, due to the limited number of camera feeds
    • fewer changes require client updates --> version control less important

------------------
EDIT:

Wow, that's one heck of a wall of text. Hope it didn't ramble on too much for you!