Imagine a similar top-down island+water view as in my previous sketch. I’m too lazy to make another.
This game is more of a resource management affair, with few action elements. All the islands are fairly small and circular in shape. You start out on one of them with a small land vehicle which you can move around. From this vehicle, you can send a probing ray aimed at the mouse cursor, used to measure range. The intended purpose is for you to drive near the edge of the island and measure how far away the next coastline is, and if it would be viable to build a bridge over there. You have a supply of resources for this purpose, like fuel, metal, and wood. There could be different kinds of bridges to suit different situations of resource supply, like wood/metal/concrete bridges.
Anyhoo, if the range checks out, you would start bridge building, which would be automatic and animated in some fashion. When the bridge is complete, you can cross it to visit the other island and harvest its native resources. One alternative system for resource handling would be that you had a swarm of builder agents that would scurry between the bridge being built, and back to connected islands with resource supplies. They would gather material and bring it back to the build site. This way there would be no global material reserve, but it seems like it could turn out kind of complex and tedious to have build time and efficiency depend on the network of bridges, possibly having the drones move all across the map to get things done.
For deciding where you want to build the next bridge, you could detach a helicopter module from the main vehicle and fly over to other islands to survey their contents. Small icons and proportional numbers (like “5 Oil” and “3 Rock”) would pop up when you get near them, or when you land. The helicopter would consume fuel, and have a limited tank capacity. That could mean you end up stranded on an island, and you would have to build bridges to reach and salvage it before it can be used again.
I’m not really sure what the overall goal would be, and if the game should either be win/lose or just sandboxy. Sandbox would probably be the best option, with some tweaking to make it apparent that you’re doing well or not, depending on how you play. One idea that could shake things up (pun!) is to have islands randomly rumble and crash into the sea, taking connected bridges with them. This would mean that you’d have to build new and longer bridges to reach some areas in case a key island was lost.
It might be possible to layer a different and primary mechanic on top of all this, where you perhaps shuttle creatures or other artifacts between islands in some way, so there’s something to aim for and always something to do (they want to go from one island to another, then back again or move on somewhere else). In that case, bridge building would become secondary but still the main important task you perform while playing.
That’s all there is, and you can’t play it. I would probably have started prorotyping one or both of these if I hadn’t made the questionable decision to spend my weekend doing really boring Sculptris code. I didn’t feel like I could relax and start on something else before that was done though, and in hindsight I’m glad that I have the boring stuff taken care of so I won’t have to do it tomorrow.
As always, good luck to those of you making games, and I look forward to browsing the entry masses later. One perk of not participating is that there’s no pressure to vote, for once 
Hmm… egg islands… hmm…
but I’m working on it!
The three islands! the red dots represent the landing area(?) to the island, clicking them takes you there, the skulls just “show” where the pirate caves are, the circles turn green and a red NO symbol appears on the skulls when you clear that island’s dungeon.