My most recent project, a browser-based game, sort of a cross between the excellent OGame, the also excellent Frontier: Elite II, and enormous second-job simulator EVE Online.

Quite a bit to do on this one, but so far it’s not a particularly tough project. Built a new generic PHP class that everything in the game inherits from. It handles all the database stuff, meaning making a new game object is as simple as making a class with the appropriate variables and an array of table links, and adding the table itself to the database.
Works great for this project, but it’s caused me a few headaches on some others, so I don’t think I’m quite as close to abstracting away the database stuff as I’d hoped I was.
The game itself places you in a spaceship. You can travel around and trade between planets. Eventually, you’ll be able to afford more ships, and pilots for them. Beyond that, you can buy plots of land on planets, build entire stations yourself, and research and trade technology and items with other players, or send your ships to attack them.
The planet side will have a nice square-based editor where you’ll have to arrange your structures yourself. Stations will work a similar way, but will require more funds to keep in operation, and be visit-able by other players if they are set as “public”. Anything you make will be tradeable here. Your planet side facilities will be able to produce resources, factories in stations or on planets will then be able to produce consumables that you can then sell to players through your stations.
Multiplayer games are an absolute bear to set up, since you need a critical mass of players before the game becomes a) fun and b) popular enough for players to spread the word. I’m aiming to de-emphasize the multiplayer aspect by adding in many AI NPCs, and a universe that expands automatically with the number of players. Ultimately, my aim is to make a game that isn’t solely about combat. It’s kind of an ambitious plan, which is why I’m not mad keen on pursuing it much further. The game really needs 80% of the elements in place before it starts to function as something genuinely new and interesting, and that’s a pretty massive feat. Add to that the difficulty involved in attracting and retaining a playerbase, and it becomes a full-time job.