(This is meant to be a funny thread in reference to a funny fictional feature of the late AGS 2.x)
So at the moment I'm really, really studying the underlying construction principles of point-n-click games. I mean I'm going beyond writing roadmaps with the number of rooms, the connections between rooms, the number of objects and that sort of standard planning.
I've extracted some very abstract core principles that end up being pretty much a bunch of lists. These are lists of entities types : types of puzzles, types of areas in a game with several locations, types of obstacles, etc.
I've reached the point where I'm thinking "If I had too much time on my hands I could create an adventure game generator". Of course I'm saying that as a joke, as the task would be HUGE, programming wise. But it's the first time that I'm thinking it could be done with a sufficient level of quality (the player wouldn't notice). With the help of those lists and metrics, the game could be randomized and generated by a computer, and still be fun -- just the same way a computer can already build a (good) randomized dungeon and looting system in an action-RPG, or just the same way some games have simple randomized quests, as in the old "desktop adventures" series from the 90's or many games today.
Imagine the face of the wannabe point-n'-click creator if he could actually create a full-length point-n-click game by just clicking a button! Then of course he/she would still need to apply a "skin" to it (well-written dialogs and nice graphics), but the underlying game would be there, would be entertaining, and would be solid (well-balanced in difficulty, clues-dropping, no dead ends, and all that).
Also from a more serious perspective, if I was making adventure games for a living I would seriously consider implementing that tool, for quality control of full-length adventure games. The tool could in turn produce the actual roadmaps (meant for developers and artists) like the ones that are normally hand-made by the project manager or the game designer. That would probably save one to two months of planning and design at the beginning of every new game.