FireLance
Legend
I think the article addresses a couple of issues that I happened to touch on in the "Examination of Differences between Editions" thread. One is the issue of player interface, and the other is the issue of what was referred to in the thread as "low-detail propositions" (e.g. "I make a Search check") and "high-detail propositions" (e.g. "I search behind the painting"). I'll just quote the relevant parts.
Sometimes, I think one of the major determinants of how smoothly a game is run is how well the players interface with the game. In more open-ended systems, like previous editions of D&D, the DM has more responsibility to be that interface and to adjudicate the players' interactions with the game world. If the DM is flexible, creative, and able to adapt the challenges on the fly to create hints or add new complications, as necessary, the players will be challenged, but will not feel frustrated. However, if the DM has only one solution in mind, or is otherwise rigid and inflexible, the players might feel that they are playing one of those text-based RPGs where you have to type in exactly the right commands, e.g. "Move painting" instead of "Search painting", "Use knife" instead of "Cut rope", etc.
More codified systems like 3e solve this problem by defining more clearly what the PCs are capable of doing. Effectively, they allow the players to interface with the game in a command-driven or menu-driven manner. By narrowing the universe of choices down to a limited number of ways in which the players can interact with the game, and implying that the solutions to problems should be defined in terms of these limited interactions, at least one source of frustration (the game interface/DM not recognizing the player's input as valid) should be eliminated.
The irony is that one of the main advantages that a table-top pen-and-paper RPG has over a computer RPG is the flexibility of the DM to accept all kinds of input. However, if I'm playing with a DM that insists on acting like a computer, I'd rather have a command-driven interface than have to guess exactly what input he will recognize as valid.
(On trying to encourage players to step out of their menu-driven or command-driven mindsets.) I think the solution to this is for the DM to accept both what you refer to as low-detail and high-detail propositions, and for successful low-detail propositions to point towards high-detail propositions.
For example, you could set your games up so that a Search check is simply a visual inspection. However, a successful Search check will uncover additional clues or hint at a course of action that the PCs could take to discover more.
Using the painting example, on a DC 20 Search check (a low-detail proposition), you could tell the PCs that there seems to be something behind it, or that it is slightly tilted, or that there are scuff marks on the wall that indicate that it had been moved repeatedly. The PCs are thus at liberty to manipulate it further (possibly setting off a trap or hazard), check it for traps (another Search check), or leave it alone. However, if a PC specifically stated that he was moving the painting (a high-detail proposition), he would find the hidden object (or set off the trap) automatically.
Similarly, for the example of the cluttered room, or the trapdoor hidden behind straw, a PC with a high Search check should realize that there might be some things that are still hidden because the straw or the clutter is in the way.