it has long been the case that the DM is expected to house rule the system, to make changes, and to making rulings - some of which may be contrary to RAW.
I think that LurkAway gets this right - it's not about making rulings that break the rules, it's about the rules being written in such a way that contextually-appropriate rulings from the GM are expected, and count as an implimentation of, rather than a departure from, the rules.I think the essential question could be reframed like so...
Which of the following is true:
1) a rule describes what can/will happen in-game (regardless of context)
2) a rule describes what is likely to happen in-game (in an average context)
<snip>
With rpgs, it's often #1 (definitely with tactical play and rules lawyers) but others often prefer #2 and view the rules as a guideline. It also varies depending on the rule, of course.
Monte gave the example of a rule that: open door = move action. Let's say that there's a heavy stone door and the DM wants that it takes a full round to open.
If the group subscribes to view #1, then the DM needs to break the rule at the risk of arguments ("Yeah, yeah, I know the rule says 1 move action to open a door, but in this case, I rule that it takes 1 round"), or the system needs an extra rule at the risk of increased complexity ("exception: stone doors take a full round to open").
However, if the group subscribes to view #2, then there is no need to break the rule or add an extra rule. The average door is likely to take 1 move action to open, but this heavy stone door is not average. The important thing is to try to be consistent -- that's why I would consider this to the mature/advanced version of D&D.
The other thing that is needed is advice to GMs on how to communicate rulings to the players, in the course of play. For example, if a door is going to take a standard action to open because of its size/weight, and this should be obvious by simple inspection (eg it's a massive stone door), then the GM should be advised to communicate this early on in the piece - in particular, well before a player gets hosed in combat because s/he assumed that the door provided an escape option that would require only a minor action to activtate, only to be told by the GM that a standard action is what's required.
What is also needed, in my view, is advice on what the appropriate stakes should be for trying to (for example) reduce the action cost to open a massive stone door. Page 42, and the improvisation suggestions in the Essentials skill entries, give a few ideas about this, but there's plenty of scope for more detailed and wide-ranging advice for both GMs and players.