A DM should be knowledgable of most of the common rules issues and describe their fices well ahead of time.
The DM will not know every issue. DMs are human. But a DM should know their game, their players, and how their player's characters work. DMs are given the power and thus have the responsibility. They should inform the players of their changes and adjudications of common things before they happen.
For me, this is a huge ask. There are other RPGs that don't impose this level of detailed technical management on the GM, and I don't see it as a virtue of D&D that it does so (if it does).
clear rules don't have to take away GM initiative. A rule can simply say, "the GM may decide what this means." That's okay - it is both clear and open-ended. What is bad is when a rule is open-ended because it is phrased poorly. That leads to unnecessary conflict. If you don't want there to be a RAW ruling on something, just say so in the rules, don't get there by creating confusion.
If there is an ambiguous rule in the PHB, one of my players may look at it and think it obviously means X, while I look at it and think it obviously means Y. It's quite probable that neither of us will realize the disconnect until the player tries to apply that rule in the game. Then the game stops while the player and I sort things out. That is precious game time being wasted because the rule is unclear. Furthermore, it means now I have to remember what we settled on and apply it consistently.
If the rule is clearly written, and I don't like the way it's written, guess what? I'm the DM. I can change it. I don't need strategic ambiguity to give me permission. But now I know that I'm changing a rule, and that I'm going to need to tell my players about it. No confusion.
The ruleset should be compact, simple, and clean. Its scope should be limited; many things should be left to the DM's judgement (trying to eliminate any need for DM judgement was the great mistake of 3E and 4E). But it should be clear when something is being left to the DM's judgement. Within the limited scope of the rules, there should be as little ambiguity as possible.
I don't agree with Dausuul's characterisation of 4e - I think it leaves a lot to GM judgement - but otherwise I strongly agree with these posts.
A related issue that I have noticed in the 5e rules that I have read is a tendency to mix rules with a wargame-style precision (eg under these circumstances, you get +X AC, or you get advantage on attacks) with natural-language descriptors, which to me at least sends mixed messages: the conditional modifiers are important enough to specify to very high levels of precision, but not so important that we need precision on when they apply.
if my character prepares contagion expecting to use it on the big boss fight we know is coming up and suddenly the DM is all no that doesn't work right away, I am upset. Rules are important, and everyone knowing those rules are consistent is also important.
Part of the issue here is that every spell in D&D (except 4e) is its own little package of action resolution, combining natural language descriptors with wargame-style mathematical mechanics. And the designers, certainly compared to some other RPGs, give very minimal meta-commentary on what some of these spells are for. So is Contagion meant to be a "strategic"-level curse, or is it meant to be a tactical bombshell? It is almost inconceivable that the designers didn't have one or the other of these options in mind, but they tend to be somewhat coy about it.
I don't think this is a place where saying "it is up to the GM" is the appropriate response. They decided that spell X uses d6 and spell Y uses d8 for a reason; they didn't just tell the GM to choose between d4s and d12s as the mood takes him/her. If Contagion is then meant to operate in the same mechanical space as those other spells, it makes no sense not to give similarly detailed advice on how it works.
5E has all kinds of RAW
<snip examples>
5E is not rules light (with 1000 pages or so of "core rules") nor is it all open to interpretation. Some of it is open to interpretation. How hard is it to open that lock, or walk on that beam, or hide behind the tree? Even these are given guidance, but yes, the DM makes the call.
D&D is traditionally a game with many moving parts and many areas of potential conflict.
I like these posts, although I think that what Sadrik says is not true for all versions of D&D - for instance, classic D&D didn't have many moving parts at all for non-combat, non-social resolution.
When talking about rules-as-written, GM interpretation etc I think it is helpful to be clear, and realistic, about how the game plays, or at least how the designers envisage it playing. If they put in lots of little crunchy bits, presumably they expect those to be used. But what does
using them consist in?
D&D has always had at least moderately intricate PC build rules (including the equipment rules as part of the PC build rules), and players are expected to make choices in relation to these rules. So there is, I think, an implicit expectation that it matters if a player spends money for better armour for his/her PC, or chooses a longer or shorter weapon (in versions of the game with weapon length), or chooses a weapon with a higher or lower damage die (in versions of the game with variable weapon damage). A GM who just ignores those rules is undermining those aspects of the game. Everything else being equal , this is probably not the best way to play D&D. (There are other RPGs out there which don't have equipment rules, or don't place so much emphasis on other aspects of PC build like spell load-out.)
Where I think the game has changed over the years is in respect of action resolution. Action resolution, both inside combat and even moreso outside of combat, is a much bigger part of the game now than it was 40 years ago. More of a player's build resources are devoted to action resolution than used to be the case. (And, conversely, exploration is less a part of the game than it once was.) GMs are expected to provide consistent adjudication on player action declarations as a key function of the game. A good DMG will give GMs advice on how this can be done, what the different approaches are. For instance, here are 3 possibilities: the GM describes the fictional situation first, then imposes appropriate DCs (this is probably the 1st ed AD&D default); the GM decides what would make a good DC, and then describes the fictional situation in an appropriate way (this is probably the 4e default); the GM describes he fictional situation first and then adjudicates the d20 roll based on what is "good for the story" without worrying about any correspondence between the DC and the fiction (at least in my experience this was often the 2nd ed AD&D default).
How rules are handled varies across these (and other) approaches. Rather than insisting in the abstract on "RAW" or "rules as interpreted/applied", I think it is more helpful for the game advice to call out the different approaches, talk about some of their pros and cons, and what sorts of impact on play and player attitudes they might have. For instance, a GM who adopts the third approach above might be better suited to players who build their PCs to "express a story" about the character, rather than players who build their PCs to generate resources that they can then deploy to increase their chances of success in action resolution.
I trust my players not to powergame and search for every scrap of advantage they can eke out of the rules and I trust them not to worry that I'm out to screw them over when I interpret the rules in a particular way.
It's not an adversarial game...there are no "two sides" there, you're all playing the game together.
For a lot of players, D&D is obviously adversarial - just look at the current "death and consequences" thread, for instance. For a significant number of D&Ders (I can't say whether or not they are a majority, but they're clearly not a tiny minority), one of the most important jobs of the players is to deploy the resources they obtain from build plus "clever play" to overcome the challenges (often but not always NPCs and monsters) which the GM has thought up and is playing as the PCs' opponents.
In that sort of play, eking out every scrap of advantage is completely rational. If you read old play reports (say in Dragon magazine), it seems to me that this was exactly how Gygax and other earlier players approached the game. That's why, in Gygaxian D&D and AD&D, regulating what equipment and spells the players gained access to was such an important part of the GM's job (because these build elements were the main resources for players).
The complication for more modern versions of the game is that build elements have become more complex, and less obviously subject to GM control because no longer acquired primarily through PC actions in the fiction but rather as part of generic PC build rules (eg feats, skills, class features), but the game still gives no reason for many players not to try and eke out every bonus.
Hopefully the DMG will have a good discussion of this issue too, and the range of options. For instance, if the players throttle back will the GM compensate by fudging? This is one well-known approach, though obviously not one that appeals to everyone. Or will the GM play the opposition at full bore, letting the dice fall where they may, but use "fail-forward" techniques to reassure the players that more relaxed or even sub-optimal choices on their part don't have to be game-ending? In my view, a good book of GMing advice will cover these and other possible approaches, changing the focus from "RAW vs RAI" to "What do we want to get out of this game, and by way of what techniques, and what does that mean for our approach to the rules?"