I think it's interesting that a lot of this stuff remains true for spellcasters in 5E.
<snip>
I think the move away from player empowerment to DM empowerment has disproportionately effected (or continues to disproportionately effect I guess, your call of your view of non martial and martial empowerment in 2/3/4) martial classes over spell-casters.
This is not true. Using thieves tools is not "narrative permission". It's a set DC. If you roll high enough, you get it. There is no DM permission needed.
<snip>
And if you say the DM can just arbitrarily raise the DC, the DM can also arbitrarily say that your knock spell doesn't work.
I don't think there's a passage in the 5e books that says that, for a spell to take effect, the GM must agree that it does. Whereas there is this passage for checks/skills (SRD p 77):
The GM calls for an ability check when a character or monster attempts an action (other than an attack) that has a chance of failure. When the outcome is uncertain, the dice determine the results.
I think that is a significant difference.
skills and abillity checks are left open ended specifically to allow players and DMs to roleplay a variety of characters. If a player wants to play a wizard with the soldier background, it makes sense that a wizard's basic combat training would differ than a standard grunt's. It certainly doesn't make sense to have a war wizard attempt to compete in a contest of strength, so they train them to use their opponent's strength against them. The player should reasonably expect to be able to make some trip attempts using dexterity, provided he isn't purposely trying to exploit the open-ended-ness of the system.
But there's no reason in principle why spells can't be handled this way too. There are even RPGs that actually take such an approach.
And the converse, too. Non-magical utility mechanics can be designed in a "player gets to declare success" way.
So what about a lock in an anti-magic zone? or a magically resistant lock? or even a lock in a silenced room? and so on.
I think that this does not speak to [MENTION=11831]The_Furious_Puffin[/MENTION]'s point. Yes, the GM can frame the fiction so as to make certain action declarations ineffective. This can happen to spells and to skills (eg the GM can declare that the PC's tools all fell out of his/her backpack back when s/he was crossing the river, and that the lock can't just be picked using one's bare hands).
The issue is - if the fiction is apt for making the action declaration, do the mechanics interpose GM discretion/decision-making between declaration and success? When it comes to most versions of D&D, the answer is "yes" for skills and "no" for utility spells.
The DM in 5e, though, can always narrate success or failure without calling for a roll, at all. So the rogue steps up to pick the lock, and the DM describes the lock clicking into place and the door opening - or not. It's less explicitly spelled out, but the same /does/ apply to spells. The DM can rule that any given portal that may seem ripe for Knocking is, for whatever reason, unaffected by the spell.
I wonder how common the second thing is. My feeling - based on a mixture of instinct and reading forum posts - is that it's not very common.
In 1e Magic-Users had to find/buy/transcribe/research all of 'em beyond their (random or DM-selected) starting spellbook
Not quite. DMG, p 39: "Naturally, magic-user player characters will do their utmost to acquire books of spells and scrolls in order to complete their own spell books. To those acquired, the magic-user will add 1 (and ONLY 1) spell when he or she actually gains an experience level". The book doesn't say who chooses; there is then the "% chance to know spells" rule.
Reading the fighter, and noting the fact that they were proficient in the weapon, it was not clear to them that actually using that bow would actually "suck". The only way to know this would be to know that a standard attack from a proficient fighter, actually is relatively weak within the 4e system. Plus, they would have to know that there were not a whole lot of options (or any) at higher levels that would make the bow a viable option. A higher level of system knowledge was required than in 5e and the 4e version was less intuitive and clear. Take your example of STR weapon vs DEX weapon, 5e makes it clear to make the decision about STR and DEX depending on the type of weapons you plan to use, no similarity really.
What about this comparison: a player builds a 5e fighter who uses STR weapons, but chooses leather armour rather than a heavier armour.
Anyone who looks at the rules for AC, the rules for encumbrance, and knows how attacks work in 5e, will see that Hide or a Chain Shirt are almost strictly superior to Leather or Studded Leather for a character with DEX 15 or less. But it's hard to work out how much better they are than heavy armours which give better ACs without the system mastery to understand how easy it might be to counteract - or just put up with - the Stealth penalty. And unless you read your class rules, how do you know that you won't get something to buff your AC wearing Leather (say, like 1st ed AD&D barbarians).
Similarly, anyone who knows the 4e rules for attacks will know that a longbow isn't that good without DEX; and anyone who looks through the fighter class abilities ("exploits") will quickly see that none of them buff bow attacks, and that nearly all of them require being in melee in order to be brought into play.
I don't think the relative lack of power 4e fighters have with bows is all that obscure; just as I don't think the relative inferiority of Leather armour for 5e STR fighters is all that obscure, even though they're proficient with it. And in either case, play will quickly clear things up: the 5e player will work out that there is no benefit to wearing Leather rather than Hide at least; and the 4e player will work out that a bow is only useful if your opponent is too far away to make closing, or throwing STR-based weapons, viable.