Sure, anyone can climb a tree or hide in a closet. Staying unseen in a shadowy corner or scaling a brick building without hand protrusions is a different story. Likewise, anyone can wave a sharp piece of metal around and poke it at things but there’s a reason why magic users used daggers on a very different combat table than a sword-swinging fighter.
Eh, proficiency in those skills doesn’t cost you anything more than your background, which is a fairly low opportunity cost. And the proficiency bonus is so low relative to the die roll that there’s not actually that much practical difference between the classes, skill-wise. I play a rogue in my group, and so far, the paladin has done better than me at sneaking and picking locks, and the half-orc barbarian with charisma as a dump stat has done better than the paladin at social interactions, just by luck of the dice.
My feeling is that with a good DM/GM, it shouldn’t really matter what rules system you use. Players shouldn’t have to know anything, necessarily, about game mechanics. I even see some advantages in keeping game mechanics secret during play (i.e., the players know what the characters know).
One of my best campaign groups (2E) included several people who had never played any kind of RPG before, at all–but they were into fiction and film and theater, and could imagine and improvise without being limited by expectations about how D&D groups were supposed to approach situations.
I’m still a bit bitter that WotC killed 4th edition after I spent ~$600 on books and adventures.
With regards to bounded accuracy, I would have preferred a shift to the d100 to free up design space.
Even with a good GM the system used changes the game quite a lot and I’d say that’s a feature and not a bug. Doing a dungeon crawl with FATE would be quite different than doing it with D&D 5th Ed or GURPS. All three would easily be tons of fun but you have to know enough about the system that you know whether a crossbow bolt to your stomach is a minor inconvenience or a mortal injury.
Not to mention in any heavier system shoveling all the paperwork on the poor, overworked GM is going to slow any combat to crawl. Players should know enough to update their own sheets at least, and in more complex systems they should be able to figure out their own skill rolls as well. Of course if you are just doing short 1-3 session one-shots with light systems or pre-made characters it’s a different story.
I disagree - background is an extremely high opportunity cost, because you only ever get ONE.
Proficiency bonus is a victim of how D&D works - if you need to be able to model levels 1-20, you can’t make the bonus you get for being “good” at something at level 1 too large, or you risk making level 20 completely laughable. This has been a problem literally since forever. “I am an awesome swordsman! I have a +5 to hit!” “Yeah, and I suck. But I rolled a 19 and you rolled a 3, so screw you.” This is just how D&D is, fundamentally, because of its “need” to model zero to hero.
How did the paladin even pick locks? Did he spend his background on being a rogue to get proficiency with thieves tools? Because otherwise, he’s not even allowed to roll for that.
Let’s spin the issue around: All it takes to become proficient in a longsword is a background, and the proficiency bonus is small. Are fighters also watered down?
This… is true and false. It’s true that the players shouldn’t need to know the rules to play, necessarily, but to imply that that means that what the rules are doesn’t matter is a completely faulty conclusion. Even at the most bland level of doing a dungeon crawl in a D&D game, a party of players playing D&D4 without seeing any of the rules will have a different experience from a party of players playing basic D&D. I mean, hell, the wizard has like 5x as many HP, and that’s going to produce a different play experience. So no, the rules matter whether the players know them or not.
Which has almost nothing to do with whether rules matter or not. 
Are they all using the same to-hit tables these days? If so, yeah (albeit to a lesser extent than rogues since many classes use swords but only thieves had thief skills). Which is what I was saying about classes being homogenized and less unique or special these days.
Back in the day, I believe all elves were nominally proficient in the longsword but, if you were a longsword wielding elf magic-user, your attack rolls were still abysmal so it mattered a lot less. A sword-swinging elf magic user was pretty much like a party member saying “I hide behind the statue when the goblins enter” – technically possible but not really infringing on class roles.
To-hit tables are gone, and while the proficiency bonus does scale at the same rate it’s not as important as class choices.
Having a proficiency in longsword isn’t great unless you also belong to a class that gets multiple attacks. A wizard with a longsword proficiency will get a single attack a round and the results will never be as good as any of his at-will spells. Yes, he’ll have the same proficiency bonus as a fighter, but even at 1st level the fighter will out-sword him and that disparity will grow every level.
Here’s an important difference: in earlier editions, saying “I want to be a wizard who can handle a sword” usually meant really big opportunity costs and without a focused expenditure of resources as you leveled, you would still absolutely suck.
In 5e, it means “pick a background with a martial weapon and be a wizard with a sword.” You’re still not going to be great at it, but you’ll always be able to apply your scaling proficiency bonus. You’ll never be great, but you’ll never be laughable.
In my opinion, 5e has done a great job of giving lots of latitude in terms of creating character concepts without bloating the creation process.
I was addressing the question of the fine choice of game mechanics being significant to novice players, not saying it made literally no difference in the course of events.
OK, that is a big difference. I would never want to play something in the former category, where commonsense understandings of the real world and traditional stories aren’t generally applicable.
“Watering down fighters” isn’t a great comparison anyway since fighters have traditionally struggled for class identity. Forget longswords, everyone is able to swing a sword, mace, staff or something that clonk the bad guys so one more guy with a sword is pretty trivial. Add in fighter off-shoot classes like paladins, rangers, barbarians and cavaliers and you struggle to find a good reason to be a fighter short of just wanting to play a simple guy. Making thief skills universal is taking away a class defining reason to play a thief/rogue but fighters never really had a universal skill to take and share. I don’t know if that’s been made better in 5e but I remember a million Dragon magazine articles about how to punch up fighters and give them a role that isn’t more than adequately filled by a half-dozen other classes.
Sorry, that wasn’t clear.
What about a crossbow bolt in the shoulder? Y’know, a designated “action movie hero hit zone”? There’s still a huge difference between “Yeah, you’re going to go into shock and die” realism and “It’s just a flesh wound, carry on.”
Anyway, the rogue thing is still false. You know why? Because if you’re going to pick someone to go sneak in somewhere, you’re going to pick the guy who:
Has a high dexterity bonus
Has proficiency in the skill
Is wearing armor that doesn’t make a ton of noise
Doesn’t have so few HP that he’s all but guaranteed to die if discovered
Only one class actually fits that niche.
It doesn’t matter if other people are “okay” at something, because the rogue is going to be better, and this is a niche that only one person needs to be good at. (Except if the entire party is trying to sneak in somewhere, in which case, god help the paladin. :P)
Edit: Also, I feel obliged to point out that the rogue hasn’t had “hiding, sneaking, and disarming traps” as his primary schtick since like 3rd edition, because fundamentally, regardless of whether other classes can easily get those skills, those skills are actually really terrible most of the time. Rogues are also “glass cannon” damage dealers in 5e.
Very old gamer axium: “Paper soldiers never die”
Games don’t generally make ‘realistic’ allowances for wounds, because you’d get hit once, cry like a baby and run away. ‘Paper people’ (PCs, NPCs, Monsters, Soldiers, et al) die at rates that would make the strongest and most cohesive army units break and rout. They don’t run for cover like you and I would under fire, they don’t fight long slow battles of throwing shit from behind that cover, because none of that would be particularly fun. They don’t get hurt, suffer a debilitating effect and lie wounded on the battlefield waiting for it to end, where they either get medical help or finished off.
Would you want to play Call of Duty where your character gets hit once in the leg and is completely out of action but still in danger? Would you want to play D&D where your Wizard takes an arrow to the knee and flees in self-preservation?
If by “only one class” you mean “all of them”. There’s definitely a party role for “guy who sneaks into places”, but what you’re saying is that you fill that role by having high Dex, proficiency in stealth, light armor, and a decent number of HP. None of that has anything to do with class. Or, rather, the proficiency has something to do with class, in that a class is one way to get it, but it’s far from the only way, and plenty of other classes than rogue can get proficiency. And I guess that the decent number of HP is class-related, too, except that most classes get as many HP as the rogue or more: Only the sorcerer and some wizards have less.
What do you mean by “making level 20 completely laughable”? Is it laughable for a high-level fighter to be able to slip his sword between the chinks in magical platemail 95% of the time? I don’t think so: I think that’s just reasonable for what you’d expect of the best damned swordsman in history (which is pretty much what a level 20 fighter is).
Which is what D&D damage modeling is all about. Depending a bit on version and your level, D&D characters can swim in lava (briefly) or hit the ground after falling from 1000 ft up or be hit by 20 arrows and not die. 6-18 seconds with a same level cleric (or whatever the combat round length is these days) and they are again at full health and can do it again.
My ranger? 
They put out more than 50 full-sized books and quite a few digital issues of both Dungeon and Dragon, each of which was stuffed with content. There is a ton of material out there.
If you love 4e, keep playing 4e. There are grognards out there who never stopped playing 1st edition.
Yeah, not in our games.
Actually, very few classes have all those things. Ranger is a possibility, if you go out of the way for it, but you’ll still be missing several of the rogue abilities.
And again - this isn’t even the rogue niche, because opening locks and finding traps is sucky and awkward, because either they are required, or irrelevant. So these are basically things that are only put in the game because there’s a class that can deal with them. Farming this stuff out actually makes the game better. It also requires rogues to get other stuff. Which they do.
That’s not what I mean at all, but it does sound awful, because it means that fighter would die instantly in a fight, because he’d be facing opponents that could also do the equivalent of that.
The problem is that the bonus spread becomes bigger and bigger with every level, and if you start out with that spread being too large, by the time you get to level 20, you end up with characters who are basically yes or no switches - there are some things that they can do basically all the time, and they are basically incapable of doing anything else at the level that would be called for to be useful.
Then…you’re not really playing D&D. I’m sure you enjoy your homegrown D&D-based game system, but it’s not the same game.
Rogues aren’t just locks and traps, no. They’re also, for instance, the advance scouts, who go check out the situation before everyone else comes up to deal with it. Nor are locks and traps a binary obstacle, like you seem to think: Without a rogue, you can still deal with them, you’re just louder on the former and spend more HP on the latter.
And how does a ranger (or monk, or bard) have to go any further out of their way to fill the rogue role than the rogue does? And what, exactly, do they miss out on? Rogues don’t get any unique ability for using skills until level 11, which many campaigns never even reach. And even then, the shadow monk or the bard are still stealthier than the rogue, thanks to Pass without Trace.
I find it bizarre to think that your DM wouldn’t have his goblins lock their doors/chests or set traps unless the party has a rogue in it. To say nothing of official modules that already have locks locked and traps set.
Whether or not the classes are homogenized and some noticeably watered down and whether or not this is a benefit to the game are two separate arguments.