Which is something I find interesting about 4e. So many people complained about 2e that "no one wants to play a cleric because they just stand around and cast cures". So 3e created the IMO, too versatile cleric. Now with 4e we swing the pendulum back a bit and make every class fill a "role".
And for the cleric, the intention was to create a type of class that can heal without having to stop contributing in other ways in order to do so. They succeeded to a degree, making healing something you can do in addition to attacking, for example.
A couple of the people I play with really love pushing the role system to breaking point. They take a character class that is one role, and choose feats, equipment, abilities, etc., to do their best to fill another. The PHB even had secondary roles already listed for lots of classes.
Again, I have limited experience with the system, so fill me in on how this works exactly. If I am a "striker" like say a ranger, can I have an interest in being a self sacrificial type who takes it on the nose for the party when I can? Or do I need to play say a fighter for that. I realize one will be more optimal for the situation than the other. But is it feasible for a player to decide they want to "Defend" but not play a "Defender" class? If not, how does that effect the players choice of interaction with the ongoing story? Curious more than anything.
You can do it, sort of. A defender usually has some sort of mechanic that punishes an enemy for attacking someone other than themselves (I believe the Pathfinder Cavalier does this as well). So unless you specifically take feats or powers that let you punish people for attacking your friends or allow you to take their blows for them, then it's not going to happen. You'd have to a) intentionally want to develop that secondary role and b) understand the system well enough to choose the right feats, powers and abilities.
This is a good example of where system mastery can limit 4E. It would be challenging to make a defender ranger if you were just a casual player.
But, perhaps it's best to ask what you're attempting to accomplish in making the Ranger a defender. Do you want a wilderness warrior who taps into some nature magic and fights in melee, protecting his friends? What if you considered the Warden class instead of Ranger? A couple two weapon fighting feats and maybe a high dex for using missile weapons and you could be pretty much there.
So yes, for my "railroading" I usually at least detail a fairly sophisticated world for players to explore first. At the very least, I outline a meta-plot of sorts that is running in the background as we play and the players occassionally bump against that meta plot
Why isn't the meta plot spontaneously formed out of the player actions? They might bump into situations that are unfolding, but why is it a pre-decided "meta-plot" and not simply a situation that not even you as the DM knows the results of yet because you don't know how the players will get involved with it?
I said that 4e IMO lends itself to a scripted adventure quite well. Again, if players can't move outside the box and they must confront the bad guys at the time and place the DM prefers then it at least feels more scripted IMO than if the players have the OPTION to teleport in a smack someone around whenever they please.
Yes. If someone wants to run 4E that way, the game will more than support them. Even the best of 4E published modules are terrible in this regard. I had heard such good things about Logan Bonner and Orcs of Stonefang Pass and then when I finally saw the thing, it was literally a straight line pass through the mountains full of orcs. With the players always arriving on scene at the dramatic moment to save the dwarves and learn about how to defeat an ancient evil. I integrated it into my campaign, heavily modified the dungeon to make it non-linear and had the situation unfold in response the the players. End result, all the dwarves died before they could arrive "just in time" and they stopped the ancient evil with a solution not present in the module.
4E absolutely supports an approach where the DM has a bunch of scenes they are going to frame prepared in advance no matter what the players do. But it works so much better when you don't do that.
When you set up situation and respond to it by framing scenes on the fly based on what happened. And the low prep DMing it allows heavily, heavily supports this. I think I posted about this in this thread where I talked about using the Story Now "bangs" technique from Sorcerer and how 4E supports that so well.
Obviously you are a very enlightened DM. You keep talking theory discussing game theory as proposed by people outside the 4e community to defend the constructs of 4e. Who's to say you aren't the DM Heinsoo refers to in his statement?
The difference is how well 4E supports the improv DM with it's low prep nature. In the example of 3.x, when a DM has a problem with "gotcha" spells, the system doesn't support resolving that issue very easily (or at least I never got it to work and neither did anyone I else I ever gamed with in a few different cities).
EDIT: Also, there's a factual error in the above quote. An author of DMG2 is also an author of Forge type games and developed some of the techniques in question. It's not just from people outside of the 4E community.
How anyone can keep saying you can't string these together with combat encounters into a good dungeon crawl is completely beyond me. Its made for it. Sounds great to me frankly.
Because that's a poor dungeon crawl. Going back to OD&D, good dungeon crawling was about players making meaningful decisions.
speak with Dead? An "I win DnD Spell?" or "Easy Button" spell? Really?
Seriously though, in practice as a DM I use that spell to develop useful riddles which players might be able to jump right to the killer but usually these tidbits just help confirm evidence as they move along. If you never read the description and just assumed "Ahhh this means free flowing two way communication with the deceased! I'm ruined!" then yes, youre screwed. 
