They're not the same game. While that seems obvious, many people take the lessons and assumptions from 3.5 and try to port them directly to 4e is a grave mistake. The best example of this are the early monsters which try to use player-available weapons and other player-assumptions to inform monster design. This has led to horrible monsters in early adventures.
It's a complex and digitally enabled game. Characters have remarkable and intimidating complexities in them. While it is certainly possible for an experienced player to make her character by hand, the digital tools out there allow for a much smoother experience. However, players tend to forget things that don't appear on power cards, so that while a digital character sheet is an excellent choice to start with, players should hand craft their own sheets, their own cards, and their own checklists. I personally recommend power2ool.com for power cards, and google docs for making a checklist. (Rationale behind the checklist: powers are hideously complex. No really, people have 3 actions a turn and only 3% of powers analyzed above only do damage. By working through the tactics ahead of time, players can actually engage in combat without trying to figure out their options as perfectly spherical adventurers).
Loot is part of character, not a reward per se. 4e features a ... pretty decent simulacra of advancement through bigger numbers without... actually changing any fundamental relationships between accuracy or damage. There are pretty useful guides (random or list, depending on which DMG you're using) in the DMGs for loot to drop, and players should generate wishlists, or you should use intrinsic bonuses. But don't monkey with the math without thinking about the implications of that. More difficult encounters should come from the environment, higher level monsters... not depriving the characters.
Don't simulate. 4e is a game of combat-as-sport. Players don't have scry & die, save-or-lose, or any of the other... quirks of 3.5's combat-as-war. This presents something of a problem when players spend too much time planning ahead. Figure out your general policy when your players choose to engage in combat-as-war and tell your players about it. If you're OK with players facerolling an encounter because of prior planning, make sure they know that you won't negate their successes. If you're not, make it part of the social contract and make sure your players won't be adversely impacted by not planning.
The party is the fundamental unit of interaction. Make sure that your players have a functioning party. The difficulty of the game is determined by how well the party works together. A good party will triumph over remarkable challenges, a bad party will have trouble with equal level challenges. Tailor the challenges to the party and allow them to plan and communicate and roleplay.
The quality of RP is a function of the DM. If you don't reward roleplaying, you won't get any. Combats can certainly be taken in a boardgamey feel or in an RP feel depending on how the DM guides them. Verbal rewards ("Cool!") are fine as is verbal description (but reduce the number of enemies accordingly, or their HP, to compensate for the increased time that combats will take.) I prefer both verbal acknowledgement and to give out a floating "+2 bonus" that they can use in the same encounter.
Skill challenges present interesting challenge to the DM Scenes are no longer resolved by "roll diplomacy" ... instead it can sometimes turn into "everyone in the party roll diplomacy many times." You need to figure out your own philosophy towards skill challenges and a way that they will be fun and interesting to your group. Many many thousands of words have been dedicated to them and I still haven't found a framework I'm happy with.