There has always been a fine balance in gaming and game design between challenge and frustration. As game designers, we want our players to constantly feel like their personal level of expertise within the confines of a given game or genre is always put to the test without allowing the player to fail that test (or at least to fail it often). If a player is playing a game as intended and isn't missing some fundamental gameplay principle or mechanic, we don't want to frustrate that player for playing the game as intended. The ideal scenario is that we want to challenge gamers, not frustrate them.
Challenge is a term that the gaming and game development collective all use and practice, but is theoretically relegated to some nebulous understanding. Challenge is the intentional introduction of gameplay forces that work against player progress as a means of encouraging skill growth or adding meaning to player achievement. If challenge is thought of as a force that impedes player progress for purposes that are beneficial to the player, then a primary reason would be enforcing a certain skill requirement that forces players to either learn new mechanics or think of new strategies of play. Designers don't want players to necessarily feel like they're better or smarter than the game at all times, or else we're ruining a player's sense of interest or accomplishment by constantly diminishing the meaning of their actions. And since challenge is the intentional introduction of frictional forces between play and progress, frustration can be a byproduct of the unintentional or undesired application of challenge elements into a game.
In Resistance 2, the player's progress through the game is marked by increasingly more "epic" set-piece battles where the game attempts to out-do its earlier efforts. This boils down to there being more and more varied enemies in a given battle that generally takes place in an increasingly large arena of battle. This is not in and of itself a problem for the player; in fact, it's generally an accepted method of progression to task the player with increasingly more dangerous and difficult scenarios as he makes his way through the game. The main issue with Resistance 2 that makes these set-piece battles is that enemy awareness for the player's presence completely defies expectations in its sensitivity and focus. When a player enters a major battlefield passively or peacefully in an attempt to get to cover before taking major action then his perceivable consequence would be one where all battlefield actors continue what they were doing when the player entered the arena -- if an enemy was attempting to kill one specific allied unit, that enemy would continue to engage in this activity. When a large quantity of enemies actively appear (because appearance is what matters, if a player does not and cannot know the reasoning for an action then it is irrelevant) to break off their current activities in order to target the player, the game instantly becomes an consequence-defying experience.
Theoretically, the distinction between challenge and frustration is pretty clear: challenge is good, frustration is bad. Practically, the difference between the two concepts is anything but pronounced and can either be a result of poor balancing and design or simply a player who has an unexpected style of play that the game is unable to course-correct. In the case of Resistance 2, the frustration comes out of a game which relies on cheap enemy tactics to unnecessarily supplement the intrinsic difficulty of the scenarios that the game supplies.
Do games still need need to be difficult? A great deal of the up and coming game developers and designers, myself included, are of the mindset that the games we all grew up playing are more intentionally challenging than a majority of current games. This is kind of a straw man in and of itself solely due to the fact that anyone who was around to play the games ten or twenty years ago has undoubtedly increased their gaming skills over ten to twenty years of playing games. Though, with that said, there is a still a truth the claim: older games were harder, but not necessarily because they were more challenging. Take the beginning of Super Mario Bros., the NES original, as an example. The very first few seconds of the game charge the player with bypassing a goomba enemy. Within the scope of the game, this event requires the most trivial of actions by the player, but if, for whatever reason, a given player was having a hard time bypassing this enemy and died three times, then the game was over. If the player died twice and had one left, then that player's progress through the rest of the game is going to be more difficult than a player who progresses past that first goomba with all three lives. Should the game be challenging because one player didn't know the necessary gameplay mechanics and lost one of his starting three lives in learning that he has to jump over or stomp on the goomba to pass a certain area?
The concept of giving a player a finite number of "lives" with which to progress through a game has gone by the wayside for genres of games that don't intend to thrive on a sense of retro gaming or nostalgia as part of their appeal. In general, this is for the best. Arbitrarily limiting a player's attempts at gameplay progression (a concept born out of coin-operated arcade machines) is a design ideology that is no longer required to challenge players and, instead, simply frustrates players. Games that requires players to manage lives, eventually, caused players to continually abandon their progress through a game because they could get past a leg of gameplay without losing one of those finite lives that would come in handy later in the game. If we're making a game that aims to challenge players, this is not behavior that we want to have challenged. We want to challenge a player's skill at the game, not their ability to perfect an early leg in the game so that they had more attempts at later levels or bosses.
A lot of games are attempting redefine the way in which players are challenged. Far Cry 2 encourages player experimentation amidst challenging scenarios by offering the player an in-game buffer through a mechanic that allows a player's "buddy" to rescue him/her when on the brink of death (thus eliminating the player's need to reload or restart from a checkpoint). The recent Prince of Persia game eliminates death entirely and the challenge in the game comes from performing a series of gameplay elements more fluidly. Fable 2 allows players to die but instantly resurrects them, creating, as Jonathan Blow coined it, faux-challenge. Flower makes a player's actions important and meaningful purely through the way the player reacts to the game world and the sense of flow that is earned through skillful gathering of flower petals; the game does not even provide the player with a failure state.
Challenge is not a bad quality of games, but it's given the success that the aforementioned recent games have had at changing player perception of challenge it is not a quality that all must possess. The worst way to foster player creativity and experimentation in games is to actively work to punish them when they go off-script. There is no reason that games should attempt to limit a player's ability to progress simply because that's how designers and players are trained to think of games. By rethinking the way that games challenge gamer skill, new attempts at making a player's interactions with their games meaningful can arise.