We've discussed this at great length. I think the area that is the most illustrative is the lack of grokking of how Keywords and the exception-based design philosophy are meant to work together.
Consider the AD&D and 3.x Fireball. These lengthy (unwieldy imo) spell descriptions are chock full of all kinds of exceptions embedded right into the spell itself. The mindset of the D&D player used to "casting spells" or adjudicating spellcasting was predicated upon interaction with this layout. Given enough time, somehow, this real-world book format was somehow then internalized as "magic" in the fiction of their gameplay.
Now consider any 4e Fire spell. It contains the Arcane and Implement Keywords. However, it also contains the Fire Keyword. This is absolutely fundamental as it is one of the two primary system components (the other being the math/noncombat resolution framework) for the exception-based design philosophy that the system is predicated upon. The 4e power statblock is meant to elegantly provide (a) only the necessities to facilitate minimal handling time in combat (single target or AoE, ranged or melee, action economy, etc) while also providing (b) the Keyword infrastructure such that when a player proposes an exception-based usage, the GM has the means to adjudicate the effect on the environment. The GM knows that Fire is "explosive bursts, fiery rays, or simple ignition." So if the player of the Wizard is in a Skill Challenge and wants to deploy the spell to ignite materials (for whatever reason), this is an exception (though clearly orthodox) to its standard stat-block usage. The GM uses the of-level math to determine a DC and the player rolls an Arcana check with the outcome being either intended effect (ignition for whatever sought end) or some kind of complication (perhaps still ignition but things may go pear-shaped because the fire/smoke becomes a hazard/impediment to be overcome).
Instead of this being intuitive/easily grokked because the Keyword system and exception-based design philosophy is comprehended, we get:
- "This powers system means there is no magic and/or martial characters are casting spells."
and/or
- "Players can't use any fire spells to cause ignition of building materials (etc) because the target line reads 'creatures'..."
Unfortunately there is no section in the rules pertaining to this. I don't even see any indication that the designers INTENDED keywords to work in this fashion. I think it is probable that it was something they THOUGHT OF at some point, in passing, but that the primary intent for keywords was the purely mechanistic interaction with other game elements (IE vulnerability/immunity, other powers, items, feats, etc).
Regardless, they failed utterly to even mention the narrative function of keywords (which is more obvious, the DM describes the fire attack as FIRE, but still not ever discussed). More critically they failed to mention the use you describe, the extension of the rules into areas that aren't covered explicitly, the use of 'Page 42' and its ilk. It was a fatal oversight.
Personally I think the coherence of the objections to 4e has less to do with 4e 'weaknesses' and more to do with fan expectations for D&D. My experience is that most D&D players want to reproduce an EXACT experience with no variation. The details vary depending on when the player was introduced to the game, and there are certainly players that are more flexible, but the typical player is playing 'D&D' and they expect, just like Monopoly, that D&D will be a specific exact thing. Its irrelevant to debate with them or even speculate on why they have specific likes and dislikes.
That isn't to say that we cannot discern the areas that 4e could be improved on by looking at what people have said. I think some forms of complaints are largely spurious when considering 4e IN AND OF ITSELF as a game. For example I put the "classes are too much the same" complaint in this category. Its not a complaint about 4e, its a complaint that arises out of an expectation from a different game.
As for the whole thing with Rituals and being an area that 'needed improvement', I think my answer to that is that the text of the rituals themselves needs improvement. I also think that perhaps even using a scroll might want to be a trained only skill use, and that certainly casting a ritual itself should require training in the requisite skill(s). I think there might be more skills involved as well, though not TOO many. This would force ritual casters to specialize a bit and make things more interesting.
As far as the ritual text itself, I think ALL rituals should be scaling. Many of them are simply due to being a 'beat some DC' sort of thing (like Knock where you have to beat the DC of the 'lock'). Many others however don't have scaling effects at all. These rituals often suffer from becoming trivial at higher levels or are too powerful all around. The other thing that should have happened is that rituals costs should scale with results. If you are getting flying steeds from your Phantom Steeds, that's kinda expensive but getting an ordinary riding horse costs next to nothing.
I think the other thing is that rituals should probably be devised to be a bit more immediately relevant in adventuring. A lot of them ARE, but a LOT of them are pretty auxiliary and it seems to me almost like the design decree from on high was "rituals should never be required to make things work, they should always be an alternate way to do things". Obviously that inherently makes the whole system redundant.