My biggest gripe with feats however is that there are too many. That's why there are so many bad ones.
I can't agree with that at all. The reason that there are so many bad feats is that they are too easy to make, causing designers to churn them out without much reflection. More agonizing should be done over your feat list than is usually done, but in my opinion the ideal number of classes is about 8-15, the ideal number of skills is like 20-40, but the ideal number of feats is probably 1000 (or more). There are as many feats as there are character concepts. I believe its perfectly possible to have 1000 well balanced, totally intriguing, interesting feats - though I can perfectly sympathize with your feelings if you don't think so given that 90% of the feats that were published were garbage.
Trim it so you have only a page or two listing a few dozen feats, make them all combat oriented so we don't have to choose between flavor and power
Again, I can't agree with that at all. Combat oriented is not the limited of character concept, nor is it a sacrifice of power necessarily to gain an out of combat ability and forgo an in combat ability. That you assert otherwise implies that you think physical combat is the only important thing that can happen, which I refuse to accept. Besides which, often a feat can be both a combat ability and an out of combat ability either because it impacts movement, and hense your tactical options, or because you've designed your skill system well so that skills provide for active abilities with the potential to impact combat.
For example, in my game all of the following skills directly impact combat either by being used in combat, or by relating to using or defending against combat manuevers: balance (movement, resists trip, circle manuever, clinch manuever), climb (movement, some clinch manuever checks), bluff (diversion manuever), disguise (distract manuever, feint manuever), escape artist (grapple defence checks), hide (increased surprise chance), jump (movement), leadership (buffs and debuffs), listen (resists surprise, resists invisible creatures), move silently (increased surprise chance), porter (reduced encumbrance penalties), run (faster speed), sleight of hand (quick draw manuever, filch manuever), sense motive (many manuever defence checks), spot (resist surprise), tumble (movement, evasion, fast stand, roll, etc.), tactics (more than I can list), use magical device, use rope (lasso) and probably more I can't remember. And that's not counting skills that lead to feats or feats that use skills. Many of those skills also have out of combat uses. At that point, well, investing in out of combat abilities is investing in combat abilities.
It's too easy for game designers to release a splatbook or Dragon article full of bad feats.
Yes, now you've hit upon it.
Keep the feat list tight and focused and we will only have good feats to choose from.
The danger of that is that it encourages designers to see a new class as being the solution to everything, or to think that only spell-casters can get good stuff because all the interesting things should be spells.