Robbie, I think you underemphasized one of your more important points.
You saide "and still contain choices which effect the game.". Allow me to rephrase that:
"Have meaningful decisions". Let's break the idea down to base parts, those of you who ever GM'd an RP game should know this.
Meaningful decisions means two things. The first is that the player has the ability to choose. If there is no choice, there is no fun (and therefore no point). I.E. Civ's research becomes pointless when you reach the future tech because you no longer need to make a choice.
The second part of the idea is that the decision the player makes matters. Having multiple choices that all gives the same result is pointless (and therefore not fun). For a decision to be meaningful the player needs to a) be aware of the possibilities and
be aware of the reprecautions. So, IMHO, a good research tree will give you the ability to choose between saveral different options, each of them giving a definite (and useful) bonus. The moment either of these criteria are not met than we have a shitty research system.
For future tech, my suggestion is this: Don't do it.
Make the tree big, make the tree complex, make the final techs exponentially harder to get, but don't bullshit us.
When a player researches *everything* (which should happen only in a 'sandbox' setting) he just get's "omnipotence" and has access to the wizard tool (able to modify the game as if he creates a premade map). Either that or Science Victory.