And this gets back to my earlier statement which seemed to have been misconstrued. I don't see the difference between Elemental and the Engine from a players POV. Certainly there is a difference, but as far as design goes, there really shouldn't be. Issues with one affect the other, and being forced to sacrifice on one side or the other in such an open environment as the beta was, isn't going to bother the players with distinguishing whether it was the engine limiting the game, or the game itself.
This is a good point. The downside to our rather...unique level of openness is that you guys aren't talking to community managers but rather developers themselves.
So to us, we look at "the game" in a very different way. To you guys, you see Elemental. To us, we see Kumquat, Havok, Miles, LBS (Land Based Strategy Engine) and Impulse::Reactor.
"Elemental" is a thin cake on the top of all this.
Years ago, I turned Galactic Civilizations for OS/2 into "Star Emperor" (think Warlords in Space) in a weekend because the "game" is just a thin layer over a lot of existing software.
90% of the work on a game is making the underlying engines.
To go back to GC2, the main gripe I always had with it was the open space decision. Now I get that without it the anomalies, culture, space docks, etc... would not have been able to work as they did, but the down side really was an AI (or AIs if you prefer) which could not manage the strategic complexity of design. Contrast that with other space 4xs where everything happens at 'nodes' and you see an immediate simplification, even in cases where there is still open movement (not limited to fixed warp lines). That's not to say that GC2 wasn't still enjoyable or even occasionally challenging, but it does underscore a design decision (in this case not engine limited I imagine, but to some extent perhaps because the engine wasn't designed to focus on a node construct) which ultimately made the AI less effective and less challenging than it otherwise may have been.
GalCiv was always inspired by Civilization. Literally Civilization in space. I wouldn't do nodes in GalCiv because I already had a favorite game that did it well - Master of Orion.
Of course take out the open space and GC2 is a different game entirely. A better one in my opinion, as unimportant as it may be since we didn't get a node based space4x from Stardock...yet (can I dream?)
Sins of a Solar Empire is node-based and is 4X.
Back to elemental then. What are the core complaints with it? Magic and tactical battles, and AI, but AI is understood to be a work in progress so I give that a pass for the time being.
The core problems with Elemental are in my view: Magic, Infinite Force Projection, and tactical battles.
The Magic was limited because of the game engine. Infinite Force Projection was due to a poor design compromise by me to try to work around the magic system problems. Tactical battles were problematic because I had to gimp the design after the continuous turns system turned out to be a disaster.
How much of the issues with magic and tac battles were driven by the engine, vs. driven by some kind of crisp and clean design mechanism? I get the feeling that Elemental was driven by the Lore Brad wanted to use to make his fantasy game, and by some new engine which was supposed to allow for ease of modding in many different areas. Pardon me for suggesting then, that perhaps this was missing a crucial aspect of what makes for a compelling game. The underlying mechanics and design decisions. I think the beta crowd echoed this sentiment, and worried that with the earlier release the 'balance and game play' beta was being truncated, if not skipped altogether.
In both cases, the engine dictated the design. Not originally of course. The original design of Elemental was essentially MOM but as budget, scope, and internal capability began to come to the fore (and this is why developing so many new engine systems at once is a bad idea) the design was altered.
Patch and polish are the usual answers, and that's fine, commendable even for the support SD provides to their products. Yet some decisions made early on send the process in a one way direction which cannot be reversed as initial design decisions become ingrained in the engine to the point where it's not possible to change them. Good or bad will depend on individual tastes ultimately, and my particular bent is of no more value than anyone elses. Yet I still maintain that moving forward with those decisions without a clear underlying concept is more likely than not to end up with a product unable to ultimate tap its vast potential.
Nothing is as simple as it seems. 
Writing the original design document for Elemental wasn't terribly difficult. Take Master of Magic, add quests and goodie huts and NPCs, and unit design make tactical battles continuous turns so that you can have really HUGE armies battling it out and voila, game of the year.
See how easy that is to design? When I see someone say "Why didn't they copy MOM more" my only response to that is "Well, duh." 
The hard part is when you don't have an engine that can do any of that and you have to develop it at the same time. You find out some things are harder to do in 2010 than others and begin to make compromises to your design. Add months of growing exhaustion and fatigue and pretty soon you start thinking "Hey, yea, enchantment slots and local mana, great idea! I'll buy that for a dollar!"