Frogboy,
I truly hope that you do as you have stated and make this into the game it could have been...
That having been said I acknowledge my sincere disappointment in the game in it's current state but, since you are asking for what may be done to improve things and if I read you correctly what we think should be improved upon to help E:WoM turn the corner and achieve something special here are some of the core elements I believe need to be worked on the most.
The feel of the land:
The world is supposed to be a post cataclysmic recovering world devastated by magic and trying to recover so how about giving us some of that feel. When a civilization collapses There are always things left behind, these things are generally obvious in the ruins that are left and the objects on the landscape showing signs of the exterminated peoples and cultures. Using this idea there are things I expect to see and ideas incorporated.
Ruins :
City ruins, Special unique structures unable to be understood, Abandon Temples, Ancient Landmarks, Oracle locations, Strange towers, Forts, Signs of the Cataclysm, Battle ruins, burned out castles, abandon schools, Ruined port on the sea, Abandon trading posts, and some kind of roads or trails showing pointers between a few of these location showing people once traveled between them.
In each type of ruin I expect there to be hints about what it was used for and perhaps a chance of discovering something random and appropriate to that type of ruin: Ancient books, Devices of forgotten technology, random amounts of gold, weapons, people who revere or worship the ancient object, hints at other locations, legends and rumors of other greater possibilities and rediscoveries.
Unlockable locations - The current mechanism for locations seems partially right but seems to miss the boat in many respects, especially where it comes to the random spawns based on research for entire groups of locations. In my opinion About 80% of all locations should be generated on the map at the beginning of the game and be visible upon entering into scouting / visible range of the square or at the very least upon entering the square. The lowest levels of locations should have between 50-100% of the visible locations able to be entered without any research much like the abandon caravans are currently. Research should then instead focus on making them appear through the Fog of War as location icons perhaps within X spaces of the actual location to account for unreliability of rumors ect... The research paths could be logical topics such as: Paths of Commerce, Ancient Battlegrounds, Schools of the Elders, Temples of _______, Homes of the Titans, Cities of the Past. Different amounts of research reveal more of these through the FoW with a small chance of making one of the other originally non-visible locations appear. Additionally to actually unlock the locations of the mid and higher levels in order to enter into them specific research must be done on those types of location. This is the research needed to overcome the runes or have the knowledge of the places to look in order to find where the items of interest might be. At the highest level of locations in addition to the normal research to make the locations show up a spell could be unlocked, a Ritual of Opening for X, when this spell is then researched and cast or simply cast it opens all of the locations of Magus the Mad or whatever. When that spell is then cast the locations become open and there is a chance for random monsters to be released from them much like the current mechanism.
Tech Tree:
I should not have it be more difficult to research a very basic idea within a tech category simply because I have spent a large amount of effort researching a more advanced idea within the same category. With the current design of the tech tree if I advance one portion of a tech tree it affects all other research costs within the same tree. Specific "dead end" research branches should be resource related i.e. mining should be a general research followed by smelting and on up the line but research into mining specific ore types or things such as beekeeping could be tied to a specific research to get full value from those resources. This would account for an improved understanding of how to effectively exploit that resource rather than merely have it present.
Interrelation of items within a tech tree, mining should need to be achieved before metal weapons and armor are available.
New topics included in the tech tree, these could then be built into the prerequisite for a variety of other structures. Among those items which are frequently lost in the fall of a civilization are the advanced forms of various skills such as carpentry, stone working, metalworking, mathematics and mechanics. These skills are often ones that are known by the eldest and most vulnerable in times of crisis so they are among the fastest lost especially when the apprentices and journeymen are sent to war and many of them lost during war. These same techs that are ones that different factions could start with in order to help with race flavor.
Map refinement:
Quest locations, Goodies huts, Notable locations and perhaps even lairs of monsters should spawn their own map locations as tactical maps rather than be simple encounters on the terrain type of the land under them. This will require the addition of a variety of additional terrain tiles and so I expect would require being delayed until the second expansion but, a series of basic wall and corner tiles and a basic floor tile would give the community something to work off of for modding into more types such as ones for specific types of encounters such as laboratories, temples, catacombs, caves and lairs.
The parts of the maps that were not initially included in the game but were intended to be present such as most notably rivers and some of the special terrain that goes with them. I would also recommend adding an additional level of magnitude of height to maps this would allow for more dramatic maps and allow for steppes, valleys, ravines as well as the look of a land devastated and torn apart by magic. Special lands showing the after effects of magic would also be a nice future enhancement at this point, lands showing strange crystals, chaotic storms of magic, floating islands of land still hanging after devastation by magic, dark corrupted forests, desserts of shifting sands that send those who enter them to random locations and any number of other refinements that could be used to enhance the game, both by the player mod base and by your designs in campaign creation.
An official random map generator, the current one made by Cephalo is a great start but it needs further refinement and shouldn't have been left to the community to develop. An official random map generator should include spawn rate of resources as well as a more complete inclusion of terrain types and perhaps sliders for climate and cliffs / beach. At the same time include rivers if they have been corrected at the point you release this or patched once you have done so.
Magic:
The current spell books lack depth and a cohesive feel to them, as well any one or two spell books will end up giving a sovereign the complete variety of spells one would need to win the game so there is no reason to take focus on magic beyond that. Taking more than one school of magic should enhance starting essence or more variety to spell choices.
AI:
AI strategy needs to have a chance of competing with players and allow for future mods by players so that players may influence the behavior of the AI in their own mods.
AI Tactical strategy needs to be developed to include something more than charge straight ahead, it should take into account it's own movement speed vs. that of the opponent as well as not closing to the adjacent square unless able to initiate an attack the same turn with the only exception being when it's own movement speed is 1.
Other issues:
Weapon balancing, unit balancing, city balancing, resource balancing, interface on the map editor (think photoshop), a quest editor and any number of the other complaints in the forum as well.
At this point I am out of things I can recall at this moment but am sure once I fire the game up again I'll come up with more since there was so much potential that seemed to fall so far from expectations and are still so many things that can be done to improve things.