In its broad lines, my idea is as follows : the gamer should be presented with a number of research areas (weaponry, shields, power generation, thrust engines, hyperspace engines, cargo hold, hull size, ...) and given characteristics of these areas (range, size, power requirement, encumbrance, ...). The gamer can then choose to evaluate the feasability of research in enhancing a given characteristic. In addition to that, the gamer should be allowed to design their own hull shapes - if he/she so wishes.
So we have a game that allows for various levels of difficulty in research. Proposing two types of games - one with a set tree, the other without a tree - could allow for vastly different game experiences.
How can one imagine such a research tree ? That seems simple. Although the research areas are defined, the characteristics can be determined and modified by time and resources. A weapon can be defined by a given power, range and encumbrance. Reducing encumbrance for a weapon allows for more weapons to be stored on a given hull. Increasing the size of the hull allows for more weapons also. Following the gamer's evaluation, he will choose either path to increase the firepower of his armada. Of course, increasing hull size will increase the load on the engines, lowering top speeds in all areas for the same engine type. But a larger hull can accomodate more engines, thus increasing speed but also increasing demand on power generators. The same process can be applied to weapon range, shield resistance, anything.
How to manage the upgrade ? Once the user knows in which direction he should go, the program should propose an evaluation : an increase of 5% should take x days and cost y. An increase of 10% would cost a given multiple of days and resources. The user could eventually input his own percentage increase. The program could eventually display a probable limit in increase, given current technological limitations. Any increase percentage approaching the theoretical limit will take exponentially longer than small increases.
In that way, the user will find himself defining his own weaponry, engines and hull sizes, implementing the models that are deemed most interesting, using the others as basis for further research.
As for the hulls, what are they anyway ? In developer terms, a hull design is a simple placeholder that can be of any shape or size, such details do not impact combat algorythms, just the way the gamer sees his vessels on the tactical screen. If a way can be provided for a gamer to create his own forms, that only impacts the rendering part of the game, not the core functionalities. The tactical viewer must be adaptable though, since gamers will no doubt strive to have the biggest hulls their research can create. So the viewer must be able to zoom and de-zoom to a fair amount of orders. Late in the game, capital ships will become monstrous and fighters will probably not be much bigger than at the beginning, so if the gamer wants to step back a bit, the fighters will probably become invisible.
Imagine such a game with a gamer who wishes to create himself StarWars-like capital ships. We are talking Star Destroyers, Super Star Destroyers and Death Stars. Sure, you cannot choose the weapon types, only their power. But a hull like the Death Star could contain a fair amount of heavy weaponry, largely enough to spell doom to any fleet crossing its path (resource issues put aside, of course). In this type of game, the gamer could satisfy this basic drive with the greatest realism. So he doesn't have turbolasers and ion cannons ? He can still make the stiletto shapes and create a deadly aura around powerful weapons. That is what counts.
Of course, if multiplayer is possible, this would be the game to end all games. Pit two players against each other for the conquest of the galaxy and you'll have to end up with two radically different designs, and capabilities. That is nirvana as far as hardcore strategy gamers are concerned - at least, I think it is.
So there you have it. I welcome comments on this idea, but I would welcome even more a game that uses it. Think about the possibilities, and send me your opinion, or suggestions. I'll be happy for any feedback on this. Meanwhile, if you want my thoughts of how it could feel to be a player in a game with such a research engine, follow this link.