[PGHArch]

I agree as well. Long ago (in a galaxy far far away) I believe I mentioned something like the following model to Malfador, but it is very possible I'm thinking of something I e-mailed a different game designer. I've been pushing this model to anyone who will listen... In any case here is my idea:

Have a "science tree" and a "technology tree" which are two seperate things. The technology tree would work largely as it does now, you apply effort to a science area or group of areas, create a prototype and build. This is the engineering aspect of it. There are few limits but practicality, eventually the costs of developing better basic armor types outweigh the benefits of the improvements.

The science tree would be where the uncertainty comes in. You have an assortment of fields you know about. You may invest in "theoretical research" in any of those fields. The trickiness is twofold. First, you are only creating a probability of making a discovery each turn, there is no set goal. Second, some discovery probabilities may be almost none unless you have appropriate simulatanious research in multiple science fields. So for emissive armor to be "discovered" you may need to be researching both physics and chemistry, and the probability is the lower amount.

Analyzing technology based on an "unknown" scientific development has a one time higher probability bonus to open up that field for your own development.

Also, some scientific research would simply open other theoretical areas, having no direct applicable products.

This creates a dynamic effects in that you can get lucky or unlucky in basic research, or you can go with the more steady but in the long run less innovative approach of engineering investment.

I think malfador is halfway there with the divisions between weapons, theoretical and applied technologies, except that the model of how they are researched is all the same, so there is no real difference other than the name. It is probably too late to make the sort of heavy change I'm suggesting, but there it is.