Well the most obvious thing is to not have discreet tech levels. So your tech becomes better even with one research point. To assuage roleplayers you could simply declare that you research better and better engine X, and at Y spot that research coalesces into engine Z. So if you research half and half you gain a benefit not on turn 10 but on turn 1, and at turn 20 you have learned to make a totally new engine type, and a totally new laser type, or maybe one takes 25 turns to represent a totally new invention. You can rationalize this as an abstraction, especially since standard tech trees are nonsensical and abstract anyways and generally represent the average time to a discovery, since in real science you advance in fits and starts and not linearly as in most videogames.

Instead you can only say which field is Primary (50%) and which is Secondary (20%), all the remaining fields would get like 5% each.

I like that actually. It means you're "nudging" scientists in a general direction without actually stopping all research. Makes sense to me, and still have the fun "I could get something researched I'm not actively focusing on as a goodie sometime in the next few turns" feel.

I can picture myself favoring offense (50%) almost all the time (new weapons, etc.) and keeping 20% on defensive upgrades for example.

Consider real-world science. Putting 100% effort into one field may simply be ineffective, e.g. a discovery in one field removes a road-block in a related field. You could have some dependencies, e.g. you can't get past 72% in field A unless you have at least 50% in field B or 63% in field C. Another possibility is considering the universities that underpin science. If all your research is into energy weapons, nobody will take biology classes, so when you switch to biology you're handicapped for some time. This would give the option of keeping a baseline level in various fields, writing off whole fields, or strategically capturing a planet which is good at biology to aid your new science goal.

or strategically capturing a planet which is good at biology to aid your new science goal.

It would be funny, admirably minimalist, and not entirely unrealistic if this were the great conqueror's ONLY way of setting a research agenda!

So here's an entirely different system than the one I sketched earlier. Your research window shows you what's possible in a given field, but you can't do anything about the relative percentages. (Okay, maybe you can put some extra investment in that will slowly grow your field-specific capacity, or something like that, but you can't just give the order "Attention all physicists! 80% of you are now biologists, and 20% are now archaeologists! Now, back to work!") The primary input you have on the research agenda is just conquering planets. Different planets provide research points in different fields, and if your long-range plan is max out Biology, then the good biotech planets should be top-priority conquests.

This system has some nice properties as well: you can't micromanage it, and players are free to ignore it, but if a player does pay attention, it enriches one of the core game choices ("Do I invade this planet or this one or...?")

I went for fields having set fixed types of priorities (Primary, Secondary or Minimum). You can have one Primary and one Secondary (also if you decide to have no priorities you will have all divided equally (20% for each of 5 fields).

As an interface suggestion, priorities might be more easily expressed by letting the user order a list. (Like having them drag and drop things to rearrange, or having up/down buttons on each field.) The

So like:

=========================================

Priority | Pts | Field |

=========================================

High | 73 | Energy (up/down) |

Medium | 54 | Construction (up/down) |