forming a discipline

There is something horribly redundant about the way work gets done in my group.
People push forward, and then months later its as if nothing was learned. The same concepts are being reimplimented and explored in new computing/people contexts. At some point it might be useful to outline the geography of the infrastructure that causes this to happen. The problem is that results are needed. So people somehow cannot afford to be outwardly interested in infrastructure.

The problem for me is that on the one hand I am good at building infrastructure, but on the other, this is not openly respected.

So my research projects move forward at a snail's pace. But what else can I do? If I push too hard in any direction, I fear that the delicate fabric will break. I see this has happened to some people. They have pushed very hard and in the end form a sub-discipline of whom there is only one expert. Then they spend years scratching their heads (or perhaps shouting and tearing themselves apart in frustration) wondering why all these idiots don't learn these perfectly obvious things that they know.

Without a robust connection to the world of ideas (which typically requires the environments that a university can offer), a subject will remain fractured and inefficient.