OK, I'm getting very close to the point where I have to commit to a structure for my new home. I have to apply for planning consent, and I have to do it soon. If - improbably - I get consent quickly, then I can build this summer. And that would solve a lot of problems.
Of course, I can't actually afford, now, to build the full structure I want for the long term, so it has to be modular: I have to be able to build some 'now' and some 'later'. So let's again review the arguments and the options.
The original conception was for an underground structure sunk into a south-facing slope, comprising four hexagonal domes each 4.8 meters in diameter: one for services, one for kitchen, one for day, one for night - where the 'service' dome held bathroom, store and spare bedroom. The roofs were domed mainly because, in concrete, that's a good self supporting shape, and form followed function. They were 4.8 metres in diameter mainly because they would be cast over fundamentally plywood forms, and plywood comes in 2.4 metre sheets.
Underground is important both for insulation, for landscape considerations, and, most importantly on the very windy site, gets down out of the wind. The alternative of, for example, a timber framed straw-bale building would be both very intrusive in the landscape and very exposed to wind.
The south slope is important because it allows a fundamentally underground dwelling to have passive solar gain through south facing windows. In any case I've deliberately bought a south-facing slope for exactly this reason. An underground dwelling on a fundamentally flattish site in Galloway would have more significant problems with drainage and with daylight. I could get a more conventional house out of the wind by sheltering it behind my wood, but then it would lose the south aspect and consequently the passive solar gain.
The case for a hexagonal grid is a bit less compelling. Mass produced furniture is designed for rectangular spaces. Deliberately choosing a non-rectangular space means that much more of the interior furniture must be custom designed, which pushes the cost, either in money or time, up. However the human eye is very good at finding linear features in a landscape. Straight lines are very obvious, very noticable; and a structure with a rectangular grid exposes longer and more obviously related straight lines. It becomes more noticable in the landscape. However, my choice of a hexagonal grid is primarily aesthetic rather than rational.
However, I've now doubtful about cast concrete. I don't think I can guarantee the engineering qualities of concrete I can make on site. I could as I originally intended cast concrete units in the void and hire Alex's Unimog to move them to site, so that is still a real possibility, but these are heavy units.
Against it, concrete has very high embodied energy, and I really would have to hire an engineer to check my structures.
For it, concrete is extremely durable - there aren't any doubts that it would stay up for my lifetime.
I've considered a wood epoxy composite structure. The problem with that is that if the epoxy encapsulation is breached it will rot, and lose structural integrity; and it's hard to imagine that it can support the overburden required for good soil insulation, so insulation would need to built into the structure. Which could be done. For the roof sections, the 'well it might rot' problem is to some extent mitigated by the fact it can't be buried deep - if a module rots, it can be unburied and repaired or replaced.
However, one of the important considerations is that this is a dwelling to grow old in. As I get older, my ability to do repairs myself reduces, and my ability to pay others to do repairs also reduces. If the structure has ongoing maintenance problems it will become unsustainable.
If I'm dealing with a structure which cannot sustain a heavy overburden, paradoxically larger modules become easier to achieve. Instead of four 4.8 metre domes, I could have fewer, bigger ones. But actually the small domes have two significant advantages. Firstly, I can build them one at a time, as I can afford them. A single 4.8 metre dome would be a small but tolerable living space for next winter. Two 4.8 metre domes - the service and the kitchen dome - would make a perfectly acceptable space. And realistically that is almost certainly as much as I can afford for just now.
Equally, if I'm not designing in a heavy material with a significant overburden, the dome is no longer form following function: it serves no functional purpose at all. It becomes, in fact, a sculptural conceit - and one which does not come for free. It makes the whole structure taller - and thus harder to bury - than flat ceilings. Like the hexagonal grid it becomes simply an aesthetic conceit. Yet it remains one that appeals to me. I believe it will make a graceful space. Furthermore, actually, a wooden dome lined with birch plywood becomes an even more graceful space than a concrete one.
So the compromise solution - concrete walls and wood/epoxy roof - seems the most attractive at present. I think. I'm almost decided.