Oregonian Haruspex wrote in
:

So you take
your rough model and start trying to account for trees, the local
topography, and pretty soon it's two years later and you haven't even
bought your antenna supplies let alone run the coax because you're
still trying to account for what happens when the geese fly over the
antenna in November.


This is true too, my take on it is to make something basic and faster than
modelling it would be, to eliminate any 'plans' that were doomed to go agley
from the off...

This discussion of whether modelling is good or not is circular, how can it
end? It's like those interminable arguments about light having to be a wave,
or a particle, when it's now clear it is neither, but is a 'thing' that can
appear as either, or even both. So instead of setting each method against
each other, make them co-operate, use their complementarity. If it works for
quantum physicist, it ought to work for us, metaphor or not.