Other factors by which the survey of processes should be ranked is whether or not they can create devices that are powered, whether they could make a molecular actuator, could they make a complex synthetic material like a phonon mediated architected room temperature superconductor, how scalable are the current options, are they a path to exponential manufacturing or would they go to terahertz operation with trillion manipulators operating in parallel. how good could a creative application of several current systems get in 2 years, 5 years, 10 years, 20 years. Other rating factors error rates, reliability of operations (some DNA insertion methods work 1 out of 5 times), range of materials that can be worked with, current limitations, fundamental limitations (Note: when brainstorming there are paths and ways to avoid solving limitations. to engineer implified but useful processes that can still achieve a goal but take longer. You can use multiple processing steps, macroscale systems can help manipulate things where the molecular scale is having problems. Flush different reactants in and out of a picoliter chamber using microchannels. Use nanopore filters to control what goes in or out. Find ways to kluge a better next stage of capability.
A quantum computer roadmap looks at ranking each of several promising technologies and roadmaps each for work that needs to be done to take each to the next level. The 2004 roadmap did seem to miss the analog (adiabatic) approach. The Sandpit should also be aware of the advanced work on scaling ion trap quantum computers. They have good control of ions.
Fit metamaterials into the developing capabilities area. When we will have useful superlenses made from metamaterials. Superlenses applied to laser manipulation, laser lithography etc… Metamaterials also give better control of terahertz radiation by fine adjustment of nanowires. It could give finer magnetic and electrical control.
How can the chemistry of some of the processes be extended? Building from DNA nanotech with more synthetic DNA, more kinds of polymers, combining that manipulation with nanotubes, silicon, metals, dopants, etc…
There is also the work for hijacking and manipulating viruses and bacteria and cells to do engineered work. Viruses making batteries (6 nanometers diameter by 880 nanometer length).
Some sweet spots for improved capability seem to be getting STMs and other microscopes able to be 5-10 times more precise and repeatable down to about 1 angstrom. Then a lot of the work that Rob Freitas and Ralph Merkle are computing would be a lot easier. Freezing things could be a short term way to make up for less ability in the tool.