Richard Jones recently wrote a blog post ending with the statement, "Matter is not digital."
Jones was responding to Eric Drexler's statement to the NAS that, just as computers are amazingly useful general-purpose computation machines (e.g. capable of displaying movies), molecular manufacturing could lead to general-purpose manufacturing systems making a very large range of products. Jones argued that, "For the idea of general purpose manufacturing to be convincing, one would need to believe that there was an analogous way in which all material things could be represented by a simple low level code. I think this leads to an insoluble dilemma - the need to find simple low level operations drives one to use a minimum number - preferably one - basic mechanosynthesis step."
A much more detailed analogy can be drawn between molecular manufacturing (MM) and computers. And it shows that the issue is not whether matter "is" digital, but whether digital design principles can be applied to matter. After all, movies are not digital. But they can be handled digitally.
Transistors, the basis of modern computers, are not digital. They are nonlinear, meaning that it takes very little effort to make them switch from mostly-on to mostly-off. That lets them ignore the error in the signals (in exchange for some energy), allowing us to design circuits as though the signals really were a digital ideal.
Forces between atoms as they bond are also nonlinear. As you push them together, they "snap" into position. That allows maintenance of mechanical precision: it's not hard, in theory, for a molecular manufacturing system to make a product fully as precise as itself. So covalent bonds between atoms are analogous to transistors. Individual bonds correspond to the ones and zeros level.
Computers are not designed in terms of individual transistors. They use microprocessors, which contain thousands to billions of transistors. The microprocessors implement a few dozen instructions called "assembly language" which are carried out repetitively. This is analogous to the mechanosynthetic reactions. You don't need as few as possible. You need a small but convenient set. A system which could carry out a few dozen synthetic reactions in programmable sequence could be very useful.
Jones continues: "But in limiting ourselves in this way, we make life very difficult for ourselves in trying to achieve the broad range of functions and actions that we are going to want these artefacts for. Material properties are multidimensional, and it’s difficult to believe that one material can meet all our needs."
By distinguishing between the nonlinear, precision-preserving level (transistors and bonding) and the level of programmable operations (assembly language and mechanosynthetic operations), it should be clear that the digital approach to mechanosynthesis is not a limitation, and in particular does not limit us to one material. But for convenience, an efficient system will probably produce only a few materials.
Now let's consider the separate question of whether a few programmable material-building operations are suitable for producing a wide range of output types.
For nanoscale operations like binding arbitrary molecules, it remains to be seen how difficult it will be to achieve near-universal competence. But most products that we use today do not involve engineered nanoscale operations. Instead, they work at a scale of many microns. If we can design and build products containing machine systems far smaller than a cell, then we should be able to make virtual materials with properties that can be engineered over a wide range.
For example, a parameterized nanoscale truss design could produce structures which on larger scales had a vast range of strength, elasticity, and energy dissipation. A nanoscale digital switch could be used to build any circuit, and when combined with an actuator and a power source, could emulate a wide range of deformable structures. A few designs for photon handling, sensing (much of which can be implemented with mechanics), and so on should be enough to build almost any reasonable macro-scale product we can design.
Of course, the design of complicated products won't be simple. No one writes a modern user interface in "down to the bare metal" assembly language, but instead uses object orientation, scripting languages, an operating system, and a ton and a half of libraries. A product like a car will likewise not be programmed "down to the bare atoms." But assembly language is not a limitation but a foundation; it is general purpose, and all the tools of software are built on top of it. Likewise, the mechanosynthetic reactions and the nanosystems they build will not be a limitation, but will provide a foundation for endless recombination, elaboration, and hierarchical design.
Yes, matter can be treated digitally: bonds make molecules, molecules make machines, machines can be engineered. This is the strength of molecular manufacturing.