Editing CRNs with limited geometry (section)

=== Computational power of polymer classes ===

'''Question''': what computational power does each class of polymers exhibit, when described as augmented CRNs?

Motivating example: CRNs in a well mixed solution cannot stably compute beyond semi-linear functions.  Can augmented CRNs?

'''Ex''': Tree polymer construction from Luca.  (CHECK.  Construct tree of height $\Theta(\log n)$ in expected time $O(\log n)$.  One spline of tree forms active molecules.  This would effectively compute f(x)=\Theta(\log x ) if correct.  Well mixed CRNs cannot.)


'''Prior work'''
Chen and Dabby, and Malchik and Winslow<ref name="dsd">Caleb Malchik, Andrew Winslow. (2014). Tight Bounds for Active Self-Assembly Using an Insertion Primitive. [arxiv | http://arxiv.org/abs/1401.0359] </ref> work on a polymer model. The latter paper shows that with $k$ DNA hairpin monomer types, we can deterministically construct polymers of length $ 2^{\Theta{(k^{3/2})} } $, quickly; and that the expressive power of these systems is exactly the context free languages.