This short paper presents a C++ library, in a generic programming style, for more efficient de novo genome assembly. The basic idea is to use a cache oblivious data structure.

The authors do not present any concrete performance improvement numbers. Their focus, rather, is on creating a basis or framework for future experimentation.

The paper does not carry the work forward to a level that would be completely and immediately useful, nor do the authors make any such claims. Rather, the authors are offering the current source code at beta level. They stress that the code does not yet handle parallel access or concurrent lock-free operations.