Dumb question: is it appropriate to describe fractal trees as an "extreme" version of a log-structured merge tree? At a glance (i.e. if you squint really hard),

Dumb question: is it appropriate to describe fractal trees as an "extreme" version of a log-structured merge tree? At a glance (i.e. if you squint really hard), the buffer flushing process in a fractal tree seems kind of similar to the tablet merging process in LSM trees in that they both batch operations to amortize secondary memory access costs, except fractal trees generalize it to all operations.