add L_1 norm

[UnixJunkie]

aka "Manhattan distance" it is supposed to behave well in high dimensions

[UnixJunkie]

cf.

Aggarwal, C. C., Hinneburg, A., & Keim, D. A. (2001, January). On the surprising behavior of distance metrics in high dimensional space. In International conference on database theory (pp. 420-434). Springer, Berlin, Heidelberg.

[UnixJunkie]

Also, implement a fractional distance metric, where f in ]0:1] could be optimized depending on the application.

[UnixJunkie]

we use Tanimoto when FPs are high dimensional