Sam Nicholls

`Gretel` only cares about `L` pairs of observations, so why store observations for pairs more than `L` SNPs apart? Can a representation that discards this information still support paired ends?

`Hansel` considers symmetric observations, thus half of the `(A^2)*(B^2)` matrix (for A symbols and B positions) is strictly zero. We can thus collapse the two spatial dimensions into one dimension:...

Add-one smoothing is pretty naive and lame, can we improve it? Under what circumstances does smoothing have much of an impact?

It seems we weren't alone in our clever naming, I've made a grim(m) assumption that in practice that `hanselx` will never bump into [`hansel` PyPI prime](https://github.com/alexsavio/hansel). Thus although the PyPI...

The pedant in me can cope no longer.

Submitting a zbase32 barcode with a `0` zero should convert the zero to an alpha `o` and raise a warning.

I'm sure this is a common design pattern somewhere