The current API asks for writer and reader types. When mutation/diploid/etc. types change, these writers and readers also need updating. Currently, they are in different parts of the library, which is a problem.
A more idiomatic approach would be to rely on argument-dependent lookup and require that types do the right thing.
This is related to #67, which proposes a change to a type's data members, causing side-effects in other parts of the library.
The current API asks for
writerandreadertypes. When mutation/diploid/etc. types change, these writers and readers also need updating. Currently, they are in different parts of the library, which is a problem.A more idiomatic approach would be to rely on argument-dependent lookup and require that types do the right thing.
This is related to #67, which proposes a change to a type's data members, causing side-effects in other parts of the library.