Daan Huybrechs

To be sure, in the current version the DomainSetsCore package does the following: - it defines domaineltype for

Hence, if IntervalSets would switch to using DomainSetsCore right now, the `domaineltype` of an interval will automatically be defined (because it is defined for `Domain`) and the `eltype` of an...

Having reread a few issues (thanks for the link to gentype, interesting related discussion): moving the definition of Domain to a new package and settling the `eltype`/`boundstype` issue are two...

Thanks for the overview, that helps to clarify the issues. I agree that these are possibilities, with pros and cons to each, but for the sake of completeness I think...

Looking at options: - about A: to me there is no genuine difference between Domain without T and all domains defining domaineltype, and having Domain{T}. It makes an implementation difference...

To add, I'm not in favour of any solution in which `AbstractInterval{T}

For the sake of the argument, I disagree with the idea that there is no iteration involved in the definition of intervals (and hence no link with `eltype`). The name...

> I thought the plan was to move to more of an interface than an abstract type? Why not use this as an opportunity to just delete `Domain{T}` once and...

Okay to register DomainSetsCore.jl? Since the discussion here I've removed `eltype` from the package, so we are in scenario C. Using DomainSetsCore as it is now is a non-breaking change...

FYI, we're preparing a breaking update of DomainSets. With some delay... :-) It might be an idea to proceed? This package has several open issues and pull requests related to...