David Widmann

This would make the types non-concrete, I think it might be better to define ```julia const IsoNormal{T

Yes, that could be done. Might be a bit more convenient though to define and work with the more general type aliases (`AbstractIsoNormal{T}` etc. maybe?) and define the existing aliases...

Thank you for the PR! Before I add more detailed comments, I wonder if we should use [an alternative parameterization](https://en.wikipedia.org/wiki/Log-logistic_distribution#Alternative_parameterization) that is consistent with the logistic distribution. Is there a...

I think for consistency we should use the same parameters as for `Logistic` here - also in the case of e.g. `LogNormal` we don't use the (possibly) more interpretable median...

I think there are probably different considerations in MeasureTheory/Base since distribution objects are viewed differently there. Distributions.jl does not work with measures but random variables when transforming distributions, as the...

It's astonishing that every other PR in Distributions ends up with a general discussion of at least partially unrelated things :smile: Discussions are very important but I think ideally we...

Test errors are unrelated, known, and caused by the RNG changes in Julia 1.7. Probably can be fixed (IMO ideally in a separate PR) by choosing a different seed or...

It's nothing personal from my side, and hence I tried to formulate it as a general comment and to resolve some possible misunderstandings. It's just a general observation that in...

As mentioned above, the general opinion was to not overload `+` for this use case. > Concerns against + were e.g. that it might look like adding two distributions (#307...

I like this suggestion 👍 However, it seems orthogonal to this PR, doesn't it? Such an informative error could be added regardless of whether a Unicode alias for `convolve` exists...