Remove DualQuaternion

[sethaxen]

This PR removes DualQuaternion and the dependency on DualNumbers. It instead adds an example to the documentation of using ForwardDiff with Quaternions to get the same functionality. The code in that example is actually the first steps one could take to define a package with just such utilities for dual quaternions using these 2 packages.

This is a breaking change, but we may want to make some more breaking changes before releasing, so the version number is marked -DEV.

[codecov]

Codecov Report

Merging #92 (620087f) into master (b1f50c9) will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff             @@
##            master       #92    +/-   ##
==========================================
  Coverage   100.00%   100.00%            
==========================================
  Files            3         2     -1     
  Lines          391       287   -104     
==========================================
- Hits           391       287   -104     

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[hyrodium]

I think it would be better to release v0.5.7 with Base.depwarn (or Base.@deprecate). The next breaking release with v0.6.0 includes the following breaking changes, so it would be helpful to add detailed deprecated messages.

• Remove linpol (This function is already deprecated with Base.@deprecate)
• Remove imag (This function is already deprecated with Base.@deprecate)
• Remove DualQuaternion (I'm now making a PR to add deprecated messages)
• Remove cis/cispi Remove cis/cispi #76 (This doesn't have deprecated messages.)

I'm not sure the following breaking changes will be included in v0.6.0.

• Remove Octonion (This should be done after creating Octonions.jl package, maybe in JuliaGeometry?)
• Remove rotation-related functions Drop inefficient functions for rotation? #86
• Remove Complex-Quaternion compatibility Should Complex be convertible to Quaternion? #62
• Change norm field to property Add isunit and change norm field to property #75 (q.norm == true does not lead (q^2).norm == true after this PR. It seems better to regard this PR as breaking changes, just in case.)

[hyrodium]

Great! Thanks for the PR. I have added minor suggestions.

[sethaxen]

I think it would be better to release v0.5.7 with Base.depwarn (or Base.@deprecate). The next breaking release with v0.6.0 includes the following breaking changes, so it would be helpful to add detailed deprecated messages.

* Remove `linpol` (This function is already deprecated with `Base.@deprecate`)

* Remove `imag` (This function is already deprecated with `Base.@deprecate`)

* Remove `DualQuaternion` (I'm now making a PR to add deprecated messages)

* Remove `cis/cispi` [Remove cis/cispi #76](https://github.com/JuliaGeometry/Quaternions.jl/pull/76) (This doesn't have deprecated messages.)

Good idea!

I'm not sure the following breaking changes will be included in v0.6.0.

* Remove `Octonion` (This should be done after creating Octonions.jl package, maybe in JuliaGeometry?)

I'll comment on this in #90.

* Remove rotation-related functions [Drop inefficient functions for rotation? #86](https://github.com/JuliaGeometry/Quaternions.jl/issues/86)

This seems fine to be in a later breaking release.

* Remove `Complex`-`Quaternion` compatibility [Should `Complex` be convertible to `Quaternion`? #62](https://github.com/JuliaGeometry/Quaternions.jl/issues/62)

This will be a little easier to do once this package contains just Quaternion, although it wouldn't be too much of a burden to do in v0.6.0. Why do you think it should wait?

* Change `norm` field to property [Add isunit and change norm field to property #75](https://github.com/JuliaGeometry/Quaternions.jl/pull/75) (`q.norm == true` does not lead `(q^2).norm == true` after this PR. It seems better to regard this PR as breaking changes, just in case.)

Yeah, dropping this is a priority, but it's a lot easier after Quaternion and DualQuaternion are gone. It could be in v0.6.0, but that already seems busy, so it can wait.

[hyrodium]

This will be a little easier to do once this package contains just Quaternion, although it wouldn't be too much of a burden to do in v0.6.0. Why do you think it should wait?

I was just concerned about the release of v0.6.0 will be delayed. It's before v1.0.0, so we can split the breaking changes.

Yeah, dropping this is a priority, but it's a lot easier after Quaternion and DualQuaternion are gone. It could be in v0.6.0, but that already seems busy, so it can wait.

I agree with that. (Quaternion is a typo and should be Octonion? 😄)

[hyrodium]

The description should be updated, but I don't have permission for that.

[sethaxen]

Good point, I've reached out to request I be upgraded to admin.

[sethaxen]

This PR should be ready for final review.

[hyrodium]

Most of the example seems fine! Just added some suggestions.

[hyrodium]

LGTM!

[sethaxen]

Thanks for the suggestions! I made a few more changes. Should be finished now.

[hyrodium]

Thanks for the updates!