Aitchison

[antagomir]

This solves the naming issue in #433

Method names in vegdist are now changed from "aitchison" and "aitchison_robust" to "Aitchison" and "rAitchison", respectively.

In addition, I added the transformations "alr" and "ilr" due to recent requests. These were discussed in the original issue as well.

In addition, I updated unit tests for all these methods. The tests include also comparisons with alternative packages, to show the compatibility (those tests with external packages are commented out, though).

Everything seems to work and properly tested, but cross-check always good.

[antagomir]

Woops, I had forgot to follow-up whether the automated builds go through. Tried to fix, hopefully all ok now and can be merged soon.

[antagomir]

All checks seem to pass now. Ok to merge?

[antagomir]

Is there anything else I could do more to get this approved?

[jarioksa]

Sorry for this. I have been busy with other packages (such as Hmsc) and I haven't had time to look at this. It is my unforgivable ignorance. I'll check the PR early next week (on Sunday you shan't work).

[gavinsimpson]

@antagomir A couple of things I noted:

1. Can you confirm that you are the original and only copyright holder of the code that you modified from the {mia} package?
2. Assuming 1., are you agreeing that the code can be released under the GPL as per vegan's license?

I note you are a contributor to the mia but it's not immediately clear who owns the specific code that was modified and included here. If you are the sole copyright holder, then assuming you agree to 2, this question is moot. If you aren't the sole copyright holder (i.e. if anyone modified the code you included here) then we have an issue of compatible licences (mia is Artistic) and we are likely violating the mia license as the copyright file is not included here in the PR.

3. I would recommend that we avoid method names that contained mixed case. None of the methods we currently support in decostand use anything but lower case or a period.

   I would suggest: aitchison and robust.aitchison if we aren't allowed underscores (I get Jari's objection but these are character strings not function names :p ).

[gavinsimpson]

Sorry - I too have been busy and meant to comment my 3. much earlier (though that is easily fixed via a patch to the code) but 1. ans 2. need addressing I believe (unless you know different @jarioksa ?)

[antagomir]

@gavinsimpson

1-2: We have in general agreed to move this code from mia to vegan among the original authors, so I think this is clear but the details are indeed important - I will cross-check and come back to this soon so that this is on record.

I can make PR with the new names aitchison and robust.aitchison if both @gavinsimpson and @jarioksa agree about the final names :-)

[jarioksa]

I lightheartedly merged this, but @gavinsimpson was younger, aware and spick & span on legalities (thank you, Gav!). The rights and licensing issues are the most crucial ones: if they are not settled, I must revert the merge. All other things (naming, mixed cases) are lightweight and can be changed on a whim.

[antagomir]

The PR #498 renames aitchison and robust.aitchison as suggested.

License:

1. I can confirm that me (Leo Lahti @antagomir), Tuomas Borman (@TuomasBorman), and Felix M. Ernst (@FelixErnst) are the original and only copyright holders of the code that I transferred here from the {mia} package. We are also the main developers of that package.

2. We agree that the code can be released under the GPL-2 as per vegan's license. TB and FE have expressed their approval that we can do this in mia issue #125.

I think I wrote the ILR also on my own based on some notes, many years ago, but could not trace back full origins of the ILR code to confirm GPL2. The ILR implementation was also otherwise slow so I just removed it to keep things simple. Better to add a more optimized implementation if needs will arise.

We wish all is ok now :-)

[gavinsimpson]

@antagomir Thanks for this.

Sorry to be a pain, but could @TuomasBorman and @FelixErnst also indicate here (via a comment) their agreement to our releasing the relevant code under GPL-2, that would be best (just in case someone deletes the {mia} repo for example...).

[antagomir]

Sure.

[TuomasBorman]

I agree that the code can be released under the GPL-2

[FelixErnst]

Agreement from my side as well.

[antagomir]

Thanks both. I wish we could proceed with this again :-)

[jarioksa]

Yep, it was merged – and it will stay merged. Next release.