Dear Nathan,

The reviewer with considerable concerns last time has now provided her report and continues to have significant concerns which need to be addressed.

I must also say that I am not satisfied with your response on the code. It is absolutely acceptable to provide just one implementation (though at this stage you did not replace the supplement, so you are still providing three). However, that one implementation needs to be of a quality that readers and users can have confidence in. Your simple statement that you are unable to meet the journal's standards doesn't excuse you from doing so. Further, the fact that it was not at all obvious to the reviewers that your code was correct highlights the need for proper tests. I note also that both reviewers also highlighted the need to produce initial conditions for a wider range of meshes. This is not particularly difficult, but it does require some care and attention to detail, and suitable testing. The authors statement that they are simply declining to do so smacks rather of an attempt to offload this work to the rest of the community. In short, it is fine to just provide a Matlab implementation - though given that that is the one proprietary platform out of the three candidates, it would be courteous to ensure that the code also runs correctly under Octave. However it's not fine to provide a half-baked implementation that fails to provide the range of functionality that a test-case user would reasonably need, and lacks its own testing.

As reviewer 2 states, this manuscript is an important contribution which should be published, but that doesn't exempt it from the need to address issues in both the text and the code.

Regards,

David

Dear Nathan,

Thank you very much for the revised manuscript, which the reviewer has now approved. Thanks for going the whole way with the Python module, I hope this will contribute to the take-up of this work. The only reason that this is "technical corrections" rather than "publish as is" is that I think the old version of the Python code is still in the supplement and I figure you probably want to fix that.

Regards,

David