Clean up my mess...

[strengejacke]

@DominiqueMakowski Due to this unfortunate setup of Github, cloud sync and two PCs for working at packages I tried to commit via webinterface, which completely messed up the file structure (with R files put in the root folder).

I now cloned the repo here, cleaned up everything and commited via GitHub app again (hope that my sync understands that it doesn't have revert the sync from GitHub after syncing from cloud at home...).

Please check the spelling, some of the performance_* functions have lower case, other upper case R.

[strengejacke]

I merged the PR, so the r2() branch now is merged into master. But that did not affect your methods (except that I have replaced the code in performance_loo() with the call to the loo()-function, which also contained the same code).

[DominiqueMakowski]

Ok let us unify this a bit :)

What about the prefix "perf_*" instead of "performance_"? This could be short enough (perf_r2 vs. r2?) and consistent with the implementation of other methods (perf_aic, perf_loo, perf_superduperindex)?

[DominiqueMakowski]

And yes, we should keep all the names in lower cases, my bad :)

[strengejacke]

It depends on the scope of the functions. If aic() calculates the AIC for new models, so it enhances other packages, I would keep the R stats name (AIC() I think), so users can use the generic function call.

r2() isn't taken yet, so we don't have to care about existing generics, but then we would break with our design-choice if one function is called AIC() and others perf_r2().

If, however, perf_aic() does something different that AIC() we could think about a naming convention we want.

[DominiqueMakowski]

hum, I think you've managed to convince me again :)

Let's remove all the performance_* in favour of direct metrics name calls