In terms of which is the “right/wrong” one, that’s probably up to you. I don’t use them, but from a quick perusal: neither seems to be actively maintained, documentation is sparse for one and non-existent for the other, and they both consist of only a single very short file of actual source code. I don’t know that I’d be excited to use either, but the registered one looks more like to actually load without drama in the first place.
Is there any “periodic culling” so that unmaintained packages are archived and do not insert themselves into Google or other searches? If not, might I suggest this as something worthy of consideration.