> Geez. I meant the concept of a cover vs the concept of an original. A cover is ultimately a performance, and is different from writing a new song. This is so well understood within the music community that a song even has separate copyrights for the song itself and a recording of its performance (even by the songwriter).

If only it were that simple. There are constantly cases bought to court about similarities in one persons work to another artists. Then you have other issues around what constitutes a derivative work. And so many original songs sample other artists songs and pay them royalties, that's not a cover either.

I think what you're trying to highlight is writing credits vs performance. Which is a lot easier to define. However even here, plenty of disputes still happen.

> But it is strange to think of a reimplementation of a piece of software that one might acquire and use easily; it doesn't really fit the concept of performance.

The real problem with these analogies is that you're comparing something consumable with something reuable. But I accept the point of an analogy isn't precision.

> I am in fact a fan of reimplementing things in order to figure out how they work, but I don't expect my implementation to have any utility beyond the pedagogical value I got from doing it, unless it is in some way different from what exists already.

Would emulation fall into this category? You're building software to run something exactly as it would run elsewhere - a reimplementation. The motive differs (to run on different hardware) but that's not a property of the product itself.

Which comes back to my earlier post: you're talking about the merit of replication without discussing motives behind it. In your latest comment you say "to figure out how they work" and that's another great example of a motive that brings value to replication.