Hi,
Il 20/04/22 16:02, Matteo Bruni ha scritto:
I was a bit confused by the major_size() / minor_size() naming, together with the coordinate and dimension flipping that we always do with matrices, but this seems to be correct.
It might be just a "me" thing but maybe names like vector_size() and vector_count() would be potentially more obvious?
While I also always need to think which one is the "major" or the "minor" size, I don't find your proposals easier to understand. The words "size" and "count" have nearly the same meaning, so it's still not really clear which one applies to the vector of vectors as a whole, and which one to each single inner vector. As you say, it's probably a very personal thing.
All in all, I still think that "major size" and "minor size" is a better choice, because at least they're more or less standard terminology. So while with both proposals you still need to think which is which, with mine at least you're immediately on the track to understand what is the point you need to think about, while I don't feel that "vector size" and "vector count" really bring you in that direction.
We should probably update the comment above split_copy() to include split_matrix_copies(). Or maybe replace the first sentence altogether with "Copy an element of a complex variable." or something.
Ok, will do in a follow up.
Adding a test with a non-square matrix might be a good idea, if anything to give a bit more reassurance that we're using the right dimensions.
Ok, makes sense.
Giovanni.