Why don't use "sparse" option when copying/extracting image files

As I understand, opennebula copy image files (loop devices) when creating images and deploying VMs.

The scrip to copy image file seems to be

It uses “cat”, “tar” commands to make image files in datastores.
Then why don’t use “sparse” option?
I believe with this, we can save much more spaces.
(I know with sparse files, it will be more complicated to know the “true” free storage space, but still I want to save resources.)