I see that the genome contains large regions of repeating sequences called interspersed or dispersed elements. The long dispersed elements (LINES) such as LINE-1, can reach up to 6-8 kb in length.
I'm wondering, given the amount of repetition that takes place in these regions, wouldn't a system where a pointer (such as in computer programming) existed be more efficient? For example, instead of including a LINE, include a unique 5 base pair sequence acting as a pointer to that LINE. A separate chromosome (containing one copy of each LINE) would then be read at the correct position once the pointer was read.
Do you think that given enough time in evolution, such a system would be more sustainable? Sort of like using a proper functional language instead of Assembler in computer programming, where f(x) can be defined once, and accessed via pointers, instead of being repeated many times?