Earlier I made a program to compress a list into a specific string format or decompress from that specific string format to a list. The compression is lossless and relies on the observation that a signed 14 digit number fits in 48-bits, whereas TI uses 8 bytes to hold the sign and the 14 digits. So with 6 bytes plus the mantissa, a Floating Point number can be compress to 7 bytes without losing any data. Usually there are more efficient ways to compress floating point numbers under certain conditions (for example, if there are only integers from 0 to 65535, you can compress each element to two bytes pretty quickly). However, this only makes the condition that it will not read complex lists. If I change the format of how the data is stored, though, I can make it recognise matrices, lists, complex lists, et cetera at the cost of one or two bytes (two for matrices)