We don’t need to store data twice...

fullFileObject

checksum

oovArray(all the file’s contents….)

Then you try to store another file (might be the same!)

look at the checksum

do we have it already?

no? make a fullFileObject

yes? Make a smallFileObject

smallFileObject

checksum

ooRef the_realfileObject

This is the current plan for binary data which we cannot parse (DSP, FPGA, ARCNET).

Works for ASCII files as well, but we might be better off storing the meaning rather than the words.