I was on a while back about using IPFS or some other CAS to cache
computation and results. I iterated on this concept with metadata, then
data, then just files as a fundamental concept. It became abundantly
clear that how we use, think, and need from and about files are not what
computer files are. They are currently unstructured inconsistent
blobs. We want them to be secure, distributed, and structured. In the
attached, I attempt to create the concepts to allow such a transition
from files as unstructured blobs to how we actually want and need them
for all use cases. The grammar is incomplete, and even wrong in some
ways. But I can't just deliver the final correct and complete paper
without feedback.
At the end, I want a
understandable, readable, meaningful, complete paper which is of
publication quality and actionable for any skilled developer.
So
I ask -- edit, review, critique, criticize, shred, and otherwise pick
out everything wrong in every way that can be done. Preferably
politely.
Unifying File Formats.pdf
Unifying File Formats.tex