Mental model for resource lifecycle

[kjetilk]

Since CS possess a Physics Envy on certain occasions, we might as well go with it to form a mental model. ;-)

I'm of course referring to the concept of atomic operations, and then what constitutes an atom, because atoms have substructure.

Atoms have a nucleus, which again consists of protons and neutrons. In Solid's first iteration, I think we should study the Hydrogen atom much like Niels Bohr did more than a hundred years ago. Thus, we have a single proton. That proton is "the resource", because it positively charged and can attract... whatever. :-)

The neutrons are the resources that lives and dies with the proton, i.e. the .acl (#58) and the .meta (#63). That is, operations on these resources are not only atomic, their lifecycle are completely tied together, and so operations on the proton may cause side effects on the neutron resources.

Beyond the nucleus, there are electrons, but as free hydride anions exist only under extreme conditions, we should restrict ourselves to only one electron per nucleus for now. The electron is typically the server-controlled metadata (#65), i.e. data that should be manipulated atomically with the nucleus, but when the nucleus is destroyed, the electron may persist. It thus to be expected that the pod graphs will be a negatively charged plasma, with free electrons representing data about long-gone resources.

Slightly tongue-in-cheek, but I think this mental model captures some essentials :-)

But the big question, does it capture consensus around resource lifecycle?

[TallTed]

Interesting. I think I like it. I'm curious to read any negative thoughts about it... with which I may well also agree.

[csarven]

Depends. Do we need to unlock quantum mechanics or string theory?

[kjetilk]

Not yet, but who knows...? There is a lot of good humour around. I think a little subatomic theory goes a long way! ;-)