lib/systems/parse: adjust isCompatible description

[siraben]

Motivation for this change

Converting the isCompatible definition into a graph shows that CPUs with multiple modes of endianness are indeed isomorphic.

Stating that CPUs and the isCompatible relation forms a category (or preorder) is correct but overtly technical. We can state it more clearly for readers unfamiliar with mathematics while retaining some keywords to be useful to technical readers.

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.

[siraben]

@Ericson2314 could you take a look?

[r-burns]

As a reader unfamiliar with mathematics, I can attest that this description is much more understandable ;)

You may be interested in NixOS/nix#2595, which touches on the nuance of platform compatibility. It can in some cases depend on more specific details of the OS or processor.

[siraben]

@Ericson2314 good to merge?

[siraben]

pinging @jonringer

[rmcgibbo]

As a reader unfamiliar with mathematics, I can attest that this description is much more understandable ;)

This is exactly my sentiment as well. 👍

[Ericson2314]

I agree "category" is over kill, but i rather call it a "partial order" than mere relation. I think "partial orders" (or "preorders) are pedestrian enough, and the vast majority of the relations that arise during packaging are partial orders, so that's the most useful concept to nudge at.

Otherwise looks good; those are partial order axioms (and we might link wikipedia) but it's fine to include them.

[siraben]

I agree "category" is over kill, but i rather call it a "partial order" than mere relation. I think "partial orders" (or "preorders) are pedestrian enough, and the vast majority of the relations that arise during packaging are partial orders, so that's the most useful concept to nudge at.

Ok. I was hesistant to add partial order because it might not be familiar to many programmers. I can find a way to fit it in somewhere.

[jonringer]

I think "partial orders" (or "preorders) are pedestrian enough

I think this is only pedestrian in circles like haskell, which has a very large mathematical influence. I'm pretty sure if I asked my co-workers over the past 6 years, I would be very hard pressed to find someone which definitively knew what that was.

Either way, I think @siraben 's wording is much more clearer, and the examples help make it much more intuitive.

[Ericson2314]

@jonringer agreed is an obscure math thing, but aren't relations also obscure? I would have thought they are about as obscure so partial order is no worse in that record.

Anyways I don't want to be the one bikeshedding this to death. It is an improvement already I agree.

[siraben]

I assume most Nix users are aware of the concept of a relation (transitive closure is already an integral concept in the Nix store at least).

[mohe2015]

I assume most Nix users are aware of the concept of a relation (transitive closure is already an integral concept in the Nix store at least).

I disagree. In Germany you usually learn about relations in university. And also I knew what is meant by a transitive closure but got it to know under another name (recursive dependencies or so)

[mohe2015]

I think we should be as inclusive as possible and if it's possible to document this otherwise (or both ways) I think we should do that.

[mohe2015]

I checked the diff and I think it's fine though

[siraben]

I disagree. In Germany you usually learn about relations in university. And also I knew what is meant by a transitive closure but got it to know under another name (recursive dependencies or so)

Fair enough, so I mentioned relations but also gave plain English explanations of the desired properties of isCompatible. I'm going to leave it as is, seeing that everyone seem to agree it's an improvement.

[jonringer]

otherwise (or both ways)

I would also be fine with having it extended. It's unfortunate that many mathematical concepts have many names depending on what specialty it derived from.