search

pi-base / viewer

Web UI for the π-Base
https://topology.pi-base.org
MIT License
9 stars 3 forks source link

Use slugs / ids for routes #10

Closed jamesdabbs closed 7 years ago

jamesdabbs commented 7 years ago

Should probably redirect to a standard UID-route but could accept slugs or names where possible.

StevenClontz commented 7 years ago

Re-emphasizing that I think UIDs should be the primary mechanism for just about everything we do. My hope is that our UIDs become the standard way to reference spaces/properties/theorems unambiguously.

jamesdabbs commented 7 years ago

Oh agreed. Right now we're using /spaces/Finite%20Discrete%20Topology and I'm proposing /spaces/S000001 (which is the only thing we ever link to) or /spaces/finite-discrete-topology (for user convenience).

On reflection - I don't think that's actually any more convenient, but the meta-question there is: where do we want to balance standardizing on ids versus user convenience, if there is a conflict between the two? (Ruminating on editing procedures mostly here)

StevenClontz commented 7 years ago

User convenience is fairly important; especially in the viewer, but also in the data repo. But I'm not seeing a scenario where having a slug is more convenient than a UID for editing, at least in the bigger picture. I'm shy about using slugs in any important way because sometimes names are in conflict (Gruenhage vs Gartside monotonically paracompact) and sometimes things are simply unnamed in the literature (I made up "Talagrand's example" in my mock PR for an example of a Corson but not Eberlein compact). In either case, the popular name may change, so standardizing slugs seems like more of a headache long-term (we can edit the name of something easily without having to go through and edit every time a slug is referenced).