> it's much easier to keep IDs stable over time ... [otherwise] you'd have to hack your slug-generation algorithm to ensure the generated slugs to be unique[.]

> it's much easier to keep IDs stable over time ... [otherwise] you'd have to hack your slug-generation algorithm to ensure the generated slugs to be unique[.] ... [A]voiding cruft ... [is] in direct tension with the point about keeping them stable given that "meaningful" non-cruft tends to change over time! Many applications have meaningful unique identifiers, creating a unique slug is not always a "hack". Any app that displays events that occurred in the past often have unique identifiers. Examples include software logs, email, government records, financial transactions, i.e., anything that displays past events. These records are inherently "write-once", and the primary key can be something more meaningful than an ID. These may be just a subset of apps out there, but in these cases there is a strong argument for removing any ID in a URL.