URLs are only really uniform not necessarily unique. The most obvious case is something like `http://localhost/file.txt`, where the resource being located is almost sure to be different on every single "localhost," but this is true of any server. The pointer is neither unique nor has any true guarantee that the resource being pointed to can or should be considered unique for any given context. It is merely unique within its calling context, which, when dealing with URLs, is often anywhere on the web where DNS responds.
Even if you presume that the resource being located is in fact a unique resource in the general case, the "unique resource" may not be unique in the way that you presume. Some URLs at a given location are intended to be idempontent and cache-able, others are not, and many are time-limited forwarders. And there's no guarantee or even expectation that two identical forwarding URL's will resolve to the same location; it may well be network-topography dependent.
Even if you presume that the resource being located is in fact a unique resource in the general case, the "unique resource" may not be unique in the way that you presume. Some URLs at a given location are intended to be idempontent and cache-able, others are not, and many are time-limited forwarders. And there's no guarantee or even expectation that two identical forwarding URL's will resolve to the same location; it may well be network-topography dependent.