Don't we use a URL to not only fetch a document on the Web but also identify it (well, at least I do!). However, this simple paradigm, when applied to certain other domains, may fail to deliver the desired result.
I wrote this little piece on how this crucial decision can limit design alternatives.