What is the difference between a URI, a URL and a URN?

Question

People talk about URLs, URIs, and URNs as if they\'re different things, but they look the same to the naked eye.

W

Answer 1

The best (technical) summary imo is this one

IRI, URI, URL, URN and their differences from Jan Martin Keil:

IRI, URI, URL, URN and their differences

Everybody dealing with the Semantic Web repeatedly comes across the terms IRI, URI, URL and URN. Nevertheless, I frequently observe that there is some confusion about their exact meaning. And, of course, others noticed that as well (see e.g. RFC3305 or search on Google). To be honest, I even was confused myself at the outset. But actually the issue is not that complex. Let’s have a look on the definitions of the mentioned terms to see what the differences are:

URI

A Uniform Resource Identifier is a compact sequence of characters that identifies an abstract or physical resource. The set of characters is limited to US-ASCII excluding some reserved characters. Characters outside the set of allowed characters can be represented using Percent-Encoding. A URI can be used as a locator, a name, or both. If a URI is a locator, it describes a resource’s primary access mechanism. If a URI is a name, it identifies a resource by giving it a unique name. The exact specifications of syntax and semantics of a URI depend on the used Scheme that is defined by the characters before the first colon. [RFC3986]

URN

A Uniform Resource Name is a URI in the scheme urn intended to serve as persistent, location-independent, resource identifier. Historically, the term also referred to any URI. [RFC3986] A URN consists of a Namespace Identifier (NID) and a Namespace Specific String (NSS): urn:: The syntax and semantics of the NSS is specific specific for each NID. Beside the registered NIDs, there exist several more NIDs, that did not go through the official registration process. [RFC2141]

URL

A Uniform Resource Locator is a URI that, in addition to identifying a resource, provides a means of locating the resource by describing its primary access mechanism [RFC3986]. As there is no exact definition of URL by means of a set of Schemes, "URL is a useful but informal concept", usually referring to a subset of URIs that do not contain URNs [RFC3305].

IRI

An Internationalized Resource Identifier is defined similarly to a URI, but the character set is extended to the Universal Coded Character Set. Therefore, it can contain any Latin and non Latin characters except the reserved characters. Instead of extending the definition of URI, the term IRI was introduced to allow for a clear distinction and avoid incompatibilities. IRIs are meant to replace URIs in identifying resources in situations where the Universal Coded Character Set is supported. By definition, every URI is an IRI. Furthermore, there is a defined surjective mapping of IRIs to URIs: Every IRI can be mapped to exactly one URI, but different IRIs might map to the same URI. Therefore, the conversion back from a URI to an IRI may not produce the original IRI. [RFC3987]

Summarizing we can say:

IRI is a superset of URI (IRI ⊃ URI)
URI is a superset of URL (URI ⊃ URL)
URI is a superset of URN (URI ⊃ URN)
URL and URN are disjoint (URL ∩ URN = ∅)

Conclusions for Semantic Web Issues

RDF explicitly allows to use IRIs to name entities [RFC3987]. This means that we can use almost every character in entity names. On the other hand, we often have to deal with early state software. Thus, it is not unlikely to run into problems using non ASCII characters. Therefore, I suggest to avoid non URI names for entities and recommend to use http URIs [LINKED-DATA]. To put it briefly: only use URLs to name your entities. Of course, we can refer to existing entities named by a URN. However, we should avoid to newly create this kind of identifiers.