- From: Patrick Stickler <patrick.stickler@nokia.com>
- Date: Tue, 22 Jan 2002 09:20:43 +0200
- To: ext Mark Baker <distobj@acm.org>
On 2002-01-21 20:35, "ext Mark Baker" <distobj@acm.org> wrote: > Patrick, I'm trying hard to understand your position here. In as few > words as possible, what's your definition of an URL? > > MB Though I probably will need to provide a more lengthy explanation... In a nutshell: I view a URL as a direct point of access to a digital resource, such that "interaction" with that point of access (whether that be retrieval, storage, query, etc.) involves the interchange of content between two systems -- a server which hosts that point of access and a client which connects to that point of access. Thus, a URL (e.g. 'http:') used to denote an abstract or non-digital resource is not a "proper" or "reasonable" URL because it can never meet the above expectations of accessibility. A URN can be viewed as an indirect point of access to a digital resource, such that -- as with a URL -- "interaction" with that point of access involves the interchange of content. Thus, a URN used to denote an abstract or non-digital resource is not a proper or reasonable URN because it can never meet the expectations of accessibility. In the case of a URN, the direct point of access must be determined based on context or other external specification, but otherwise, the interaction of a client with a URN denoted resource is analogous to interaction with a URL denoted resource. This does not mean that every URL or URN must always and forever resolve in every context for all time to some digital resource -- only that the original and intended purpose of every minted URL and URN *is* to resolve to a digital resource. In my I-D draft-pstickler-uri-taxonomy-00, I define a set of primary distinctions between URI classes based on the denotation of authorities/agencies in the URI scheme, such that, for a URL both the minting authority and resolution (access) agency are defined by the URL itself. For a URN, only the minting authority is defined, leaving the resolution agency open to contextualization. Without the distinction between directly resolvable (URL), indirectly resolvable (URN), and non-resolvable (URP) inherent in the semantics of each URI Scheme (and ideally, though not necessarily, organized into a taxonomy of URI classes where significant intersection of semantics occurs) applications are unnable to automatically differentiate between a true access error and an "intentional" access error -- the latter being the case in the use of URLs (or URNs) for denotation of abstract and non-digital resources. With regards to resolution, URLs and URNs are closely related, in comparison to URPs -- as both URLs and URNs resolve to digital resources (where URPs do not), and the primary difference between them is simply whether access/resolution is direct or indirect based on whether the resolution agency is specified in the URI itself. With regards to named authorities/agencies, URNs and URTs are closely related, with the primary difference between them being whether they denote digital resources or non-digital resources (resolvable versus non-resolvable) -- with both naming the minting authority but not any resolution agency. URVs are unique in that they name neither the minting authority nor resolution agency (the latter of course because they are URPs and non-resolvable) -- such that anyone can mint them so long as they conform to the lexical requirements for the URV scheme, and they become completely "free" and "unbounded" digital resources without reference nor tie to creator or resolver. Any means to derive a source for a given URV (e.g. analyzing a UUID, etc.) is an artifact of a given URV scheme and not an inherent part of its interpretation. Does that help clarify my understanding of URL, URN, etc.? Cheers, Patrick -- Patrick Stickler Phone: +358 50 483 9453 Senior Research Scientist Fax: +358 7180 35409 Nokia Research Center Email: patrick.stickler@nokia.com
Received on Tuesday, 22 January 2002 02:20:11 UTC