AI Glossary / Graphs and symbolic AI / Intermediate
HTTPRange-14
Definition
HTTPRange-14 is the name of a historical World Wide Web Consortium Technical Architecture Group issue about what HTTP identifiers can identify and how clients distinguish a thing from information describing it. It became influential in Semantic Web and linked-data design, particularly in the use of separate identifiers, fragment identifiers, and HTTP 303 redirects.
What it means
A web identifier might be intended to name a document, a person, a physical object, or another resource. Retrieving a document about a person does not mean the person and the document are the same entity. That distinction matters when structured data attaches properties to identifiers.
For example, a person can have a birth date, while a profile document can have a publication date and file format. Using one identifier ambiguously can mix claims about those different subjects. HTTPRange-14 concerns that architectural interpretation; it is not an HTTP byte-range feature or an AI model parameter.
How it works
The 2005 TAG resolution described conventions for interpreting responses to HTTP GET requests. Under that advice, a successful 2xx response indicated an information resource; a 303 response allowed the original identifier to name any resource; and a 4xx response left the resource's nature undetermined.
The linked-data pattern often uses an identifier for a thing that redirects with 303 See Other to a document describing it. Another pattern uses a fragment identifier, with the base document supplying relevant descriptions. These patterns require clear data modeling. The historical resolution and later guidance should be read in their stated context, rather than turned into a universal claim that every redirect proves a particular ontological category.
Three practical examples
- A museum dataset distinguishes a sculpture's identifier from the web document describing its material and history.
- An organization publishes separate identifiers for a person and that person's profile document so document metadata does not become a claim about the person.
- A linked-data teaching example compares a fragment-based identifier with a 303-based pattern and traces the resource described by each document.
Think of it this way
Think of a library catalog record about a painting. The record can be revised or translated without changing which painting it describes. The analogy is limited because web identifiers and HTTP responses follow explicit technical conventions; a human reader's intuitive distinction does not by itself tell software which resource an identifier denotes.
Worked example
Use fictional example identifiers. Let https://example.org/id/sculpture-7 identify a sculpture, and https://example.org/doc/sculpture-7 identify its descriptive document. A client requests the first identifier. The server returns 303 See Other with the document identifier as the location to retrieve. The client then requests that document and receives a successful response containing the description.
The dataset can attach material = bronze to the sculpture identifier and document format = HTML to the document identifier. It should not infer that the sculpture itself has an HTML file format merely because its description is delivered that way.
Now revise the descriptive document to correct a spelling error. The document's modification metadata can change while the sculpture's identity remains the same. The example requires two requests in this simple uncached redirect trace, but that count is not a universal performance benchmark. Caching, content negotiation, and actual deployment details can change the request pattern. The identifiers are illustrative and do not refer to a real published collection.
Illustrative case study
In a fictional catalog project, Linden Museum discovers that a data import assigns a page's revision date as the creation date of the artwork it describes. The team traces the error to an ambiguous identifier mapping.
It separates artwork and document identifiers and documents the relationship between them. Reviewers then inspect which properties belong to which subject. The invented case illustrates entity modeling and metadata provenance, without claiming that adopting one URI pattern automatically corrects every data-integration problem.
Common misconceptions and mistakes
A 303 response does not by itself prove that the original identifier names a physical object. The historical resolution explicitly allowed any resource in that case. Nor does a 404 establish that the identified thing cannot exist; it concerns the request and available response, with the resource's nature left unresolved by that convention.
Do not confuse a representation with a description of a different entity, or treat every URI containing a fragment as automatically identifying a person or object. Context and the relevant representation semantics matter. The issue's name also has nothing to do with requesting byte ranges through HTTP Range headers.
Business use
The concept is useful background for knowledge graphs, linked-data publishing, and systems that combine records about real-world entities with documents about those entities. It encourages explicit identity choices so properties are assigned to the intended subject.
For a project, document what each identifier denotes, how descriptive documents are linked, and how identifiers remain stable when documents change. Check consumers' expectations and the applicable publishing guidance. Treat HTTPRange-14 as architectural and historical context, not a shortcut that replaces an ontology, provenance model, or carefully tested data interface.
People also ask
Is HTTPRange-14 an HTTP status code?
No. It is an issue identifier; HTTP 303 is one response code discussed in its resolution.
Does a 303 redirect mean the original resource is definitely not a document?
No. The original resolution says it could be any resource, so that stronger inference is not justified by 303 alone.
Why does the distinction matter for AI?
Knowledge-graph and retrieval systems can make incorrect connections if they confuse an entity with a document describing it.
Check your understanding
In the worked example, which identifier should receive the claim that the sculpture is bronze?
Answer: The sculpture identifier, https://example.org/id/sculpture-7. The document identifier describes a separate information resource and should receive document-specific metadata.
Related terms
- Representation
- Linked data
- RDF
- URI
- resource
- Semantic Web
- HTTP 303
- fragment identifier
- entity identity
Learn next
Read the original resolution alongside the W3C linked-data examples, paying attention to the distinction between any resource and a non-document resource. Sketch separate identifiers for one object and two documents about it, then assign each property to its correct subject.
More in Graphs and symbolic AI
- Hopscotch hashing - Hopscotch hashing is a hash-table technique that keeps each stored entry within a small neighborhood of its home bucket.
- Horizon effect - The horizon effect is a problem in limited-depth search where an important consequence lies beyond the search cutoff and is therefore misjudged.
- Horn clause - A Horn clause is a disjunction of literals containing at most one positive literal.
- Hub - A hub is a node that connects to other nodes in a way that makes it useful for navigating a network.
- Human-level AI - Human-level AI is a term for artificial intelligence whose abilities are comparable to human abilities under a specified comparison.
- Hungarian algorithm - The Hungarian algorithm solves a linear assignment problem by finding a one-to-one pairing with minimum total cost.
Sources
Sources: W3C TAG, original httpRange-14 resolution message: https://lists.w3.org/Archives/Public/www-tag/2005Jun/0039.html W3C, Cool URIs for the Semantic Web: https://www.w3.org/TR/cooluris/
Previous: Hough transform Next: Hub