NamespaceHijacking

From W3C Wiki
Jump to: navigation, search

Mostly refactored into NamespaceSquatting and NamespaceDistortion.

Example below still needs to migrate, or something.


For now, this page raises far more questions than it answers. But they are important questions for people who think they understand the role of URIs in the Semantic Web. Please add comments which might interupt the flow of this text to the end, in PPR:ThreadMode.

Example Scenario

A group of student enthusiasts reads about RDF and OWL. They decide to use these new Semantic Web technologies in their ongoing project to catalog and describe the social cliques (and social networking) in their school.

They end up writing RDF/N3 like:


@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>.
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix c: <http://www.w3.org/2002/02/30/7197/10888/4828/clique-ontology#>

c:Student a rdfs:Class.
c:Jock rdfs:subClassOf c:Student,
       rdfs:comment "Typically an athlete; always makes fun of Stoners".
c:Stoner rdfs:subClassOf c:Student,
       rdfs:comment "Typically fond of Marijuana; always makes fun of Jocks".
       owl:disjointWith c:Jock.


Social commentary aside, the thing to notice here is the "c:" namespace.

It's conceivable this vocabulary could catch on and be more popular than FOAF or DC in a matter of months.

What Harm Is Done?

Is there anything wrong with them doing this? What harmful effects might it have? (it could be done more than once by independent parties; then a third party might want to use both of those bits of work, but they use the same URI differently. HTTP service is about coordination as much as browsability) Is there a W3C Recommendation that tells them not to do this? Is there a W3C Recommendation that tells other people or software what to do if they see this? Should there be?

Clearly http://www.w3.org/2002/02/30/7197/10888/4828/clique-ontology#Stoner is not a great URI (as in GoodURIs) since it's not Browsable. But for people who have no option of making a Persistent, Browsable URI (there are no such people; purl service is free), maybe that's good enough. Maybe it's even better than a Browsable one where the institutional committment to persistence is potentially malicious (where people actively distrust the people who will control the web server in the future).

Terminology

This name, "Namespace Hijacking", suggests this practice is antisocial. It could also be "Namespace Squating", which suggests that the practice is neutral on society but antithetical to "property" rights. More neutral still is "Unauthorized Third Party Namespace Use", but that name is laden with other assumptions. Is there some "first party" which has special rights to use a namespace? What does it mean to "use" a namespace? Exactly which uses of rdf:type ought to require the consent of W3C?

Discussion Elsewhere


I don't understand what the problem is.

Is it the use of the letter "c," as if the letter "c" is "used up?"

I don't understand that- isn't "c" just a local shorthand for http://www.w3.org/2002/02/30/7197/10888/4828/clique-ontology# in this case?

And so, doesn't it not matter, since if someone else uses "c" differently, it'll just substitute differently in their document, since they bound it to some other URL in their own document?

-- LionKimbro DateTime(2004-06-14T02:30:16)