Chairs' decision on triple terms in the subject position

After careful consideration of the results of the recent vote [1] and comments from WG members, the chairs have decided in favor of restricting triple terms in the object position. Only 25% of the voters prefer to not have this restriction, and 85% of the voters “can live with” the restriction.

Specifically, the following reasons/arguments were considered:

a) Complexity and “mis-modeling”: RDF has always been accused of being “academic” and hard to understand. Whether this is true or not, adding to the complexity does not seem like a clever idea. Specifically, the choice of an abstract vs. a concrete triple in the subject position will be something many people will get wrong, leading to all kinds of RDF that does not actually represent what the authors believe it does. Most, if not all, of the collected use cases where the triple term is in the subject position are in fact better represented with a reifier of a triple term in that position. Interestingly, a comment was submitted in favor of allowing triple terms as subjects that actually argues the complexity case, pointing out that mistakes would be made.

b) Symmetry: Proponents of triple terms in the subject position often state symmetry as an argument. But RDF does not provide for symmetry today (in its metamodel or abstract syntax, specifically wrt. literals). Merely wanting symmetry for symmetry’s sake is not a convincing argument, and as observed above, the use cases that seem to require symmetry in fact do not have this requirement.

c) Future: The prohibition of triple terms in the subject position is a “two-way door” decision, meaning that if, in the future, we find that a compelling reason to not prohibit has emerged, the decision can be undone without harm to (then) existing RDF. The same cannot be said about the converse: Allowing triple terms as subjects now will lead to RDF that would “break” were we to reverse that decision (a comment about early adopters was made that supports this observation).

d) Comparison with other “symbol systems”: The chairs failed to understand what the relevance of these submitted arguments was to the immediate question at hand.

[1] https://www.w3.org/2002/09/wbs/139681/2025-rdf-star-tripleterms-subject/results



--
Dr. Ora Lassila
Principal Technologist, Amazon Neptune