search

w3c / rdf-star

RDF-star specification
https://w3c.github.io/rdf-star/
Other
119 stars 23 forks source link

Tidy up intro #171

Closed pchampin closed 3 years ago

pchampin commented 3 years ago

addresses #94 and #165

Preview | Diff

rat10 commented 3 years ago

[sorry for not being able to suggest the following change via proper GIT commands!]

In section 1.2, at the end of the second paragraph, I propose to replace

The concept of reification has always been part of RDF, but expressing it in RDF concrete syntaxes such as Turtle, N-Triples, and RDF/XML, as well as processing or querying it with SPARQL, has been verbose and cumbersome.

with

The concept of reification has always been part of RDF, but expressing it in RDF concrete syntaxes such as Turtle and N-Triples, as well as processing or querying it with SPARQL, has been verbose and cumbersome. RDF/XML provides syntactic sugar for reification by means of an :ID attribute but for other reasons has become mostly a legacy format.

As Jerven Bollmann mentioned, the rdf:ID attribute is the reason why they still use RDF/XML in Uniprot.

rat10 commented 3 years ago

I would have been fine with removing "but for other reasons has become mostly a legacy format" as @afs suggested but not with removing the reference to RDF/XML itself. With that sub sentence I didn't want to hurt anyones feelings but provide some sort of reference to the fact that RDF/XML is not anymore the go-to format for all things RDF that it has once been, so it's practical importance is limited nowadays. But as I said, I'm fine with removing any such statement or rewording it in the nicest possible way.

What is important however is the fact that RDF/XML is the only standardized syntax that provides syntactic sugar for RDF reification in form of the ID attribute. Providing a less cumbersome alternative to the verbose standard reification quadlet was not too long ago - before the proposed semantics was introduced - the central value proposition of RDF-star. So RDF/XML is of very high relevance in the context of RDF-star, and much higher than all the other syntaxes mentioned.

Also one of the most concrete use cases ilisted for RDF-star is that of UniProt which would like to get away from RDF/XML but makes heavy use of the ID attribute to sweeten reification syntactically. UniProt, as expressed by Jerven Bolleman would like to switch to RDF-star but would also need its model-theoretic semantics to support referential transparency.

So this is indeed an interesting topic with very relevant aspects and deleting it from the report entirely because a sub sentence might hurt somebodies feelings is a very debatable editorial decision. RDF/XML is now only mentioned once, in section 3.5 "Other Concrete Syntaxes" in the context of extendendig those to support RDF-star.

In academic papers an important section is always that of "Related Work". I don't know for certain what the customs are w.r.t. W3C Recommendations or Community Reports but I would expect some of the same attitude towards openness and constructive discussion and dialogue. That certainly includes mentioning the closest related approach to the problem.