[IMPORTANT] Necessary revision of the sugar proposal

After a lot of thought, I believe that in order to pursue a “syntactic sugar” proposal, we need to fix the original proposal (as documented by pfps), as follows:
https://github.com/w3c/rdf-star-wg/wiki/Revision-of-the-sugar-proposal


[it is in a referrable document, so that it will not be embedded in emails]

[BTW, I do not necessarily endorse this proposal, but in order to make it work, I believe it should be changed as I am proposing]

[If we do not fix it, we will continue forever the discussion about well-formedness, since the original proposal confuses the representative of the triple (which is indeed unique) and its “occurrences".]

I paste the proposal below.

## Named occurrences of triples as RDF reification

This syntax is simply a shorthand for RDF reification (or some variation thereof).
In other words, any occurrence in the graph of a triple

    << :e | :s :p :o >> :p1 :o1 .

would be syntactic sugar for the following triples:

    :e rdf-star:is-reification-of _:b
    _:b rdf:subject :s .
    _:b rdf:predicate :p .
    _:b rdf:object :o .
    :e :p1 :o1 .

with `_:b` a fresh new blank node, unique within the RDF graph.
The abstract syntax would not need to be extended.

It is necessary to add a notion of syntactically "well-formed" RDF:

**Definition**: An RDF graph is reification well-formed iff:

  1.  The object of each `rdf-star:is-reification-of` triple is a blank node.
  2.  Any blank node appearing as object of a `rdf-star:is-reification-of` triple occurs only in that triple and exactly once as subject of a `rdf:subject` triple, a `rdf:predicate` triple, and a `rdf:object` triple.

The notion of well-formedness does not impact reification-star-free RDF 1.1 graphs, and so it is fully backwards compatible.