Canonical form - avoid redundancy

[pchampin]

Just an idea:
if w3c/rdf-n-quads#17 is merged into N-Quads,
the "Canonical form" section of N-Triples should probably simply refer to N-Quads and state "canonical N-Triples is the corresponding subset of canonical N-Quads" -- possibly copy-pasting the relevant text in a non-normative subsection.

We don't want to have duplicate notmative text that may eventually diverge out-of-sync.

[gkellogg]

Yes, that seems reasonable. We just need to preserve any anchors that might be used in external references.

The N-Quads section on canonicalization does not directly reference the N-Triples section.

[gkellogg]

Thinking again, we have a dependency chain, in that N-Quads is based on N-Triples and many aspects of N-Quads are defined in terms of N-Triples. Inverting the dependency to have N-Triples depend on N-Quads for canonicalization may not be worth it, given that the text is constrained to a single section.

[pchampin]

Makes sense.
An alternative way may be to have a normative subsection describing how to extend canonical N-triples into canonical N-quads, then a non-normative section copying and extending the text from N-triples, just in order to have a self-sufficient description.

[domel]

I agree with @gkellogg. Perhaps it is enough for the grammar to have rules named the same way.