Change the section on Quoted Triples to Triple Terms.

[gkellogg]

• Updates the diagrams to use a reifier.
• These represent changes based on the minimal baseline

Fixes #81.

---

Preview | Diff

[domel]

Just a little comment on the SVG images. I see that everything is converted to paths. It seems a bit redundant. All the elements are simple primitives. The text, which is a path, also seems a bit exotic.

[gkellogg]

Just a little comment on the SVG images. I see that everything is converted to paths. It seems a bit redundant. All the elements are simple primitives. The text, which is a path, also seems a bit exotic.

This is simply the result of exporting the Google Doc image as SVG. Hand-coding the SVG might avoid this, but that's beyond my skill set. In any case, the Google Doc image is easy to edit, if necessary, and re-export the SVG.

[niklasl]

I simplified the old ones a while ago (just put them in a gist). I can make such versions of these new ones if you want to?

In any case I'd like to tweak them so that the arrow in the new ones go to the reified edge (in the current transparent version it looks like the reifier just links to the subject).

[gkellogg]

I simplified the old ones a while ago (just put them in a gist). I can make such versions of these new ones if you want to?

In any case I'd like to tweak them so that the arrow in the new ones go to the reified edge (in the current transparent version it looks like the reifier just links to the subject).

I show that _:reif01 is the subject for both the triple term and :bob, which essentially represents what the previous version with "Quoted Triples" does.

The examples in your gist don't use a reifier, which I thought was important for these examples. But, perhaps an example that represents something like the syntactic sugar << _:reif01 | <alice> :name "Alice" >> :accordingTo :bob might illustrate the intended usage better. Perhaps, but adding _:reif01 as a label to the grey box.

[niklasl]

Yes, my gist example is from last year, they do not reflect what we need. I just put them there to show an example of the primitive shapes asked for (see the sources).

[gkellogg]

I'd be happy to use simplified versions of the diagrams. Looking at the source for each, they would seem easy enough to maintain by hand. Feel free to push a commit with newer versions of these diagrams, although we should probably continue to reference (in the comment) the Google Doc version as an alternative representation.

[niklasl]

I don't seem to have permission to push to this repo. Meanwhile I put proposed new versions in another gist. Feel free to grab those copies if the permissions are hard to sort out.

(I put the reifier in the box, akin to what you suggested, but showing as a real node (to clearly show that it is). I also kept the rdf:reifies relation label, moved to the border of the box, since the box "encircles" the relationship which is reified. Remove the rdf:reifies if we want the implicit version it.)

[gkellogg]

I don't seem to have permission to push to this repo. Meanwhile I put proposed new versions in another gist. Feel free to grab those copies if the permissions are hard to sort out.

Thanks, those are great. I'll push them up. Probably still needs a paragraph or two on the rationale for using reifiers.

[gkellogg]

This could be the place to define a term such as "reified triple term" to describe the syntactic structure that optionally includes a "reifier" and a "triple term".

[hartig]

Here are a few edit suggestions.

[hartig]

Given that this section is non-normative, I wonder whether it makes sense to put a definition here (<dfn>reifier</dfn>) ?

[gkellogg]

As a reifier is really just a concept, and not a component of the abstract syntax itself, it seems to me that a non-normative definition is fine. I think we also need a concept for "reified triple term", or whatever we come up with. In Turtle, both "reifier" and "reified triple term" will be normatively defined as part of the grammar, and it may be elsewhere, so having the source definition in RDF Concepts makes sense.

But, I'm open to suggestions to handle this better.

[hartig]

Makes sense.