Meeting minutes
<TallTed> whatever we put here will be captured in a log, if not proper minutes...
<TallTed> I cannot scribe, but if someone else can, that would be a win.
discussion about the terminology in order to reach a shared agreement
niklasl: does anyone want to change the agenda?
… we need to iterate on the terminology
pchampin: my suggestion is to go through the terminology
… for things that turn out to be too controversial
… let's discuss them a bit, but not get stuck
gkellogg: which doc exactly are we working on?
niklasl: I cannot edit the wiki
… or the docs in the repo. Therefore, I created a fork of the repo
… I try to share my screen with my editor
… looking at the terminology doc now
AndyS: we need only/primarily the definitions in this doc, no need for extensive explanations or justifications
niklasl: think we need to define "abstract triple" / "triple type"
gkellogg: isn't that defined in RDF Concept?
<pfps> From concepts: An RDF triple encodes a statement—a simple logical expression, or claim about the world
niklasl: can I remove the section about "RDF Triple"?
tl: no wait
… reader should understand the difference between triple and occurrence
… also, I want to rename "occurrence" to "instance"
olaf: For me, the term RDF Triple is basically a Triple Type or Abstract Triple.
<pchampin> +1
pfps: we should not change the terminology of the RDF universe
<doerthe> +1 to pfps
pchampin: +1 to pfps
… regarding "instance" vs "occurrence"
<AndyS> +1 to pfps
pchampin: let's keep that for a different entry in the terminology
<pfps> From RDF Semantics: The subject of a reification is intended to denote a concrete realization of an RDF triple, such as a document in a surface syntax, rather than a triple considered as an abstract object. This supports use cases where properties such as dates of composition or provenance information are applied to the reified triple, which are meaningful only when thought of as denoting a particular instance or token of a triple.
enrico: regarding "instance" vs "occurrence", that's a delicate topic
… we should first understand what they are and then pick a suitable name
niklasl: We still need a term for something that we say all the time now.
… perhaps "statement" ?
<pfps> My view is that "instance" is a poor word to use. "occurence" or "token" are much better.
AndyS: What is the audience for this material here?
… I thought the audience was the wider WG, and eventually the readers of the specs
… The wider WG may not have read all the ground material (mailing list threds etc)
<pfps> My view is that the audience for the current document is the working group. This document may be the genesis of something for wider distribution, but any "missing" stuff for that audience can be added later.
<tl> +1 to peter
pchampin: see what pfps wrote in the chat
niklasl: yes the spec about RDF reification talks about tokens
… this was the gist that I was trying to capture when I wrote this down
… the spec singles out "realization", "token", ...
… I would like to clean up the definitions of these to make things clearer
pchampin: I agree that ultimately these definitions may be material for the specs
… If we think the definitions we have from RDF 1.1 are not satisfactory, we can clarify them
… but we should reuse what we already have in the specs
… may for our internal material, we should also have a list of terms that we think are not so suitable
… including synonyms that are less good for the words that are better
niklasl: make a SKOS vocab for it? ;-)
pchampin: Regarding the term "RDF triple" let's quote the definition from RDF concepts
… because I don't think anyone wants to change that definition / the use of that term
niklasl: How to mention the "less good" synonyms?
tl: no objection on "RDF triple"
… we should move to the difficult ones, in particular "occurrence"
pchampin: next one is "RDF statement"
<AndyS> "An RDF statement is the statement made by a token of an RDF triple." : https://
niklasl: My thought when adding "RDF statement" was that this term shows up in spec text related to reification
<AndyS> https://
gkellogg: The NQuads grammar uses "statement"
<pchampin> and quad is not the appropriate term either, because it does not always have 4 components :->
gkellogg: "triple" is not the appropriate term there because it is not about triples but quads
<AndyS> https://
gkellogg: but it would mirror the grammar rule of "triple" in a single-graph serialization format
<AndyS> https://
gkellogg: we could say that this word "statement" has a meaning different from the notion of "statement" as defined in RF Concepts
AndyS: "statement" is mentioned in RDF Concepts and in RDF Schema
niklasl: looking at RDF concepts now
… I will use that definition of "RDF statement" from RDF Concepts
<TallTed> a quad always has 4 elements. sometimes the 4th is implicit.
pchampin: The links in that doc should point to the current 1.2 drafts
… because these are supposed to capture the current consensus of the WG
<pchampin> TallTed: meh... when the 4th element is "default graph", it is not a term...
tl: this is editorial work, we don't need to do that with 10 ppl
… let's move on to discuss "triple occurrence"!
pchampin: was "triple instance" and "triple occurrence" meant to be different?
niklasl: no, they are there because both where used
gkellog: Andy's definition is great and we should put it in there
pchampin: put it into RDF concepts?
gkellogg: for the moment, I would like to simply capture it because it is useful to refer to it
pchampin: please add some text to say that we may want to include these definitions in the spec
pfps: if there is no observable difference, we need only one term
pchampin: does anyone see a difference?
niklasl: I wanted to capture the assertion of a triple
tl: I do see a difference
… we need a name for the triple in the graph
… which is only once
--- because of set semantics
pchampin: for me the assertion is in the syntactic realm, the statement is in the semantic realm
<Zakim> gkellogg, you wanted to suggest that an RDF Statement which is in the set of triples associated with an RDF Graph is considered to be asserted.
gkellogg: that's why graph is an important concept
… it is a set of triples
… a triple / statement in itself is a concept without ground
… it becomes asserted when it is put into a graph
AndyS: yes, that's the problem
… asserting is an action
<TallTed> a triple is asserted (i.e., becomes an assertion) when ...
AndyS: a graph is immutable
gkellogg: Then, better if it is a constituant of a graph
<doerthe> at some point we talked about triple terms vs. asserted triples (longer ago)
pchampin: I suggest to replace "constituent" by "element"
gkellogg: yes
pchampin: This term (asserted triple) has a meaning only in the context of a graph
olaf: is an "asserted triple" a type or token?
niklasl: I will attempt to answer your version of this question in the mailinglist
… but I don't think the answer to this question can be added into the terminology
<pchampin> my take to olaf's question: "asserted" is a "contextual" property of a triple (relative to a graph); an asserted triple is therefore whatever a triple is
AndyS: just to say "RDF type" already exists - it is an informal term of RDF class
tl: I think "token" doesn't apply
… the distinction between different assertion events is crucial
niklasl: "asserted triple" as defined here is for _a_ graph
AndyS: but you missed the "w.r.t. a graph" part
tl: can we have "RDF Triple Occurrence / Instance" to indicate that we are not sure about the name
gkellogg: "occurrence" for me means to be an element of a graph
AndyS: spot on!
… "instance" has the same problem
niklasl: We have "asserted triple" and "RDF statement"
… niklas: "occurrence" is ambigous
pchampin: "occurrence" serves a purpose similar to what original reification was about
niklasl: I like that
tl: the definition of reification is pretty clear but the practical use is different
… So, if we use new terms now (claims and facts), it would be more correct but it would make it unusable for people coming to RDF
… I don't kow what to do
<Zakim> AndyS, you wanted to suggest "potential occurrence" or "possible occurrence")
tl: perhaps cover all the meanings by the annotation syntax
<pchampin> I like "possible use"
AndyS: reification names something in the domain of discourse
<Zakim> gkellogg, you wanted to note RDF Schema description of rdf:Statement
gkellogg: about the specific statement in RDF Schema
https://
<AndyS> "The reification only says that the triple token exists and what it is about, not that it is true, so it does not entail the triple."
TallTed: What we were just looking at in the Schema spec lends itself to be something for the new syntax we developed
tl: It is clear to me. Yet, the language of "tokens" in that spec text makes it clear to me that the triple is asserted
TallTed: No, it is explicitly there!
<niklasl> +1 to Ted, in that I am more and more certain that RDF-star triple term syntax should *at least* be for RDF reification (but I wonder if we *may* do more to make it "better" (in some way))
niklasl: what I wanted to say is what you also said regarding that ...
<pchampin> to tl, a token / materialization of a triple does not have to be "in a graph"
niklasl: I wanted to emphasize this supports part of the definition