Re: strange terms defined by reference in RDF Semantics

I’ll answer several different questions from different emails, some of which I believe I’ve touched on before:

The “index” section is added automatically by ReSpec when there is a section with id=“index”. In this case, `<section id=“index”></section>`; it’s defined in [1]. It seems that when there is a citation of a spec but no terms from this spec are referenced, it instead adds an entry that allows you to find where those citations occur. For example, RDF Semantics references [RDF12-N-QUADS], and the bullet item underneath it in "Terms defined by reference”, shows the permalink URL of the spec, and the two references in “Set of Documents” and “Informative references”. Just because this shows up in the index, does not make this a term. The first citation looks like the following:

<cite>
  <a data-cite="RDF12-N-QUADS”
    data-matched-text="[[[RDF12-N-QUADS]]]”
    href="https://w3c.github.io/rdf-n-quads/spec/“
    id="ref-for-index-term-rdf-1-2-n-quads-1”>
    RDF 1.2 N-Quads
  </a>
</cite>

Not a term Definition.

In contrast, it references the term “denotes” from RDF12-CONCEPTS, which looks like the following:

<a href="https://w3c.github.io/rdf-concepts/spec/#dfn-denote" id="ref-for-index-term-denotes-1">denotes</a>

A term that it does define is “class”, which looks like the following:

<dfn id="dfn-class" tabindex="0" aria-haspopup="dialog" data-dfn-type="dfn">class</dfn>

Note the <dfn> element and data-dfn-type=“dfn” attribute, which does make this a definition.

A hypothetical change to ReSpec might eliminate the citations used without creating index entries. Feel free to raise an issue at github.com/w3c/respec <http://github.com/w3c/respec> about this.

Regarding data-cite [2] vs. xref [3]: once we’ve published, terms we define can be accessed without explicit data-cite. For example, RDF12-SEMANTICS defines “class” which can be referenced using `<a data-cite=“RDF12-SEMANTICS#dfn-class”>class</a>`. Once we these specs are published, they can be accessed just using `<a>class</a>` (which ReSpec also says is not recommended, but I prefer this style), or `[=class]` which they say is preferred. This will attempt to resolve” class by exported term definitions. The set of specs across which it will search can be refined using the “xref” property [3] in respecConfig, which is a way to deal with conflicts when similar terms are defined in different specifications. We could consider making all of our references simpler after FPWD, but in my experience this can create transient errors, and the data-cite mechanism is less prone to these kinds of errors.

With regards to “hallucinating citations”, as mentioned above, you can use the index to find where each citation is used. In the case of RDF12-XML, it is used in the “Set of Documents” cross-index, and because of that in the “Informative references” section. The “Set of Documents” references all documents produced by the working group, which is common when groups publish a set of documents.

Gregg Kellogg
gregg@greggkellogg.net

[1] https://github.com/w3c/respec/wiki/index
[2] https://github.com/w3c/respec/wiki/data-cite
[3] https://github.com/w3c/respec/wiki/xref

> On Apr 7, 2023, at 5:06 AM, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote:
> 
> On 4/4/23 17:17, Gregg Kellogg wrote:
>>> On Apr 4, 2023, at 2:13 PM, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote:
>>> 
>>> 
>>> 
>>> On 4/4/23 17:09, Gregg Kellogg wrote:
>>>>> On Apr 4, 2023, at 1:40 PM, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote:
>>>>> 
>>>>> RDF Semantics has citations of RDF Schema and SPARQL Entailment Regimes.
>>>>> 
>>>>> In H.2 Terms defined by reference there ends up
>>>>> 
>>>>> [ RDF12-SCHEMA ] defines the following:
>>>>> 
>>>>>    RDF 1.2 Schema
>>>>> 
>>>>> [ SPARQL12-ENTAILMENT ] defines the following:
>>>>> 
>>>>>    SPARQL 1.2 Entailment Regimes
>>>>> 
>>>>> This looks like a bug.  How can I get rid of the terms?
>>>> If you click on the link in the “Terms defined by reference” section, it should take give you a link to each place that term is used. For RDF 1.2 Schema, it is used in the Abstract, Set of Documents, and Informative references sections. As every documnent in our set if iin the Set of Documents section, they will all show as informative references.
>>>> Gregg
>>> 
>>> 
>>> But how do I keep the citations and get rid of these extraneous terms?
>> The terms are directly related to the citations, as I mentioned elsewhere. I don’t believe there’s any way to cite documents without ReSpec creating something like a term.
>> Note that the two sections in the Index: “Terms defined by this specification” and “Terms defined by reference” could be considered more of an editing aid, and we may decide to not keep them long term; but, they are useful to see what terms are being defined or used. Personally, I find this useful and don’t think it’s a problem.
>> Gregg
> 
> If there is no way to use ReSpec links without creating these extraneous terms how can I remove them from the documents I am editing?   I don't want to having something published with my name on it that I find embarrassing.
> 
> peter