W3C home > Mailing lists > Public > public-openannotation@w3.org > October 2019

Re: inverse property of oa:hasTarget

From: Marilena Daquino <marilena.daquino2@unibo.it>
Date: Thu, 3 Oct 2019 11:23:24 +0000
To: "public-openannotation@w3.org" <public-openannotation@w3.org>
Message-ID: <AM6PR01MB4312C8D87794297BCDA324DAD29F0@AM6PR01MB4312.eurprd01.prod.exchangelabs.com>
Hi all,

> My gut reaction to the proposal of the inverse was that the same thing

> might be achievable by swapping the subject and object while keeping the

> oa:hasTarget predicate.

Apologies, the example was just to give you all an idea of how I will use it.

We serve data in JSON-LD, and having annotations nested in the element of interest is more intuitive for developers that will just consume the JSON. See for instance the following example, where “annotation” is our custom property/inverse property of oa:hasTarget.


  "@context": "[JSON-LD context URL]",

  "iri": "grp:04207",

  "a": "reference_pointer",

  "content": "Agarwal et al. 2010",

  "denoted_entry": "gbe:04201",

  "annotation": {

      "iri": "gan:04201",

      "a": "note",

      "note_content": {

        "iri": "gci:04201-04025/1”,

        "a": "citation_relationship",

        "citing_document": "gbr:04201",

        "cited_document": "gbr:04205",

        "citation_creation_date": { "value": "2015", "a": "year" },

        "citation_time_span": "P5Y",

        "characterisation": "f_extends"




I know that technical/implementation concerns should not influence the modelling activity, but still, it’s a thing.

>> On one hand, there's the problem of multiple representations for the same

>> thing (cf. http://richard.cyganiak.de/blog/2006/06/an-rdf-design-pattern-inverse-property-labels/).

>> But in my own work with OA, I have introduced a 'hasAnnotation' property

>> (not in the oa namespace) for the specific purpose of having a tracking

>> (some of) its annotations.  The annotations still use 'hasTarget'.

I understand the concern about interoperability and redundancy. My2cents, as you also imply, is that there might be dozens of inverse properties scattered around the world that do exactly the same job. I assume that having one property with the oa: namespace would make the ontology fit plenty of scenarios that you haven’t addressed yet (or just please adopters), and would in any case stem the proliferation of redundant properties. To this extent, I definitely agree with Robert that having inverse properties is a good practise that doesn’t hurt.

Many thanks,


Marilena Daquino
Research assistant /DH.arc, Digital Humanities Advanced Research Centre.
Department of Classic Philology and Italian Studies, University of Bologna,
40126, Bologna. Italy

Received on Thursday, 3 October 2019 11:23:48 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:38:31 UTC