Representing Periodicals in OPDS 2

[JamesEnglish1028]

We have Magazine and Newspaper periodical providers using OPDS 2. There are no examples as to how this should be propoerly represented in OPDS 2. However, there seem to be most of the.

Because a periodical publication is published overtime as a Publication Issue - the Publication metadata collection @type property can use https://schema.org/PublicationIssue

The Publication Metadata collection's "belongTo" property can be used to create the associaltion of the PublicationIssue to the Periodical.
Example:

{
"belongsTo": { 
  "schema:periodical": {    
    "name":  "People", 
    "identifier": "uri:issn.org/0093-7673",
    "schema:issueNumber": "52" 
  }
}

In the above example the preferred Identifier of the periodical is the serial publication's ISSN. However, any unique persistent ID of the periodical publication could be used. It should not be used and the Publication Issues identifier property. That identifier property should be unique to the issue.
Additionally, becasue periodical are published as issues, it would be more preferable to include the issue number explicity as a property e,.g" "schema:issueNumber": "52"

the position property is to vague. for example a Periodical such as a Journal may have an article that holds the first "positioned" in an issue. therefore the potion property becasue confusing. It is best used when the publication holds a unique position with in the respective grouping conveyed such as in a series or collection if relevant.

additionally, becasue scheme.org does not provide a periodical scheme property to distinguish a Magazine from another type of periodical (such as a Newspaper or Journal). It may be useful to convey the Publication Issue is in a Magazine collection with the collection property. "collection":"magazines"

[JamesEnglish1028]

Looking at this closer in order to be able to support other type of periodicals and the way they are referenced in libraries, we should consider volume numbers as well.

A periodical can and typically does have volumes, where a volume usually represents a collection of issues published within a specific year, with each issue being numbered within that volume; meaning a periodical will be referenced using both a volume number and an issue number.

Key points about periodicals and volumes:

Definition:

A periodical is a publication that is released on a regular basis, like a journal or magazine. 

Volume meaning:

A volume refers to a grouping of issues published within a specific period, usually a year. 

Issue number:

Each individual publication within a volume is called an issue and has its own unique number

Therefore in the belongsTo collection we should ad or ask providers to add additional context regarding the volume to which a publication belongs using schema.org/PublicationVolume.
Maybe something like this.

{
“belongsTo”{
  ”schema:periodical” {    
    ”name”:  "People",
    ”identifier": "uri:issn.org/0093-7673",
  },
  "schema:publicationVolume" {
    ”schema:issueNumber”: 52
    “schema:volumeNumber”: 2020
  }
}

[JamesEnglish1028]

The above use case was for minimum information to Harvest a collection from a provider e.g. a server client. In the use case for a user client application, additional properties could be added to help with discovery and presentation using a link property in the respective belongTo collection objects.

{
“belongsTo”{
”schema:periodical” {
”name”: "People",
”identifier": "uri:issn.org/0093-7673",
{
"links" : {"rel": "self", "href": "https://example.com/0093-6673", "type": "application/opds+json"}
},
"schema:publicationVolume" {
”schema:issueNumber”: 52
“schema:volumeNumber”: 2020
}
"links": {"rel": "self", "href": "https://example.com/0093-6673/2020", "type": "application/opds+json"}
}

[JamesEnglish1028]

So as a best practice, we use the appropriate scheme.org publication type

There are specific types for Newspapers e.g scheme.org/Newspaper could be added for context

{
“belongsTo”{
”schema:periodical” {
"@tyoe": "schema.org/Newspaper"
”name”: "New York Times",
”identifier": "uri:issn.org/1553-8095",
{
"links" : {"rel": "self", "href": "https://example.com/1553-8095", "type": "application/opds+json"}
},
"schema:publicationVolume" {
”schema:issueNumber”: 365
“schema:volumeNumber”: 2020
}
"links": {"rel": "self", "href": "https://example.com/1553-8095/2020", "type": "application/opds+json"}
}

[JamesEnglish1028]

Journal Use Case

Journals can be more complicated becasue they contain independently published articles that are published as stand alone publications as well as part of a published periodical. However that should only impact the publication's metadata @type property to identify it not as a publication issue, but rather a journal article.

For an independently distributed Article would need to add to the belongsTo property the context of the an issue
Example

{
“belongsTo”{
”schema:periodical” {
”name”: "Nature",
”identifier": "uri:issn.org/1476-4687",
{
"links" : {"rel": "self", "href": "https://example.com/1476-4687", "type": "application/opds+json"}
},
"schema:publicationVolume" {
”schema:issueNumber”: 52
“schema:volumeNumber”: 2020
},
"links": 
{"rel": "self", "href": "https://example.com/1553-8095/2020", "type": "application/opds+json"}
},

If the publication is the complete issue and not the individual article, then
it would have a publications metadata collection "@type": "schema.org/PublicationIssue"

One note, a Jounral Articles and Publication Issue shares the ISSN of the periodical in which it was publlshed, therefore, the publication metadata collection identifier property should NOT use ISSN becasue it will not be unique beteen issues or other article issues. Another URI should be used at the Issue and Article level.

[JamesEnglish1028]

While i was not intending to go into article or story level expression via OPDS for metadata harvest and client presentation, it is apparently a more common use case need.

Wading into these waters I see two possible approaches.

Because the spec says:

"OPDS 2.0 is based on the same abstract model as the Readium Web Publication Manifest."

AND
_"In the Readium Web Publication Manifest model, everything has to be a collection:

a collection is identified by its role
a collection contains both metadata and links, along with sub-collections
specific collection roles that only contain links are called compact collections and use a compact syntax (array of Link Objects)"_

I interpret that the "compact collection" should be used.
The Links object in the Periodical Issue will contain this compact collection

The question is what does it link to?

RWPM has two things that could contain convey the contents - The resources collection, or the readingOrder collection

Example readingOrder

"readingOrder": [
    { 
      "href": "https://example.com/1553-8095/2020)/1l", 
      "type": "text/html", 
      "title": "Article 1 - Title"}, 
    {
     "href": "https://example.com/1553-8095/2020)/2", 
     "type": "text/html", 
     "title": "Article 2 - Title"}, 
    { 
      "href": "https://example.com/1553-8095/2020)/3",
      "type": "text/html", 
      "title": "Article 3 - Title"}, 
    {
    "href": "https://example.com/1553-8095/2020/4)", 
    "type": "text/html",
    "title": "Article 4 - Title"}, 
    {
     "href": "https://example.com/1553-8095/2020/5", 
      "type": "text/html",
      "title": "Article 5 - Title"}, 
    {
      "href": "https://example.com/1553-8095/2020/6", 
      "type": "text/html", 
      "title": "Article 6 - Title"}
  ],

the other example is within the resource collection

{"href": "html/toc.html", "rel": "contents", "type": "text/html", "title": "Table of Contents"}

the above would basically be a web document with links to the respective articles or stories.

Another question would be how to provide context to the two collections such as
Journal or Magazine Article >. https://schema.org/Article
Newspaper "story" / Article > https://schema.org/NewsArticle

Which methods are better for the two use cases - harvesting vs client presentation?