Defining a list of roles for Divina

[HadrienGardeur]

The initial list of roles for this project is directly based on DPUB ARIA 1.0, which doesn't really cover anything comic-specific.

Some roles remain useful (cover or pagebreak for example) but there's nothing really meant to convey the structure within a page.

Looking at the EPUB 3 Structual Semantics Vocabulary 1.1, there's a short list of roles dedicated to comics inherited from the work on EPUB Region-Based Navigation 1.0:

balloon

An area in a comic panel that contains the words, spoken or thought, of a character.

panel

An individual frame, or drawing.

panel-group

A group of panels (e.g., a strip).

sound-area

An area of text in a comic panel that represents a sound.

text-area

An area of text in a comic panel. Used to represent titles, narrative text, character dialogue (inside a balloon or not) and similar.

Here's my take on this:

• the concept of page and spread are missing and even though spread can be inferred from the presence of spread we could also have comics where a single image is used for a spread
• panel and panel-group are fine as structural elements but we might need something even more generic than panel, since a number of comics do not use panels at all
• balloon, sound-area and text-area feel confusing. text-area is probably a superset of bubble (I prefer that term to balloon) but we need to further qualify different types of bubbles and potential other text areas

I expect this to be a long and iterative process, in which we should probably try to identify the higer-level structural elements first before we dive any deeper.

[HadrienGardeur]

Here's a first list of suggested roles.

Structure

This covers the basic structural building block of a Divina publication.

Role	EPUB type equivalent
spread
page
panelGroup	panel-group
panel	panel

I think it's worth pointing out that img in ARIA is quite similar spread, since it's meant to be used over multiple images that should be presented together:

<div role="img" aria-label="Description of the overall image">
  <img src="graphic1.png" alt="" />
  <img src="graphic2.png" />
</div>

Text

Role	EPUB type equivalent
bubble	balloon
caption
sound	sound-area
shoutBubble
speechBubble
thoughtBubble

bubble is meant to be used for bubbles/balloons who do not have a specific role (speechBubble, thoughtBubble or shoutBubble) to identify them.

caption is meant to either identify a caption below a panel or within a panel (usually a narrator). It would be inherited from <caption> and <figcaption> in HTML.

[llemeurfr]

Apart from that I'm ok with this vocabulary (as a start).

[HadrienGardeur]

I'm not a fan of "caption" for the information given by a narrator

Based on what I've been able to identify so far, "caption" seems to be the correct term in English. This page does a great job of providing a full vocabulary for comics: https://blambot.com/pages/comic-book-grammar-tradition

[llemeurfr]

ok then; I had not find a proper English lexicon.

[crigaud]

Regarding text, shoutBubble, speechBubble and thoughtBubble being at the same role level than bubble is strange for me. Would it be possible to qualify bubble with shout, speech and thought instead?

Also, I feel comic characters are missing, along with their link to bubble and text.

[HadrienGardeur]

Regarding text, shoutBubble, speechBubble and thoughtBubble being at the same role level than bubble is strange for me. Would it be possible to qualify bubble with shout, speech and thought instead?

Since roles are represented as an array of strings, we can't represent any kind of structure.

In this proposal bubble is the generic (and fallback) role for bubbles, while shoutBubble, speechBubble and thoughtBubble are meant to be specific (we could extend the list if you think that's necessary).

Also, I feel comic characters are missing, along with their link to bubble and text.

This would be handled through a dedicated character property which I have yet to submit a proposal for. At a bare minimum, this should also allow an array of strings (the names of these characters), but I'd like to explore an object representation as well which could contain: name, identifier, gender and age.

Take a look at the following comment if you're looking for an example: #8 (comment)

[HadrienGardeur]

I've created a complete example at: https://github.com/readium/guided-navigation/blob/main/examples/Pepper%20%26%20Carrot/guided.json

In terms of semantics, this was a very straightforward example where only panel, speechBubble and shoutBubble were needed.