Bearing entity of digital (immaterial) content.

[trapperkeeper]

I know that the CCO has 3 branches for something like a graphic, intended to capture the representational entity, the directive entity, and the bearing entity (not the exact names). For something like a graphic, I believe these would be applied as follows:

representational: The information about how something is to be drawn to represent an object in the world, i.e. "a question mark shape represents an ear".

directive: The exact instructions for drawing the representational entity. For example, an SVG file that defines the stroke, stroke width, color, fills, coordinates, transforms, canvas size, etc.

bearing: The bearing entity is the perceptual elements that result from rendering the directives. The SVG rendered in the display. The colors you see, where you see them, ...

If we are correctly applying the above, then a confusion we have is in regards to bearing entities for digital artifacts (such as pixels or collections of pixels). Bearing entities are classified as material entities. That is, things with mass. Of course a pencil mark, drawn according to a directive, leaves behind graphite on paper. That fits fine within bearing entity. But what about digital artifacts, rendered as pixels? Photons do not have mass, but pixels are perceived based on those photons. I don't think a pixel would be said to include the emitters, the glass over the emitters, nor the lens of the eye. But then, where should these immaterial digital bearing entities be classified?

[alanruttenberg]

There's a lot to say about this subject, and there are many facets, as you point out. There's also ambiguity about the word pixel. That word has been used to mean the individual sensor in an array, a glowing bit or set of bits on a monitor, and the piece of information that is the intermediary - the bits.

Let's narrow this to a raster image like a digital photograph. That's defined partially structurally (raster vs vector). One thing I've been working on is splitting off the structure part.

As far as bearers go, the bearer story in CCO is messed up and on the way to being refactored, see #564

That something like a photograph is repeatable means it is an ICE of some sort. A modern digitial photograph has no standard bearer. Each time it is displayed (it is concretized) it can be on a different monitor or screen (or piece of paper). Those are bearers. In most work we aren't interested in that bearer. Monitors aren't unique to a photograph(the monitor can display different photographs). The bearer is only interesting when there's something interesting about it. For instance, if we were doing a proof on a specialty monitor and we want to have some provenance chain that included how the final was approved, we might want to mention that monitor. Or if we had a signed photograph, that's worth something, so we might want to record that instance.

If I had to guess I'd venture that you aren't interested in bearers at all, just content and format, neither of which will be material.

(this isn't an answer, more like an attempt to start the conversation)

[trapperkeeper]

Thank you Alan, I appreciate your contributions to this discussion. I deleted my last comment as it was not well formed. Let me try again.

Each graphic, in our context of interest, "represents" something in the world (i.e. a unilateral complete cleft lip). The object (cleft lipe) being the referent. Thus I think it is fair to call this graphic a Referential Content Entity. The instructions, for how to draw this graphic, might be recordable in a number of formats (i.e. as an SVG file). There can be many copies of the SVG, but they all contain instructions for drawing the graphic. I believe the SVG would fall under Directive Content Entity. Perhaps, as you say, the bearing entity for rendering this graphic, is the monitor. But what are the marks?

Is the bearing entity for a pencil drawing the paper? Or is it the paper plus the graphic deposited by the pencil? If the latter, then the marks (lines, dots, etc.) are part of the bearing entity. But what of digital marks?

On the topic of marks, where do they belong in the CCO before they carry information. That is, I might draw a triangle to represent a tree. But before there is a triangle, there is the left side line. It doesn't yet carry any information alone, but it needs to be represented (for us at least, as we are building graphics as collections of marks).

I admit to being fairly new to the CCO, so perhaps these questions are more obvious for some. I'm just looking for the intentions behind how the CCO should be applied.

[alanruttenberg]

So I was starting to try to reply in detail and it was getting way too TL;DR.

A short answer is that CCO and IAO are first starts on what is, I think, an incredibly multi-faceted and complicated domain. There are parts of CCO's information ontology that clearly address some use cases, other cases where it's not clear that it can or where its ambigous, or wrong (#564), and other cases where there's effectively been no development and people just make stuff up, or abuse what's there because it seems close-ish to what they are working on. We'll talk offline some time soon and I can elaborate more on what I think is the full scope of the ontological issues are.

But here, I want to try to bring it back to something more concrete. What are your use cases? What are the questions you want your representations (RDF/OWL) to be able to answer? What's the data you have that could support answering those questions?

Let's see if what you need is within what CCO users would tend to agree is reasonable to do with CCO as it is, or whether what you need requires a more complicated theory. For instance, you ask about where marks like the first line of the triangle are in the ontology. But do you have data and questions that require you to know that?

Also, not sure how much you've looked at OBI (Ontology for Biomedical Investigations) but it's worth being at least aware of what they have done. I worked on OBI and OBO's IAO, which predated CCO, was developed as an offshoot of OBI.

[trapperkeeper]

We have graphics (think representative drawings, not photographs or charts) that are built compositionally from graphical components (elements, groups, units, and graphics). For the sake of brevity, I'll skip for now the aspects of graphics that we would like to represent. Let us for now focus on only the marks. In our context, marks comprise units (e.g. sub-graphics) or graphics, units or groups comprise graphics. The distinction between those types of things isn't important for this, other than to know we have a hierarchical structure, for each graphic, of how it is composed. The leaves in the graph are the Elements, which are the graphical marks. Now, to your question, what is associated with these leaves?

First of all, they are components of larger elements in a graphic. So we need to represent that. Additionally, each mark is associated with one or more role. So we want to represent that as well. And we need to answer the question, what roles does this mark have, as that affects how it will be drawn.

To make that more clear, let me give an example. Consider this graphic: https://graphics.endlessforms.info/graphic?id=HUMAN0000813

If you hover over the elements in "components" box, it will highlight those parts of the graphic. Note that HUMAN0000442 is a dotted line at the edge of the philtrum. In the OWL that represents this graphic, it is associated with a style "ridge" (basically a css style that says that it will be a dashed line, with color black, with width 0.75, etc.). The style "ridges" is associated with a role in our Graphic Descriptor Ontology and or the extension Graphic Descriptor Ontology Anatomy extension (together GDO/A). The role in this case is "ridge line" (https://gdo.endlessforms.info/view/classes?uuid=GDO_0000072), which is a line depicting where a ridge occurs in the referent. So, in a figure with multiple ridge lines, they would all be styled uniformly. Similarly for regions. So, our other question that we wish to ask of marks is, what are their roles. In that sense, I guess, the mark already bears information.

It is in the context of the Graphic Descriptor Ontology that this question arises. Please note that the version of the ontology linked above is in beta and is receiving some structural changes at this time.