Definition of image / represents #593

Current definition:

An Information Bearing Artifact that is designed to bear some specific Information Content Entity that represents some entity.

I'll skip of the IBE aspect of this, and assume the IBE/ICE refactor so that we're talking about ICEs.

The definition uses (in the text) represents, and I'm assuming that means the "represents" relation. That relation is defined as:

"Isomorphism between the carrier of x and the represented entity can be via a direct similarity relation, e.g., grooves in a vinyl record corresponding to sound waves, or linguistic convention, e.g., a court stenographer's transcription of spoken words, as well as others, such as encoding processes for images."

If we are to take this definition then the definition of image is too narrow. Why? Hand drawings will be expected to be images but there is typically no isomorphism. Even a photograph is not isomorphic to what it is a photograph of. This makes me think that, the isomorphism language should be dropped.

On the other hand it is too wide in that an audio recording would seem to satisfy the definition.

The chart definition:

An Image that is designed to represent an Information Content Entity by means of Written Symbols in order to convey that information in a readily understandable format.

Saying it uses symbols it seems to inconsistent with the "direct similarity" that the represents relation describes.

(as an aside, I'm not actually sure what CCO's chart is supposed to mean)
(another aside: Why is Written Symbols, capitalized?)

A certain class of image is central to my work. I'm having some trouble coming up with a good definition. But getting CCO's house in order would help. The class I work with are digital photograph-like things - that derive in whole or part from an array measurements of electromagnetic radiation incident on a sensor. "In part" because the digitized thing might have blank sections (as when you rotate an image within a rectangular frame, or there are bad pixels), or might have markup. Or it might be the measurements are simulations, as in a computer generated scene. I say derive because it can also includes things like false-colored images synthesized from those measurements.