In 1992 Bill Kent published a short essay titled "Attributes? Why?" that I want to re-read and maybe understand better. For many years I have held the view that there are two primary properties when talking about data elements, namely, kinship relations or attribute relations. This notion is from a 1967 paper by George Mealy "Another Look at Data" (Looking at Mealy's theory of data leads to further thinking about relations and attributes.)
Kent's first claim in the essay is:
Things in object models should be there for useful reasons.
And I have read the piece to the end and remain confused. I am thinking about this in a wrong way. But Kent does distinguish between stored data and object specifications. So this note is about computation, but I need more context for his ideas about the use or even need of attributes.
Updated 2023-01-12: added open access link to Mealy paper.