Document variability
This is work in progress.
Can a document change? This is a nasty bit of ontology.
A variable document has the same gestalt as a changing thing in the world, such as a university or a cat: It is possible to agree on which variable document is meant, even as it varies over time. What is needed is to provide necessary and sufficient conditions for something to be this particular variable document, as opposed to any other. One might
- specify rigid properties such as author, date of first publication, and title, or
- provide a summary of what the variable document persistently says (in the sense of "means"), or
- specify the ongoing process by which new editions of the variable document are produced.
When documenting a URI that names a variable document, you can't just say that it can change and expect your readers to know how. Documentation should explain what sort of changes to expect, and users of the document's URI need to take care to restrict statements to those that are either scoped to a particular time or that remain true even if the documented possible changes happen.
For rigorous determination of provenance, such as that required for repeating a document-based experimental analysis, it is necessary to be able to name unvarying documents, which might correspond to the state of some variable document at some particular time.
(Alternative ontological designs: (1) it is the URI-to-referent association that changes; (2) variable document and unvarying document state (contents) are disjoint types. Either of these would be awkward and would break the analogy with continuants.)
