Meta: Documents vs Document descriptions

For source analysis, it seems like it will be useful to say that there exists a document for every link in a communication path. These "documents" can be more ephemeral that typical documents. For example, the storage medium for the document might be someone's recollections of an event.

Viewed this way, "document" objects in our database in the general case may just be "descriptions" of documents. For example, a URL of a web page, a transcript of an audio or video program, or just notes someone took of a meeting they were in. The document description could even be as vague as "some book by Isaac Asimov". Of course, it is fair to say that these descriptions are also documents themselves, but they are directly derived from other documents, and often the document they are derived from has the most interesting information.

With this viewpoint in mind, it becomes useful to be able to identify document descriptions that are describing the same document and allow them to be linked together. It would also be useful to find the descriptions which contained large amounts of the info from the originating document (e.g. find a transcript of a radio show instead of just the name and time of the show).

The question also arises should we then create a document for every communication medium supplied in communication links. It would be useful to allow for source linking as described above, but we would also end up creating a lot of "documents" with very little actual data in them. Maybe we could store "light" document descriptions of this sort in a different table than actual documents, and allow them to be linked to "real" documents with substantial information content.

If so, we could also create a page to view recent "light" documents to allow for people to find and attach them to "real" documents.