I like the term computationalist because, as you suggest, computer analysis can be performed on more than just text. The terms ‘descriptionist’ and ‘fulltext analytics’ went through several iterations in drafts of this post, and I was never quite comfortable with them. Perhaps my remaining doubts are that the terms relate to the actors doing the description. The term ‘descriptionist’ (in my mind, at least) implies the human element. A ‘computationalist’ is a human actor as well, but one that creates the algorithms. It is the execution of the algorithms themselves that generates the description — akin to the activity that the descriptionist performs.

That the computationalist and the descriptionist bring different behaviors to the act of creating a surrogate for the target item is why I think a combination of the two is important.

I think the computationalists can do a decent job of relating their assumptions to a broad range of theories in computer and cognitive sciences, statistics, and linguistics.

I should like to see the descriptionists do likewise in attaching their assertions to fields of inquiry beyond that of the library (but amply represented within the library).

Descriptionists might try avoiding the easy reach for its philosophical roots and pay a great deal more attention to psychology, linguistics, anthropology, and sociology.

Each of these fields have much to say about the kinds of behavior indulged in when one describes a Resource for oneself or on behalf of others.