Manovich has done some interesting work. The projects he and his team have produced through the Software Studies lab at UCSD are inspirational. I'm primarily engaged with the how of what he does, rather than the objects of his study. The emphasis on pattern analysis is important, what is sometimes less interesting is the material chosen for analysis, or at least, that is the case from the art historical point of view. Very little of what is analyzed or looked at is art, instead it falls within the realms of visual culture or social media. Applying methods of computation, including algorithms, to works of art raises some specific issues and challenges that non-art material does not. There are questions and methods that need to be thought through; later blog entries will be about what I think those might be.
The list of sub-groups and examples of computational art history is useful, and captures many of the kinds of things being done in art history under the larger umbrella term of digital humanities. These mostly are about the movement, reception, and interaction with objects in cultural systems (patronage, exchange, collection, dissemination); not so much about objects themselves.
Manovich does cite Douglas Engelbart's notion of augmenting human intellect, which is fundamental: always we hope we're doing work to improve or add to knowledge...not merely to re-arrange what we already know. Is there potential for boot-strapping, too?