To be making paintings through a computer raises interesting questions about control, expertise and research. In computer graphics "artifacts" are unwelcome blisters on images. In painting such accidents can be very welcome at the end of a frustrating day. You could say that all a paint program does is produce unnatural by-products. And painters? Yes, we do the same. Some argue that painting expertise is built into a good paint program, and the difference between the amateur and the professional is now eradicated. Others would say the opposite, that mastering both digital and painting know-how is quite a conundrum in itself, precisely the kind of "knowledge" that a research project should tackle. My initial thought was that such research in digital painting just has to be practical. Like software, which can be built on years of testing and refinement, a painting either works or it doesn't. This can be a matter of absolute precision. Eureka moments in painting seem to come about empirically. They have been written up, so to speak, retrospectively. Clearly, we should be wary of confusing the write-up, especially if written in advance, with the substance. Some pseudo-science surrounding new media art has been somewhat immodest, and something of a smokescreen. It never explains how the art is to be put to the test. Digital painting, like each niche research area in art, does need a campaigning voice, but it also needs a sense of critical measure. Unfortunately, the artifacts normally do not do this for themselves. They cannot speak to a critic on the phone, release press releases or write their own grant applications. Language? Yes, I am afraid we do need it. The smart/dumb case can be made either way with a few examples: a renaissance painting can demonstrate a "research active" artist, while another painting equally magnificent as art could indicate an airhead. A painter can, so to speak, "embed" an argument in the visual texture of a work. Equally, some painters, even digital ones, will say that the role of material, of the "artifact", is to resist and obstruct "interpretation". A "stupid" painting with no subject can of course release as much intellectual energy as a research-drenched piece of visual pedantry. One project may involve condensing a huge amount of data, another no data at all. With digital paraphernalia a painter can "research" topics that were previously much less accessible, and among examples I have tackled I shall show one about newsgathering, and one about a frog. Research outcomes? Inconclusive.