If you want to average the light falling on a plane running through the subject that is parallel to the film plane, then point the dome at the camera.
But why and how often would you actually want to do this as a matter of course?
Yes; instruction manuals say to point the meter at the camera. What a handicap and a disservice! Sometimes it takes a thinking and constantly working independent user in the field to figure out how to use something to its best benefit, not a designer. Collins was not the first to figure out the best way to use an incident meter, and he is certainly not the only one who uses them that way. Everyone who has ever taught me anything worth knowing about metering has said the same. It is certainly how incident meter use is taught at the Art Center College of Design, perhaps the most well-renowned commercial photography school in the nation, for one.
I'd just suggest that anyone who is curious shoots the six test shots I mentioned. Then make up your mind about what to do based on those shots. If you always shoot negative film, metering down the middle won't kill you, though it is not ideal. Most people don't notice, and if they do, most people actually think that slightly overexposed negs look good, so think nothing of it. As I said, it ain't a big deal with negs...though I still like to have my ideal exposure when possible. But shoot this way on positive materials and you are hurting yourself.
If you want to average main and fill, meter down the middle, of course! But think about why you want to do that first. What averaging does is to compromise between two ideal exposures for two respective parts of the image. It gets a printable neg in most situations. But if you want to preserve the lighting ratio that exists at the scene without overexposing, especially important when shooting positives, you meter the brightest light that is illuminating the part of the subject for which you want to expose.
If you don't like what metering the brightest side does to the subject, then you probably aren't properly visualizing what the ratio of light you are shooting will look like on film. Doing so will give you an accurate exposure for the most-lit part of the subject, while causing the less lit parts to fall where they ought to fall. If you are shooting in contrasty light, you should know how your film will react to that, and expect a contrasty picture. The exposure is not wrong if it has too much contrast when shot in contrasty light. It is absolutely correct, within the limits of the film when exposed and processed normally. If you don't like it, i.e. if you want to alter the ratio that is there, and capture something different on the film, that is when you use exposure and development changes, or perhaps take an averaged reading to give you some wiggle room in printing. Just because you are using an incident meter does not mean you are "stuck" with what you've got. That is not what I am saying.