Slide exposures are necessarily keyed to the highlights; that is to say, the exposure is determined by the brightest area in the shot. If this is overexposed it will 'blow' to a featureless white.
Incident light metering used to be known as the 'artificial highlight' method, which gives the game away: the incident light dome/Invercone is indeed an artificial highlight.
For landscapes, you can read the light falling on the subject (or its equivalent -- the light a hundred yards away is likely to be much the same as at the metering position, allowing for moving clouds) and this is far and away the easiest and best approach. Just about ANY meter will do: with incident light readings (again allowing for moving clouds) there aren't going to be two readings to average.
All adjustments are intuitive: you give a little more exposure for a lighter tranny, a little less for a darker one.
You can also do it the hard way and take spot readings of highlights, and there are times when this is better (for example if you are in the shade and cannot take an equivalent light reading, or in movies where you want to key your exposure to the star's face), but in the vast majority of landscape photography there is no need for this.
Negative exposures are normally keyed to the shadows which is why spot readings of the shadows are best, but you were not asking about that. Nor is there much to be gained (when exposing transparency) by reading both highlights and shadows, except to find out how much shadow detail you are going to lose: your exposure is constrained solely by your assessment of how light you dare let the shadows go.
You might also care to look at the free 'Bracketing' module in the Photo School at
www.rogerandfrances.com.
Cheers,
Roger