Strict melodic keyboard-style voice-leading involves the composition of two primary musical lines—the melody and the bass line. The inner voices are secondary and serve largely as harmonic “filler.” All principles of good basso continuo voice-leading hold for melodic keyboard-style writing. However, because of the inclusion of a melody, several additional principles of composition must be observed.
The outer voices (melody and bass) draw the most attention, and therefore they should make good counterpoint with each other. The melody should largely follow the principles of composing a cantus firmus or a first-species counterpoint line. In a strict keyboard-style melody that means:
The melody should begin on a member of the tonic triad.
The melody should end on do.
The melody should have a single climax and good, smooth shape.
The melody should be “singable” (even though it will be played on the keyboard).
These melodic constraints may make following the law of the shortest way and contrary/oblique motion with the bass difficult, and at times impossible. When that happens, keep the voice motion as smooth (and playable) as possible, and be very careful not to compose voice-leading errors such as forbidden parallels.
In general, if you follow the figures, double the correct chord tone, move the upper voices as little as possible and in contrary or oblique motion to the bass, and take special care when the melody makes the latter impossible, your voice leading will sound smooth and will be fairly easy to perform. Those are the goals of strict keyboard-style voice-leading.
A number of specific situations come up frequently enough that they are worth tucking away as “stock patterns” to be pulled out when appropriate. See the Keyboard-style voice-leading schemata resource.