With composing or improvising we often reach the conclusion that less is more, just like blocking out some of the glare helps us see better all of the light that's there (photo).
When we're writing or improvising a coda, for example, the shorter it can be, usually the better it is.
Long, extended codas are out of place when we get the sense that the piece already ended several measures back.
If we're torn between using a shorter or longer coda therefore, the shorter version is usually preferable.
If we're writing a 4 voice fugue, for example, and the exposition is complete with the 4th entry of the subject, we might contemplate introducing a (5th) redundant entry (as in the D Major Fugue from Op. 18).
This runs the risk however of disturbing the architecture for the listener, and the benefit of providing the 4 voices an additional chance to enter into counterpoint with each other should greatly outweigh this risk or it would be best not to introduce a redundant entry at all.
Any needless complication which tends to make it more difficult for the listener to follow the moving lines in a fugue should be scrutinized carefully before it is introduced.
This includes the insertion of ornaments or cadenzas foreign to the composer's score under the guise of increasing its stylistic authenticity; this practice of changing the composer's music puts us on a slippery slope [See blog, Getting Started With Writing, Part XXXVIII].
Certain fugue themes, as well, when they have an intermediate note in the middle of a fast octave leap in the hands, often benefit by some form of shortening in their melodic curve to eliminate the awkwardness when the theme appears in the bass.
At times it's helpful to simplify the bass line in these situations by taking out that middle note if the change makes it easier to play it or it sounds smoother that way.
A well known example of this is in Bach's great a minor fugue, where the theme enters in the bass line in the 1st exposition in the dominant key; here the theme was rewritten by Bach to eliminate the F# in the middle of the quick octave leap from B1 to B2 in the feet.
This deliberate omission is hardly noticeable to the ear and is maintained by Bach, with fine effect, every time the theme enters in the pedal.
That note is missing from every manuscript copy for good reason; the composer wanted a smooth pedal line and nothing in the way to rob it of the clarity upon which the effect of every fugue depends.
Any argument therefore that the music should be changed to add that note to the subject every time the subject appears in the bass, in this case, is based upon a flawed premise.
Bach did not mean for that note to appear in the bass line and did not perform it that way.
If Bach the practical musician were alive today, he would take issue with any organist who tried to change his manuscript to include this note in the pedal line for the sake of consistency or to demonstrate how well they can manage these kinds of acrobatics with their feet.
That may be virtuosity, but it's not serving the music.
He would find having to practice it that way, even with heel-and-toe pedal technique, too awkward and destructive to the rhythm.