2D

• perpendicular striping — this lets you see which direction rope is going, particularly for over/under crossings. Examples: [1]
  • the bonus is that this is relatively easy to automate in Inkscape
• diagonal striping — this appears more realistic, but when two segments cross each other, it makes it more difficult to see which crosses over and which crosses under. Diagonal stripes are less clear than perpendicular stripes. Examples: [1] [2] [3] [4] [5]
• simple line drawings — Rope segments have almost no width. This allows drawings to be very compact, which is probably one of the reasons that Ashley's Book of Knots chose it. However, IMHO, it can be difficult to interpret for neophytes. Examples: [1]
• "cell shading" — Rope is depicted with two lines that outline the outside of the segment only, with the interior of the rope being a flat color. Slightly more clearer than simple line drawings, but not much. Examples: [1] [2] [3] [4]
• pseudo-3D — Hand-drawn to depict some shadowing. This helps immensely to understand which segments cross over and which cross under. However, they take much more skill and more time to create. Examples: [1]. Diagrams of rope splices often end up being forced to be this, because they need to convey a lot of detail.

3D

• photographs — Pull out a nice piece of real rope, make your knot, and photograph it. This seems to be the approach that is taken on many Wikipedia articles. lots of examples
• 3D computer rendering — This has the advantage of using realistic shadows to depict crossing points. Also, the model can be rotated, possibly by the user, to give a better feel for the overall structure of the knot. Downside: it takes more work to create these.