We discuss here some crucial issues related to the validity of the variable stripe length (VSL) method to measure optical gain in semiconductor materials and we especially point out the main experimental as well as conceptual difficulties arising when the VSL method is barely applied to low gain materials such as silicon nanocrystals (Si-nc) devised in a planar waveguide geometry. The variable stripe length method is revised carefully considering all the basic assumptions of its underlying one dimensional optical amplifier model, such as the gain saturation problem, the pump diffraction effects, and the inhomogeneous collection coupling of the amplified spontaneous emission. We will show how the standard one dimensional optical amplifier model has to be generalized in order to avoid undue artefacts as well as fundamental flaws. We indicate that, once the model has been properly modified, a more refined experimental analysis can be performed on the VSL experimental data yielding unambiguous gain values even in the case of planar waveguides made of low gain materials.