The Casimir-Lifshitz force acts between neutral material bodies and is due to the fluctuations (around zero) of the electrical polarizations of the bodies. This force is a macroscopic manifestation of the van der Waals forces between atoms and molecules. In addition to being of fundamental interest, the Casimir-Lifshitz force plays an important role in surface physics, nanotechnology and biophysics. There are two different approaches in the theory of this force. One is centered on the fluctuations inside the bodies, as the source of the fluctuational electromagnetic fields and forces. The second approach is based on finding the eigenmodes of the field, while the material bodies are assumed to be passive and non-fluctuating. In spite of the fact that both approaches have a long history, there are still some misconceptions in the literature. In particular, there are claims that (hypothetical) materials with a strictly real dielectric function $\varepsilon(\omega)$ can give rise to fluctuational Casimir-Lifshitz forces. We review and compare the two approaches, using the simple example of the force in the absence of retardation. We point out that also in the second (the "field-oriented") approach one cannot avoid introducing an infinitesimal imaginary part into the dielectric function, i.e. introducing some dissipation. Furthermore, we emphasize that the requirement of analyticity of $ \varepsilon(\omega)$ in the upper half of the complex $\omega$ plane is not the only one for a viable dielectric function. There are other requirements as well. In particular, models that use a strictly real $\varepsilon(\omega)$ (for all real positive $\omega)$ are inadmissible and lead to various contradictions and inconsistencies. Specifically, we present a critical discussion of the "dissipation-less plasma model". Our emphasis is not on the most recent developments in the field but on some conceptual, not fully resolved issues.