The situation of yield stress fluids in physics is very original: they borrow their properties partly to solids and partly to liquids, both material types which have strikingly different structures, and it is assumed that they can undergo a simple transition from one state to the other at a critical shear stress. This also implies difficulties for the mechanical description of their behavior: in the solid regime one usually follows the stress vs deformation while in the liquid regime one follows the stress vs shear rate. The experimental difficulties for determining the yield stress are inherent to this peculiar behaviour: one needs to identify the critical stress for which a flow occurs, which implies to detect a flow of a material with a viscosity tending towards infinity; as the flow curve tends to exhibit a plateau at low shear rates the uncertainty on measurements are obviously large and this is complicated by possible slight stress heterogeneities. Local measurements, i.e. inside the materials, are thus needed to determine the effective constitutive equation of the material. From MRI data we show that a certain class of materials appears to be simple yield stress fluids, except below a critical shear rate for which we cannot get any relevant data and which seems to be the “Bermuda Triangle” of pastes. For another class of materials the transition between the solid and the liquid regime is more abrupt: it occurs at a finite viscosity, so that no flow at a shear rate below a critical value can be observed in steady state. For these materials shear-banding in steady-state flows is the rule whatever the flow geometry. In addition this behaviour is associated with time-dependent properties: the shear-banding effect is obtained as a viscosity bifurcation in time. The behaviour of simple yield stress fluids (first class) in more complex geometries (extrusion, squeezing, object displacement, flow through a porous medium, etc) has received some attention in recent years and it is remarkable that the force vs velocity data are always well described by a model of the Herschel-Bulkley type, in analogy with their simple shear behavior. However the behaviour of yield stress fluids along a solid interface still constitutes a challenging field although of great interest for controlling various processes such as mud adhesion or drying, cream or gel coating, fouling deposits, welding of pottery pieces, adhesion of dental cements or glues, etc. We show some first observations with model materials which illustrate the complexity of this field and some unexplained results.