popay, the formula at #8 assumes a homogeneous fog, which does not consider stratification (fog layers during formation or shallow fog) or cumulo forms (fog ‘clouds’ when dispersing), in these conditions the horizontal visibility can vary with altitude (cf #5) and hence change the slant visual range (SVR). These are the more complicated (hazardous) fog situations where a decision can be made based on the required visual scene and then it reduces; they are more often encountered in Cat 2 conditions. Thus SVR is not a consistent parameter for determining the visual segment.
The complications caused by this were simplified by using RVR (measured at 15m altitude) and a predetermined visual segment (what the pilot requires to assess position / fly the aircraft) to provide a probability of approach and landing success; IIRC something like 98% prob of not loosing the visual requirement. This was based on the research at BLEU in the 1970s; however some of the more recent changes (JAA etc) reducing minima might not be as reliable. One such operation is ‘super fail passive’ Cat 3 (150m dual autoland) where the manual landing option when below 200m is removed to reduce the overall risk. Other changes may have omitted the effect of pilot's eye height.
Thus landing minima (RVR) is a function of the visual requirements to check position / land the aircraft, which can be a function of aircraft size and manoeuvrability (for manual landing).
Similarly take off minima are based on the visual requirements, but the task is different, to either conduct / continue the take off, or reject the take off. These requirements are also a function of aircraft type (ease of handling / equipment fit), which gives greater variability in the minima, and thus even more difference with the landing minima.
Visual assessment: when stationary at the threshold and knowing the runway / centreline lighting spacing, the RVR can be approximated by counting the number of lights seen – actually quite accurate.