Visual minima timings of eclipsing binaries: To use or not to use?

Plenty of mid-eclipse timings of short-periodic eclipsing binaries derived from series of visual observations appear to be an acceptable source of information for orbital period analyses, namely if they were done in time-intervals not covered by other types of observations. However, our thorough period analysis of the nearly contact eclipsing binary BS Vulpeculae proves that visually determined timings done in 1979--2003 were undoubtedly biased to accommodate the existing linear ephemeris. The heavily subjective character of visual observations disqualifies them as a source of true phase information apt for fine eclipsing binary period analyses. Consequently we warn against the use of visual timings without a preceding careful verification.