More than 1000 quantitative gas-solid and solid-solid reactions in more than 25 reaction types all across chemistry have been exploratory developed in Oldenburg since 1985 and the waste-free preparative use also at the kg scale became possible due to the experimental three-step "phase-rebuilding mechanism" as deduced from our nanotechnological investigations (AFM, SNOM, GID, Nanoscratching). None of these reactions were foreseen by highly acclaimed Schmidt's "topochemistry" hypothesis that strangely claims "minimal atomic and molecular movements" for isomerizations and photoreactions despite very poor predictive power (e.g. too short distances might be impeding, etc). Conversely, long-range anisotropic molecular migrations have been secured by all of the various nanotechniques used. This applies yet to all reaction types among the different solid-state reaction techniques. Only truly topotactic reactions without geometric change do not require molecular migrations as secured by AFM down to the molecular level. The reason for the molecular migrations is the necessity for release of the enormous pressure that is created by the geometric change of the molecules upon chemical reaction within the crystal bulk. Such release requires the presence of cleavage planes or channels or voids in the crystal: no reactivity is found in their absence. Therefore (unlike very few exceptional topotactic reactions without geometric change) reactivity predictions cannot be based on distances of reacting centers but they must analyze the crystal packing on the basis of the crystal structures. This provides the answers to the anisotropy of the reactions with single crystals. The shapes of cleavage planes and channels (11 basic types) will be classified, limiting cases pointed out. The influence of gaseous or solid reagents is comprehended by their molecular size. All three steps in the solid-state mechanism (1. phase rebuilding, 2. phase transformation, 3. crystal disintegration) must occur in proper solid-state reactions. The strict application of this mechanistic knowledge indicates how the gas-solid and stoichiometric solid-solid reactions have to be performed to give 100% yield in a short time due to favorable kinetics, or how they should be engineered if the steps 2 and/or 3 provide difficulties. Local melting can be detected by AFM. It stops reaction in most cases, as melt reactions require much higher temperatures. But cooling down below eutectics will enable then.