Patterns recurring at multiple scales.
Zooming in / out does not change the essential structure of the system / the function looks the same (possibly after rescaling appropriately).
the whole resembles some of its parts at any scale
A strike is like a mini revolution - all the elemnts of a revolution are present in a strike. People who you thought would never move become transformed in front of your very eyes, women in particular.
Link to original
And strikes happen for the same reasons that revolutions happen. If a system is under stress, eruptions are pre-programmed. The size and trigger for specific events are unpredictable, but the causes and overall dynamics are not. One may try to delay / reduce the impact in controlled releases through various safety valves, but sooner or later, if the internal contradictions are not resolved and the systems means to alleviate the stress are exhausted, revolutions will happen (see this for more / the general). ^335c11
Terminology: self-similar vs scale-free vs scale invarance vs homogeneity vs fractal
Homogenous functions: (exact rule that applies to a function)
Scale invariant model: After coarse-graining and rescaling, the model looks statistically the same (same distribution of properties at different scales).
Self-similar object/process: Looks the same at different scales (exact or statistical).
Scal-free distribution: No characteristic scale, tail follows a power law.
Fractal: Object with detailed structure (may change with scale in a non-trivial way) at arbitrary small scales, often appearing self-similar at various scales.
Self-similarity doesn’t force heavy tails.
Heavy tails don’t imply a self-similar process.
Scale-invariant systems typically generate: homogeneous observables + self-similar geometry + scale-free distributions.
Fractal Self-similar + non-integer dimension + roughness across scales