Good science relies on tested data- and this tested data stays the same. The rules derived from it, may become more detailed - or in some cases even completely

Good science relies on tested data- and this tested data stays the same. The rules derived from it, may become more detailed - or in some cases even completely overthrown with a new ruleset, which will provide a reference frame that can produce such data. Best example for science is the Newton Physics, which allowed for pretty close calculations on planetary movements. The data gathered stayed the same, but the unexplained edge-cases and small deviations, where better explained by Einsteins theory.