Modularity is also worth considering. At the level of the individual, modularity is adjacent to repairability and to functional specificity of subunits. But when you're looking at a natural history, modularity also becomes apparent at the population level.
If you have a collection of sewing machines (or remains thereof), sorted by age, you'll see what might look like a lineage by ancestry with variation, that might look consistent with an evolutionary model based on reproduction with variation plus selection. But then you notice specific features such as electric motors, plastic parts, and microcontroller chips, that don't have ancestral antecedents in the sewing machine lineage, but that had appeared previously in other non-sewing-machine items. Or actually, let's be more specific: within the sewing machine lineage they had functional antecedents that were not structural antecedents. Certain metal parts disappear to be replaced with plastic parts of identical function and near-identical shape. Elaborate cam-based automata disappear, to be entirely replaced with microcontrollers that are morphologically completely different but perform the same function (variant stitch patterns).
Such observations don't prove intelligent design, but they do prove that some process outside of reproduction-with-random-genetic-variation and selection has been going on. More importantly, when they're not observed, it's strong evidence against a generally intelligent designer. It's valid to doubt that an intelligent designer would have any reason to refrain from transferring "modules" (such as visual organs, tooth forms, venom defenses, or manipulator appendages) arbitrarily across lineages.
If you have a collection of sewing machines (or remains thereof), sorted by age, you'll see what might look like a lineage by ancestry with variation, that might look consistent with an evolutionary model based on reproduction with variation plus selection. But then you notice specific features such as electric motors, plastic parts, and microcontroller chips, that don't have ancestral antecedents in the sewing machine lineage, but that had appeared previously in other non-sewing-machine items. Or actually, let's be more specific: within the sewing machine lineage they had functional antecedents that were not structural antecedents. Certain metal parts disappear to be replaced with plastic parts of identical function and near-identical shape. Elaborate cam-based automata disappear, to be entirely replaced with microcontrollers that are morphologically completely different but perform the same function (variant stitch patterns).
Such observations don't prove intelligent design, but they do prove that some process outside of reproduction-with-random-genetic-variation and selection has been going on. More importantly, when they're not observed, it's strong evidence against a generally intelligent designer. It's valid to doubt that an intelligent designer would have any reason to refrain from transferring "modules" (such as visual organs, tooth forms, venom defenses, or manipulator appendages) arbitrarily across lineages.