The yields and spectra of strange hadrons have been measured as functions of centrality in 200 GeV AuAu collisions by STAR. Comparing with pp, dAu and CuCu, the system size dependence of these measurements is studied. Short-lived resonances, such as Sigma* and Lambda*, which may decay and regenerate in the medium, are used to examine the dynamical evolution between production and freeze-out for these systems. Particle production is investigated by comparing to thermal models, which assume a simple scaling between N(part) and the correlation volume, in order to calculate the strangeness enhancement. Our hyperon measurements in AuAu indicate that N(bin) may be a more appropriate scale for the strangeness correlation volume, in which case canonical suppression could not be simply parameterized with the geometrical overlap volume, but will depend on the individual quark content of each particle. This theory is tested by comparing data from different collision systems and centralities.