Abstract

The concept of libraries of stellar spectra is by no means new, though access to on-line ones is a relatively recent achievement. The road to the present state has been rocky, and we are still far short of what is needed and what can easily be attained. Spectra as by-products of individual research projects are inhomogeneous, biassed, and can be dangerously inadequate for modelling complex stellar systems. Archival products are eclectic, but unique in the time domain. Getting telescope time for the required level of homogeneity, inclusivity and completeness for new libraries requires strong scientific arguments that must be competitive. Using synthetic spectra builds misconceptions into the modelling. Attempts to set up the initial requirements (archives of observed spectra) encountered dogged resistance, much of which has never been resolved. Those struggles, and the indelible effects they have upon our science, will be reviewed, and the basics of a promotional programme outlined.