Afficher les documents scientifiques sur le Web

Every day CERN handles a large number of research documents, mostly marked up in \LaTeX{} and coming from many Internet servers. Our aim is to make them easily locatable on the Web with the help of the CERN Library's \emph{Preprint Catalogue} in several formats (\PS, \PDF, \GIF). We review the conversion procedures and give some details on some massive production trial runs to directly genera te \HTML{} from the \TeX{} sources. We conclude with a discussion of recent developments in the framework of the \XML{} (and \MML) efforts which should ease the support of mathematics formulae in W eb browsers.