eScience can be defined as joint research in Computer Science and other domains, to let scientists from these domains conduct their research faster, better or in a different way, while at the same time advancing the state of the art in Computer Science. As such, it involves joint cooperation of computer scientists with researchers from,e.g., (A)rchaeology to (Z)oology, covering all knowledge domains. While at the same time offering many exciting research opportunities for computer scientists, this poses very many challenges, including those inherent to the very nature of interdisciplinary research – such as issues of distinct work methodologies, very different vocabularies, and heterogeneity at all levels (from work profiles to infrastructures, from data to devices,software and people-ware). This talk will discuss one of these challenges – that of enabling experiment reproducibility for such heterogeneous research environments – and the many open problems whose solution benefit from innovative computer science research. It will draw from real life examples, and concentrate on the design and development of a suite of workflow-centered tools.