There is no uniform datamanagement solution for scientific database. They are often embedded in complete application scenarios, programming frameworks, workflow systems and web-services. Often developed with the limitations of the underlying platform as a fact of life. However, this hinders a clear view on the data management challenges at hand.

Therefore, for a few science domains we give a concrete micro-world of functional requirements and performance challenges. In the spirit of Jim Gray's adagium we take the core of the database definition, the 20 most important queries and progress from a working to working system. The queries abstract away from the application specifics without loosing touch with the growth encountered in real-life use. Taking small steps uncovers the hurdles to be taken. Experience with the SkyServer demonstrator illustrate that science nuggets will be found on the way.

The SciLens project capitalizes upon knowledge, skills and insights of volunteers from different science fields. The key demonstrators are each shephered by a member of the MonetDB research team.