Keywords: expertise sharing, collective memory, organizational memory, crowdsourcing, pervasive environments, collaborative help, escalier, collective helpEscalier is a new architecture for community-sourcing configuration settings, and by extention other types of pervasive data such as activity traces. The next generation of computational environments are likely to be so complex that users will not easily be able to keep up.Using Escalier, users can find out what system configurations will be stable in return for giving the community their configuration data. By doing this, both parties get something valuable: End-users can determine whether their system configurations will work, and the community as a whole gets a collaboratively-built datastore, or collective memory.Escalier operates by making probabilistic assessments of users' reports and of the users' expertise levels. The community collectively constructs a Bayesian map of configurations' stabilities.Escalier consists of a core backend and web services that allow web-based applications. These applications allow users to assess their configuration for stability and obsolescence, as well as make recommendations for additional enhancements.