Another major benefit of HAC is its ability to cleanly access other HAC file systems, and other CBA mechanisms (possibly remote). In this section, we shall use name space to denote either a traditional file system (which provides path name-based access), a CBA mechanism, or a HAC file system. Connecting different file systems across a distributed system can be done with mount points ([neum:92], [rp:93]). Mount points define new name spaces within which path names can be resolved. They allow different file systems to share certain directories so that they can access each other. HAC supports such mount points, which we call syntactic mount points. But we want to do more. We want to connect "semantically" so that we can evaluate queries against different name spaces, even if these name spaces do not allow us to organize information hierarchically (e.g., commercial search engines on the web). We want to allow users to use data from anywhere, create semantic directories anywhere, and in general treat the remote file systems and CBA mechanisms as if they were local. To achieve such rich, transparent connection, we must "decouple" the part of HAC that provides path name based access from the part that provides content based access, so that both can be used independently of each other. This is close to impossible to do if we restrict ourselves to syntactic mount points. We therefore introduced semantic mount points in HAC.