Updating mxd documets that drive web services in and enterprise environment can be hard to manage sometimes. At 10, updating of data sources became easier with the use of pyton. However, if the mxd is built using stand alone tables joined to layers, those joins do not and cannot be easily updated when data sources change. This is especially troublesome for users in a multi-tiered environment where the mxds and services need to move from test to staging to production with data source updates at each move. Now, the joins do not get updated with the data source change and the mxd needs to be manually updated (join removed and recreated) which requires work and introduces the chance of errors being introduced in the change.