The principles and procedures for managing technical data changes specifically addressed by Product Lifecycle Management systems are examined in the context of fast changing and serially produced technical objects. Attempts to combine functionalities provided by PLM systems lead to choices that are irreconcilable with fully automated management. The expected outcomes should thus be put into perspective. The theoretical issues relate to the foundations of identity and discernibility, oneness and multiplicity, changes and invariances explored in this paper. Along the lines of E.F. Codd's algebra, a formal approach to object naming is proposed in order to communicate strategies for managing relationships. The interchangeability of objects and relations, compositions and use contexts is relatively and conditionally redefined. The strategies for managing configuration changes are optimized using two generic principles that aim to preserve invariances or design new constructible ones. Those principles come into play depending on how industrial settings combine redesign (innovation) and production (serial production) rates. These investigations drew upon multiannual research conducted in manufacturing (in particular, aeronautics) as well as health care, and led to experimental corporate applications. The major PLM solutions in the market were examined.
