Metadata Definition

Standardizing metadata and terms doesn’t require that the same metadata and terms are used across the whole site collection or ECM repository (like OpenText Content Server) but only that the metadata and terms used are effective in describing the content to which they are applied.  Much time can be lost on trying to standardize metadata and terms and the results do not generally yield comparable value for the time invested.  It is best to consider metadata and terms from a relational perspective.  That is to say, don’t consider how each value can uniquely describe documents but how all the values ascribed to a document uniquely describe them.  Rather than focusing on metadata and term standardization, a better exercise is focusing on what would constitute the smallest set of values to create that unique definition with the understanding that it is the value together as a group which are important.

Meaning is transitory.  Though it is useful to try to simplify the metadata attributes and terms which are being used over time, the effort required to develop a common vocabulary across a large group may not be worth the effort involved.  When we have studied efforts like this, we find that these projects consume a great deal of time and resources without achieving the envisioned results (an organization holding all of their content in single, common, system using the same terms and descriptors).  The problem is change.

Organizational restructuring and personnel changes will always result in a group using a system from being different from the group which devised that system.  The definitions for terminology developed by one group will not be the terms desired by a new group.  Therefore, as a group changes, the terms increasingly have less relevance.

An implementation team should then be more focused on creating descriptors and encouraging people to use them then to spend significant effort on standardizing the descriptors.  Since it is always possible to modify descriptors later or create proxies that tie one descriptor to another, conflicting terms can always be resolved.  What is essential is that the content is described.  As long as this happens then it will always be possible to query and retrieve content later.

In short, an implementation team should attack the low hanging fruit, solving the obvious problems, without going into interminable cycles of metadata and classification definition that have been common within organizations in the past.