Change Management

Even the most carefully curated collections experience change for a variety of reasons, and those changes can impact the preservation outlook for the collection. For example, after a preservation package has been created, ingested into a preservation repository, and stored, a curator may need to add or change a file to that package, particularly if it is a meaningful body of content (i.e., a collection, not just a batch of unrelated files). Perhaps a file will need to be re-digitized and re-incorporated into the preserved collection or a metadata record or OCR file will need correction. Such changes can be significant and may impact the preservation package.

Prior to moving content into preservation storage, institutions should grapple with the implications of file changes and how to document and store changed files within a preservation repository and its given workflows. In some cases this may mean that the file name is ascribed (manually or by the system) with a new creation date or an incremented version count. In other cases there may be simple non-numerical re-naming conventions that can be ascribed to indicate the change. Relationships (inter-linkages), and/or inheritances of any previously assigned metadata also need to be considered.

As a first step, institutions must decide whether updated files will simply replace any previously collected and stored file—if so, a new file may not require the ascription of any change information. However, the issue of information loss and irretrievability in such over-writing should be weighed carefully. Growth of collections is also important when it comes to change management—particularly when working with external preservation partners. Such on-going exchanges of new and/or refreshed data need to be well coordinated, synchronized and documented across all the partners (metadata can sometimes help with this but other documents may also play a role). Whatever the situation, a reliable, consistent and clearly communicated policy and practice should be the goal. Creation or modified-since date information (YYYY_MM_DD or YYYYMMDD) is one of the easiest solutions and has already been particularly useful and standard when it comes to distinguishing digital newspaper content—it also has some use value for indicating changes (as already mentioned above).