Metadata discovery
Metadata is the ‘glue’ of coherence which actuates people, processes and systems across ecosystems & beyond. Metadata is often described as ‘data about data’ so on that basis, why doesn’t metadata receive the same due care and attention as ‘data’? If anything, metadata is more important and valuable than ‘data’, yet it’s treated like a Cinderella. Everything we said about data, applies equally to metadata! Ignore your metadata at your peril.
On the roadmap to the future of computing, it’s important to know your metadata and its condition. Typically, this will be patchy according to its nature:
- object or programmatic metadata is likely to be top quality and eminently usable;
- application metadata: may be extensive and of good quality but not easily discoverable or usable owing to the opacity of some widely-used packaged applications (ERP & CRM), which also affects visibility of their data structures & dictionaries, and an organisation’s ability to use and integrate those;
- dark & unstructured data: typically lacks business-relevant metadata;
- archived data: typically has poor or inappropriate metadata, consigned to digital oblivion;
- legacy estate: typically has hard-to-extract metadata (especially relating to lineage) originating from legacy, end-of-life or proprietary applications & which predate cloud. The difficult of securing such metadata is inversely proportionate to its potential comparative value.
Missing metadata are problematic. Remediating gaps or errors in metadata is also problematic. There are constraints regarding the extent to which metadata can, or should, be remediated – and still remain valid:
- some metadata are editable, others are not;
- some metadata are not editable, nor should they be as they form part of a record;
- some metadata are set by the system automatically and typically form part of a record – or become pointless if they are edited;
- inaccurate metadata may persist but ought to be corrected e.g. inherited from previous saves in back office applications;
- custom metadata may not persist but should be retained e.g. after resources, instances or libraries are moved -and may be difficult to retrieve.