Metadata - an extended definition
A philosophical note on metadata as needed a priori and impossible to avoid a posteriori
In this post present an extended version of the definition of metadata that I put forward in my upcoming book Fundamentals of Metadata Management. It’s a philosophical dimension that is useful for you - it’s like a secret key to my book.
A priori and a posteriori
Knowledge can be categorized into two:
A priori
A posteriori
This distinction - in Latin - has existed since antiquity and is most famously discussed by the German philosopher Immanuel Kant.
A priori (before) knowledge and thinking is found disciplines such as logic and mathematics. This is thinking that has no need to be verified empirically, by observing nature or humanly created objects, it occurs before such activities.
A posteriori (after) knowledge and thinking is found in disciplines such as physics and history. These disciplines are relying on observations in nature and culture and because of that, this type of knowledge can only occur after empirical observations.
In metadata management, there is a constant push and pull between a priori and a posteriori. We design standards, technologies, structures, reference architectures - and we have meetings about how metadata management should be performed. This is metadata management a priori - beautiful, logical… and impossible.
What I put forward in Fundamentals of Metadata Management is the ugly truth of Metadata Management a posteriori. It’s a long essay of how metadata management is actually performed in companies - observations throughout my career and confirmed by a substantial amount of my readers and more broadly in scientific studies. Metadata management is performed throughout multiple instances of many different types of applications, that all mirror the IT landscape at the metadata layer. Observing real metadata management, how it actually happens in companies, it far from the ideal architectures we think about a priori - and trying to solve the problem of metadata management only with more a priori advice, will fail. Companies will always have complicated ecosystems of metadata repositories for all kind of purposes, inherited for all kinds of reasons.
The definition of metadata extended
This is the definition of metadata that I find in my book:
Metadata is a description that is both attached with what is described and placed somewhere else, to make what is described discoverable and manageable.
Taking into account a priori, and a posteriori, the following can be added to the definition:
Metadata is a description that is both attached with what is described and placed somewhere else, to make what is described discoverable and manageable. Metadata needs preconceived structures, technologies and architectures that are all ideal, but metadata will always exist in a reality that is not ideal. Accepting this is the only way to improve metadata.