Meta data is data about data. Data is meaningless without a context, and that is what meta data supplies. Meta data is used all over the place. It is called something like 'label', 'table heading', 'unit' (of measurement), or 'description'. Meta data is actually data, and itself has meta data. While meta data can be manipulated as data, care is needed if the relationship to its data is to be maintained.
As meta data is data the same rules apply to it. There is not an indefinitely long chain of meta data because meta data is a 'model', and a (useful) model will always be simpler than the thing it describes. So, the chain rapidly stops. Meta data cannot (directly, usefully) describe itself; a lot of 'paradoxes' go away when this rule is applied.
Meta data is classified into two types: 'system' and 'user'. The context of meta data is important. The original person who is responsible for data being in a system is critical 'context' meta data.
There are two sorts of meta data, properties and associations.
Properties are a basic part of a thing, some initial (entry to a system) measurement or characteristic. Such as: size of the data, data sequence (e.g. time-ordering), where and how the data originated (which includes the person who originated the data), or the person who owns the data.
Associations are relationships to other data that people make, which are convenient for their purpose. Such as: the 'why' of the data, pre-entry (to a system) associations (e.g. ownership) with the data, and the internal structure of the data.
Properties are 'system meta data', associations are 'user meta data'.
Meta data can be quite high level. However, it cannot reliably describe anything beyond the bounds of a system in which it operates. 'System' can be a very general concept. Without any aliens or artificial intelligences 'product of human society' will be meta data that applies to anything. Theists may differ. Meta data which describes the precise instrument which created the data is often useful.
Within human society '20th Century English' might describe a great deal of data, possibly qualified by 'UK English'. Then there might be specialised vocabulary or jargon for a specific profession, such as 'Information Technology', an organisational role, such as 'Systems Analyst', or another grouping, such as a club, friends, or an (extended) family.
Note that while 'language' is closely related to 'culture' they are not identical.
Who is the originator, the person responsible for putting the data into a system, is critical meta data. Though, the originator can be a role or other grouping, such as a (medical) doctor/patient relationship, rather than an individual.
The originator is responsible for the data being in a system. On that basis they should say how the data is used. The default is that the data is 'private', only usable to them; they can change or delete it as they like. So that data is usable to (specific, or all) others it is 'published'. This new data has meta data which says how it (and its meta data) can be used. It is possible to publish just meta data, describing potentially published data.
Meta data is critical to effective use of data. The two must be bound together, as neither makes sense on its own. Unless it is known where data comes from, what other data it is usefully related to, and who originated it, it can't be best used.
Meta Data on Wikipedia
(c) ROMsys Ltd, December 2013, permission given to use for non-profit making purposes
Back to 'Notes & Papers' index