The news in the UK is currently full of the Supreme Court’s ruling on the prorogation of parliament. The prorogation has been judged to be ‘void and of no effect’. The consequent actions thus never happened: as the judge stated, ‘ when the Royal Commissioners walked into the House of Lords it was as if they walked in with a blank sheet of paper ‘.
What has this to do with meta-data?
Courts are higher authorities to which we turn to resolve disputes. They interpret the facts from apparently opposing viewpoints and determine the ‘right’ outcome, the ‘right’ interpretation of reality. I suggest they are a useful metaphor for what is needed in our management of data.
As ‘big data’ technologies are applied, increasing amounts of time and effort are spent on understanding, analysing, cleaning, formatting and cross referencing data items. This ‘preparation’ often dominates the time for fruitful data science; more importantly, the meaning of the resulting data is often not known. Each data set is an island and the meaning and linking of the data is simply assumed.
Similarly, as applications and user services evolve, their data sets evolve with them, compounding these problems. Data cannot be analysed without contextual knowledge, nor can it be reused across services without risk of errors due to the lack of that context. The user is not offered appropriate safe reuse of her data, or worse, receives incorrect service due to errors and misrepresentations.
Together, meta-data and reasoning over meta-data are the means to represent this missing context, and enable automatic analysis and appropriate reuse. The required characteristics are described below, elaborated by the legal metaphor.
Syntax – what format is a data item? How is the item encoded, serialised and decoded? For example, what is the syntax of an address? e.g. a unique property reference number, a post code, a ‘three words location’.
In legal metaphor – syntax is the basic evidence, without a standard record of a communication there would be no material to argue. Without a common view of syntax both sides couldn’t even read the material and so start to determine its value.
Semantics – what is the meaning of the data item? A given data item may be seen to be syntactically a date, but is it a date of birth?
An item may be an address, but is it a correspondence address, a residential address, a business, a holiday address or a holiday home?
In legal metaphor – semantics presents meaning of the evidence by presenting the context, who, what and why. Semantics is the attribution of meaning to the basic evidence.
Timeliness – is the data item current? is it historic or is it ‘expired’? Does it have a ‘natural decay over time’? Mr and Mrs Jones were married in June: are they married in December? Of what year? I lived in Washington in 2001, do I still live there?
In legal metaphor – the context of ‘when’, provides currency of evidence, its relevance to a decision and indicates causal and dependency relationships. In the extreme, evidence may be too old to be admissible, or, as in the prorogation ruling, it can render consequent actions void!
Assurance – is the data item correct? Is it correct enough to rely on for our purpose? What are the standards of ‘correctness’ and the auditors of compliance? Mr and Mrs Jones were married by a registrar in the presence of a witness in accordance with the jurisdiction of the location. My identity is assured to a certain standard to enable me to open a bank account. An online retailer ‘knows’ I am at an address because I receive parcels there. A retailer decides I am old enough to buy alcohol.
Assurance may be applied both to individual items of data and across data items. For example, a certain address is a valid address, it is holiday home, and Mr Jones resided there for a period of time in June 2019.
In legal metaphor – given evidence, its meaning, timing and assurance, what is ‘true’?
In the UK’s prorogation case, applying the law and case history, analysing the evidence, its meaning, timing and levels of trustworthiness, eleven senior justices unanimously determined that prorogation was unlawful and thus ‘void and of no effect’.
Although secondary to the main intention of this note – which is to stimulate discussion about the importance of meta-data – the question of implementation may be raised.
The above characteristics of meta-data can be readily applied to current business problems in building digital services, protecting organisational assets and learning, and safely reusing data assets, thus providing better and more effective user journeys.
Implementation will be incremental and there is no need to wait for universal standards: it is sufficient to make a business case in the context of an organisation with data siloes (for example a government department), or for an eco-system of parties who wish to inter-operate (for example KYC/AML financial services).
Existing semantic web standards (RDF, OWL, SKOS, FOAF, etc) can be used as the basis of such scoped standardisation to ensure openness and inter-operation.
Organisations or eco-systems can represent the meta-data and the rules necessary to appropriately reuse data items across its services.
We can ask and be assured: what is ‘true’? We can avoid ‘walking into the room as if with a blank piece of paper’.
If you would like to talk to us about meta-data please contact via the form https://vinesolutions.co.uk/contact-us/, or see more in Vine’s service offering at https://vinesolutions.co.uk/services/. Thank you.