OASIS Classification of Everyday Living (COEL) TC

The purpose of the COEL TC is to create the standards necessary for the successful formation and growth of a business ecosystem aimed at providing personalised services.

Personalised services, such as those that emerge from loyalty card schemes or branded wearable devices, form a large and growing market. However, these solutions exist only as narrow, vertically-integrated propositions. As such, information exchange between various domains is hampered and the cost of creating complete vertical solutions restricts uptake to large players, excluding the small to medium enterprise market and the public sector (where personalised services could offer the greatest human value). All these solutions share one thing, the need to understand what we do as humans and what we might do next. This behavioural aspect of personalised services is at the core of the COEL TC aims.

Any system or device that collects data for the provision of personalised services is likely to be handling information that individuals would classify as personal or sensitive. The objective of the COEL TC in this domain will be to create transparent technical frameworks that respects the data of individuals, can be trusted by individuals and are compatible with the needs of regulators in a wide range of jurisdictions.

Building the infrastructure for these services requires many distinct capabilities, including capture of behavioural data, a comprehensive framework for describing behaviour, secure transmission and storage of this data, maintenance of a separation between this data and directly identifying personal information, the development of complex data analytics to interpret the data, and imaginative strategies to deliver value to both business and the end user through personalised service offerings. Standards and best practices are required to address each of these so that an ecosystem can function in which each actor focuses on their strengths providing value to the rest of the ecosystem.

The COEL TC will develop and publish the standards necessary to realise an ecosystem for personalised services. The specific projects targeted by this TC are:

• development of a taxonomy of human activity (the Classification of Everyday Living),
• a format and protocol for the representation, transmission, storage and referencing of individual instances of human behaviour (called Behavioural Atoms) and their associated context,
• an ecosystem architecture delineating the roles and responsibilities of ecosystem players, and
• a set of technical requirements for implementing this architecture.

In an ecosystem of this type, business-to-business interactions can be defined and conducted to share data in appropriate ways about personal behaviours (such as consumer brand consumption). The source of this model for a business-to-business ecosystem of this kind is Coelition <coelition.org> (see the Contributions referenced below). Within such an ecosystem, participants could agree to share and/or license consumer-facing brand and behaviour instance data (Behavioural Atoms). A complete specification for the necessary processes for such an ecosystem may include defining intermediary roles, the patterns for interactions among participants, and methods for determining compliance with these specifications.

The COEL TC will develop the documentation, in the form of OASIS specifications, necessary for operating the Coelition ecosystem from the technical and process perspectives. A complete description of the Coelition concept can be found here [1] and on the Coelition website [2]. The deliverables described in the next section address the following three specific domains, which are within the scope of this TC: A model, a process and a protocol.

Classification Model: The TC will develop a classification of everyday living. The aim is to produce an agreed taxonomy of things that people do. The hierarchical structure of the taxonomy will provide decreasing granularity of activity with depth in the hierarchy. For example, at the highest level a person might ‘Travel’. As the activity becomes more completely described, the granularity decreases and we have ‘Travelling by land’, then by ‘Car’, and finally by ‘Sports car’. The TC will also document the guidelines used to develop this taxonomy, e.g. that each level should be complete and mutually exclusive, or what the finest granularity of activity should encompass. The normative representations, and any serializations, of the classification model will be determined by the TC.

Process Architecture: The TC will describe the architecture of this business ecosystem. This will include descriptions of the business-to-business actors in a conforming ecosystem, and their respective roles. The architecture will include a privacy model that specifies how actors follow certain technical requirements in order to ensure that individual privacy is adequately protected. These technical requirements will be documented by the TC and should be detailed enough to facilitate audit or accreditation.

Protocols: In terms of technical protocols, the TC will describe protocols for those interactions among components of an ecosystem that are mandatory in order to comply with and sustain the process architecture. The expected protocols are described under "Deliverables" below. It is intended that the TC and its deliverables only describe what is absolutely required for the ecosystem to function, as the aim is to leave as much scope for innovation and business differentiation as possible.

The manifesto chapter of “Data to Life” [1] is available as a description of the envisioned human behaviour data ecosystem architecture and guidelines.

The following are specifically out of scope:

• The TC will not develop or provide branding or promotional content.
• The TC will not develop or provide consumer-facing aspects of an ecosystem; its deliverables are intended to specify business-to-business data exchanges and sharing arrangements.
• The TC will not develop or provide legal advice.
• The TC will not make specific recommendations on compliance with laws or regulations in different countries or regions.
• The TC will not address issues related to how Behavioural Atoms are technically detected and measured. For example, the protocol will structure how information about sleep or sleep quality is described and shared, but it will not specify how a wearables device detects or measures sleep quality.
• The TC will not address the types of complex queries that are likely to be performed against a store of Behavioural Atom data.
• The TC will not certify implementations of any standards.

The following deliverables will be produced by the COEL TC. Each is expected to be a separate document, although some may be divided further or refactored by the TC to facilitate development. The order listed is expected to be the order of development, although work is likely to proceed in parallel and the TC may determine which projects are completed at what time.

• The COEL Model: the Coelition's contributed model (see below) is described as a four-level taxonomy, with nested levels describing everyday human activities in increasing detail. This deliverable is intended to be capable of representing and meaningfully classifying any and all routine human behaviours. It also will include guidelines for creating classifications itself, e.g. completeness, mutual exclusivity, and the granularity of each level.
• The Behavioural Atom Format and Exchange Protocol: this deliverable will define the message format (probably JSON, together with a compressed binary format to facilitate device integration) and a protocol (probably REST style and Bluetooth) for data exchanges that are capable of describing, querying and reporting about a human activity, using the COEL model classification, as well as the context in which it took place, i.e. time, location, etc. The format will also include extension capabilities. Some predefined, common extensions will also be defined (possibly as a separate document).
• The Ecosystem Process Architecture and Technical Requirements: this deliverable will define and describe the roles and responsibilities for the various actors required or typically present in an ecosystem of business-to-business data exchanges using the COEL model and Behavioural Atom data, particularly focusing on what information each of them should handle, and how privacy data should be managed. These technical requirements are expected to be usable as a basis for a standardised audit framework in support of certification of ecosystem actors. Actors will be encouraged to host low cost development services to facilitate adoption of the standard.
• The Identity Authority Interface Specification: defines the interface protocol for an Identity Authority (IDA). The IDA is a central web-based service, required in any ecosystem that conforms to the Ecosystem Process Architecture, that statelessly provides unique, signed IDs. These IDs are used to register actors within the ecosystem and are then requested by an actor wishing to register an individual person within the ecosystem (and thus enter into data exchanges about that person's Behavioural Atoms). The IDA's operation must ensure the integrity, universality and interoperability of the ecosystem. To ensure no conflict of interest, the IDA must be controlled by an independent actor who does not provide Behavioural Atom services and it is expected that IDAs will be non-profit organisations. Coelition anticipates establishing a single IDA service for its ecosystem.
• The Minimal Interface: A minimal interface will be defined for the transmission of Behavoural Atoms to the data engine, and the registration of parties, and to allow appropriate access to Behavoural Atoms that have been previously stored, including limited query capabilities. However, as this interface is expected to be a high-value and differentiating aspect for Behavoural Atom storage actors, it should be minimally described, and will not be defined or constrained in great detail.

The TC may also produce such explanatory or educational materials as it feels is necessary or useful in helping users understand and implement the specification.

This TC will operate under the RF on RAND Terms IPR mode as defined in the OASIS Intellectual Property Rights (IPR) Policy.

The audience for this TC includes any party interested in defining the nature of the Coelition ecosystem, and consensual, privacy-respecting exchanges of human behaviour generally. Most particularly, these projects should be valuable to parties interested in implementation of Behavoural Atom information stores; the development of devices and systems that can identify and create Behavoural Atom data; and those looking to develop services around the data flow of individual and aggregate Behavoural Atom data gathered. Within the scope of the COEL taxonomy, social scientists and psychologists from industry, government or academia may also be interested in participating in this TC.

The COEL TC will conduct proceedings in English and publish the deliverables in (British) English.