• No results found

In order to constrain the emerging terms within the developing heavyweight ontology logical rules were required. These rules would have to constrain the terms within the combined contexts of both the MI systems domain and the requirements for interoperability.

The research described in Section 5.5 which resulted in the definition of MI generated some basic defining statements which could be used to help define logic and relationships later in the process. These were relatively loose descriptive statements that proved to be useful in creating a clear MI definition but required a level of interpretation to generate logic and relationships as described in Section 6.6.

Based on the findings from the literature review and industrial investigation (see Chapters 2 and 3) the challenges to systems interoperability could be grouped into ‘themes’ and these themes were then be used to identify constraining logic, to which a type or instance within the ontology must conform if it is to meet the requirements for interoperability. The logic themes and example logic are shown in Figure 6-38:

Figure 6-38 - Logic development themes developed from the challenges to interoperability research

The next approach was to focus the subject matter experts on the purpose of the ontology and list rules and relationships between individual terms using the UML taxonomies as a prompt. The development of these rules was structured around the key themes previously

identified as the challenges to system interoperability. This work resulted in the following statements:

 Each data instance has a technical authority level

 Each data instance has an operational authority level

 Decisions about systems can only be made by super users or administrators

 Non standard systems specifications require manager and super user approval

 A Person has a technical authority level

 A Person has an operational authority level

 If a person’s operational authority level is less than the data authority level they cannot access the data

 If a person’s technical authority level is less than the data authority level they cannot access the data

 Having technical authority does not give a person operational authority

 A system should use a standard data structure

 Anything with obsolete or deleted status should not be used

 For something to be approved, all data must be proven or approved.

 A change decision makes proven data unproven.

 If data is not approved it is be default unapproved

 The use of unapproved data within a system renders the system output unapproved

 Users or employees cannot change unproven to proven or unapproved to approved

 Outside of its validity timescale data becomes unapproved

 Equipment maintenance requires maintenance data

 Output data should must have a timestamp

 Finish data Timestamps cannot precede a start timestamp.

 Increasing a person’s authority requires some sort of training

 For a person to work with a system there must be a HMI

 For two or more systems to work together there must be a system-system interface defined

 For 2 systems to measure the same metric they must use the same data types.

 For decisions to be consistent across people and systems the same data or metrics must be used.

 A system requires sustainment processes to remain consistent.

 People require training to remain consistent

 The durations between training to maintain person consistency must be defined.

 Maintenance intervals for a equipment must be defined

 The consistency of people and systems should be monitored.

 Metrics must be defined for any form of monitoring

 Targets must be defined for any form of monitoring.

 Reactions must be defined for any form of monitoring

 A system must be described by a functional and non functional spec

 If a system uses IT there must be an IT spec.

 Proactive feedback uses input data

 Reactive feedback uses output data

 Responses require output data

 Using non standard data structures requires super user authority

 Specifications must be approved by people with leaders and super users authority or higher

 System inputs, outputs must be specified for a system

 System resource and constraints should be specified for a system

 Metric targets have a validity timescale

 Metric data has a validity timescale

 A system has a data type if it is an input, output, resource or constraint

 If a system outputs a program it is a programming system

 If a system has data inputs but no data outputs it is a data storage system

 If a system is the only store of a data type it is an archiving system

 Data should not be stored in 2 archiving systems

 If data is stored in 2 archiving systems one must be declared the authoritative data store

 Data should have a data retention category (mandatory/ non mandatory)

 Data should have and export control category (export controlled/ non export controlled)

 Data should have a security category (public, private, secret, top secret)

 systems should have a security, export control and data retention category

 If an system doesn’t have a security category if should be set to public

 If an system doesn’t have an export control category it should be set to non export controlled

 If an system doesn’t have a data retention category if should be set to non mandatory

 systems that have security category of public can only input or output public category data

 systems that have security category of private can only input or output public or private category data

 systems that have security category of secret can only input or output public, private or secret category data

 systems that have security category of top secret can only input or output public, private, secret or top secret category data

 systems with an export control category of non export control can only input or output non export control data

 systems with an data retention category of non mandatory control can only input or output non mandatory retention data

 If a data item is the input of one system and the output of another these systems are series associated

 If a system outputs data to another system it is an input for that other system

 A system functional spec must define a performance metric and metric target

 Systems that share the same input, outputs resources or constraints are parallel associated

 If a system outputs data that is not input into it, it is the data source for that data The term pool, taxonomy and relationship development the logic development were all iterative and interlinked, with developments in one area requiring a form of regression

assessment to ensure all logic and relationships still held true and comply to the constraining logic. This logic was also checked against the BPMNs (representing real systems and

process) to ensure the correlation with reality. At this stage this was a manual activity, and the complexity due to the number of terms and relationships was highlighted by the number of issues that would be highlighted during the ontology coding activity, where these checks were carried out by the software in a very rigorous manner.