Interoperable Data Exchange

The application schema is the basis of a successful data transfer between different systems as it defines the possible content and structure of the exchanged spatial data, which means it covers both semantic and syntactic interoperability. Applications (software) and users (people) should interpret data and information in the same manner to ensure they are understood as it was planned by the producer of the data.

The general idea of the data interchange between two different systems is shown in Figure 1. System A wants to send a dataset to system B, what follows, system B has to be able to use data from system A. To ensure a successful result of this process, it is necessary for both systems to determine a common application schema I, an encoding rule R and a transfer protocol (ISO/TC 211, 2011).

Figure 1

Overview of the interoperable data exchange (source: own elaboration on the basis of ISO 19118 (ISO/TC 211, 2011) and the CEN/TR 15449-3 (CEN/TC 287, 2012b))

According to the ISO 19100 suite of standards, the application schema used for encoding should be written in the UML conceptual schema language, in compliance with ISO 19103 (ISO/TC 211, 2015a) and ISO 19109 (ISO/TC 211, 2015c). These documents provide a set of rules for how to properly write the application schema, including the usage of standardized schemas to define feature types. a sender and a receiver of spatial data must have an access to the application schema. It is even recommended to transfer it before data interchange proceeded, to allow both ends of this transaction to prepare their systems by implementing appropriate mappings and data structures corresponding to the application schema (ISO/TC 211, 2011).

The encoding rule is an identifiable collection of conversion rules that defines the encoding for a particular data structure (ISO/TC 211, 2011). In accordance with ISO 19118, it specifies, among others, the syntax and structure of the resulting data structure and is applied to the application schema specific data structures to produce system independent data structures suitable for transport or storage. The conversion rule, by contrast, defines how a data instance in the input data structure is converted to a data instance in the output data structure (ISO/TC 211, 2011).

To conclude, the UML application schema determines the possible content and structure of the interchanged spatial data, whereas the encoding rule defines the conversion rules for how to code the data into a system independent data structure.

GML Encoding Rule

A good example of encoding rule mentioned above is the XML-based encoding rule for neutral data interchange described in detail in ISO 19118 standard (ISO/TC 211, 2011). It is compatible with the UML and defines the encoding rule based on the XML (eXtensible Markup Language). An overview of this encoding rule is illustrated in Figure 2. Within the XML-based encoding rule exists two sets of conversion rules. The first one specifies a mapping from the UML class definitions in the application schema to the type declarations in the XML Schema. The second one defines a mapping from the objects in the instance model to the corresponding element structures in the XML document.

Figure 2

XML-based encoding rule (source: own elaboration on the basis of ISO 19118 (ISO/TC 211, 2011))

The discussed XML-based encoding rule is applied in the GML standard that is an XML encoding based on principles specified in ISO 19118. The GML provides a common XML encoding for spatial data, along with an open, vendor-neutral framework for the description of geospatial application schemas for the transport and storage of geographic information in the XML (ISO/TC 211, 2007b).

Whereas the GML application schema is an application schema written in the XML Schema in accordance with the rules specified in ISO 19136 (ISO/TC 211, 2007b). Additionally, it has to import the GML schema that compromises XML encodings of a number of the conceptual classes defined in the ISO 19100 series of International Standards.

In the case of INSPIRE, the GML is usually recommended encoding, unless otherwise specified for a specific data theme. By way of illustration, for large volume coverage data such as orthoimagery or computer simulations (e.g., weather forecasts), other, more efficient, file-based encodings (e.g., geoTIFF) may be defined as the default encoding language (Tóth et al., 2012).

Data Specifications

Both, UML and GML application schemas are commonly used in the European SDI, as well as in the National SDIs (e.g., in Poland). As data models, they are included in the data specifications that usually contain other relevant requirements about data, such as rules for data capture, encoding, and delivery, as well as provisions of data quality and consistency, metadata and so on.

In the broader sense, data specification can refer to both the data product specification and the interoperability target specification in SDI (Tóth et al., 2012).

The data product specification is a detailed description of a dataset or dataset series together with additional information that will enable it to be created, supplied to and used by another party (ISO/TC 211, 2007a). In line with the ISO 19131, it provides a definition of the universe of discourse and a specification for mapping the universe of discourse to a dataset. It may be used for production, sales, end-use or other purposes (ISO/TC 211, 2007a).

One of the key and mandatory items included in the data product specification is the data content and structure information, which means an application schema. In case of a feature-based data product, this element is described in terms of an application schema and a feature catalogue (this issue is beyond the scope of this paper and will not be discussed later, for more information, see the ISO 19110 (ISO/TC 211, 2015b)). The application schema provides the formal description of the data structure and content of the data product. It is a conceptual model expressed in the UML (conceptual schema language in terms of the ISO 19100 suite of geographic information standards) in accordance with the ISO 19109. This model should include the representation of feature and property types (including attribute types), feature operations and associations, inheritance relations and constraints, where attribute types cover descriptive, geometric and temporal properties, whereas associations include spatial and temporal relationships such as topological as well as non-spatial relations (e.g., ownership) that occur between feature types (ISO/TC 211, 2007a). In case of coverage-based and imagery data, a coverage is considered a subtype of a feature and behaves like a function, which returns one or more feature attribute values from a direct position within a spatiotemporal domain (ISO/TC 211, 2007a).

The interoperability target specification is used for transforming existing data so that they share common characteristics (Tóth et al., 2012). In the case of INSPIRE, to achieve interoperability in the European SDI, data specifications have been established for the 34 spatial data themes. The Member States of the EU can use these documents to create new datasets or to transform existing datasets according to the specifications by mapping the existing models to the models defined in the data specification guidelines (CEN/TC 287, 2012b).

The Head Office of Geodesy and Cartography, the main coordinator of SDI creating and functioning in Poland, adopted a similar approach to implement the INSPIRE Directive and establish the National SDI. The law of the infrastructure for spatial information (Act, 2010) in Poland was approved. It is a transposition of the INSPIRE Directive to the national law. This action involved a necessity of many acts and relative laws’ changes, among others, the law on geodesy and cartography (Act, 1989). The existing (very often obsolete) instructions and guidelines were replaced by regulations of the Cabinet or the responsible Minister. These documents became annexes to the law on geodesy and cartography as well as put some of the INSPIRE Directive recommendations into action.

An integral part of these developed regulations are the UML and GML application schemas that define information structures of spatial databases, corresponding to each regulation. These schemas were prepared in accordance with the ISO 19100 series of geographic information standards to ensure interoperability of spatial data sets and GIS applications. Regulations, data product specifications in terms of ISO 19131, cover the whole legal and technical issues regarding the geodetic domain in Poland. This was a very ambitious challenge due to the methodology of the conceptual modelling and the usage of the UML and GML notations to describe the information content of databases, were applied for the first time in Poland.

Quality According to ISO 19100

The ISO 19100 series of geographic information standards share a common definition of quality, previously defined in ISO 9000, cited above (ISO/TC 176/SC 1, 2015). In general, quality issues are currently discussed in the following normative documents:

ISO 19157 provides the principles for the description of geographic data quality and specifies components for reporting quality information as well as procedures for the evaluation of geographic data quality.

In turn, ISO/TS 19157-2 concerns data quality encoding in the XML. It defines an XML Schema implementation of ISO 19157 and the data quality related concepts from ISO 19115-2 (ISO/TC 211, 2009).

ISO/TS 19158 establishes a quality assurance framework specific to geographic information that is based upon the quality principles and quality evaluation procedures identified in ISO 19157 and the general quality management principles defined in ISO 9000.

Interestingly, this standard introduces the concepts of ‘customer’, ‘supplier’ and ‘product’ derived from ISO 9000. According to both these documents, the customer is defined as an organization or person that receives a product, the supplier is an organization or person that provides a product and the product is a result of a process (defined as a set of interrelated or interacting activities, which transforms inputs into outputs) (ISO/TC 176/SC 1, 2015). Additionally, it was clarified that the supplier has provided the product via the process that can have some impact on the quality (ISO/TC 211, 2012). The ISO/TS 19158 arrangements are applicable to customers and suppliers of all geographic information.

However, the above standards consider quality mainly at the data level and the subject of the research activities is taking into consideration quality at the higher level of abstraction, that is, the level of data structures. It may be assumed that by ensuring the suitable quality of spatial data structures (application schemas), the quality of spatial data, including spatial sets and databases, can be increased significantly.

Quality Aspects in SDISD

I relies on standards and specifications in the domain of geographic information and information technology. Therefore, the above discussed ISO standards regarding quality are also relevant in terms of any SDI development and implementation.

In the CEN/TR 15449-1 view (CEN/TC 287, 2012a), quality issues, particularly checking the quality, are an important consideration for achieving interoperability. However, quality is not considered in absolute terms, but when it comes to user requirements (CEN/TC 287, 2012b). Users, first of all, require information about the quality of datasets to assess whether the datasets are useful for them or not (‘fitness for purpose’). The quality levels for each spatial dataset are defined using the criteria established by the ISO 19100 suite of standards, including completeness, consistency, currency, accuracy and usability. In turn, quality information concerning individual spatial objects is part of the metadata associated with the respective spatial objects and generally should be described as part of the application schema (CEN/TC 287, 2012b).

According to Tóth et al. (2012), from the point of view of SDI, poor data quality may compromise interoperability. Therefore, it can be generally stated that poor application schema quality also may impact negatively on achieving interoperability within SDI.

Producer Perspective

The quality of UML and GML application schemas at the producer level, that is the manufacturer or supplier level (in the considered case, the EU and the Head Office of Geodesy and Cartography), can be defined as a ‘formal quality’ or ‘legal quality’. This means that these application schemas (products) conform to the relevant standards of the ISO 19100 series of geographic information standards, as well as to the appropriate legal regulations concerning specific thematic issue. Additionally, such product characteristics as profitability (cost, market size) and competitiveness, including among others, technological leadership or an ongoing development of the product and process (Frąś, 2000), can also be taken into account. The concept of quality management, called the Deming cycle (Deming, 1986), known as the PDCA cycle, is linked with the latter feature. The continuous improvement of products and processes is an iterative four-step management method including Plan, Do, Check and Act stage. Similar approach was applied in the methodology for data specification development, presented by Tóth et al. (2012), with the effect that INSPIRE data specifications are revised regularly.

It is recommended that primarily, the methodology of formal quality examination for the UML application schema should consider its verification against the ISO 19103 (ISO/TC 211, 2015a), for example, choosing the proper data types, including geometry types, as well as ISO 19109 (ISO/TC 211, 2015c), generally correctness of the schema design. Conformity with relevant legal provisions in the scope of the given topic also should be taken into consideration.

In respect of evaluating the GML application schema quality, it is necessary to check its conformance with ISO 19136 (ISO/TC 211, 2007b) recommendations, mainly validation of schema structure, and also verification of their range of information.

When it comes to the evaluation of formal quality of application schemas, some previously prepared ATS can be applied. Examples of such ATS can be found in INSPIRE data specifications.

Consumer Perspective

The quality of UML and GML application schemas at the consumer level, which means their user (in the considered case, all users of the European SDI, as well as surveying and cartographic administration in Poland and also ordinary users of spatial data), can be called a ‘technical quality’. This type of quality includes some functional and non-functional requirements for application schemas. By way of illustration, economy, reliability, maintainability or generally user comfort (Frąś, 2000) can be recognized as functional user needs. On the other hand, aesthetic design or image building (e.g., maintaining database and generating data that conform to obligatory application schema) are examples of non-functional requirements.

It is recommended that technical quality examination for the UML application schema should include, among others, verification of association roles, which are used in the UML for describing relations between classes. Their lack in the UML application schema causes the impossibility of encoding such association in the XML Schema. Besides, checking whether there exists double reference between classes (classes can indicate to each other what leads to recursion), whether applied «voidable» stereotype and its correctness and validity of use, occurrence of abstract classes, review of geometry types (e.g., in some cases, defined geometry can be too general or too detailed) and the way of specifying enumeration types (CodeList and Enumeration; in some cases, giving only codes without any additional description can be ambiguous for users).

Regarding technical quality examination for the GML application schema, verification of the correct usage of types with ‘PropertyType’ and ‘Type’ suffix, encoding attributes with assigned «voidable» stereotype in the UML (in the GML ‘nilReason’ corresponding attribute), relations between classes (e.g., the lack of association roles in the UML entail the lack of information about relation in the GML code) and the manner of their encoding, as well as examination how enumerations and code lists are encoded in the XML Schema, should be mainly considered.

Other Aspects of Quality Examination

In addition to the already discussed aspects of evaluating application schemas quality, the considered methodology should also include some characteristics that can be simply measured or even calculated. An illustrative example is the entropy and complexity of application schema. It is worth investigating how complexity of application schema (e.g., the number of classes and how they are linked) influences its quality.

For this purpose, it is necessary to apply adequate complexity metrics. a good example of such a measure is a cyclomatic complexity, that is, a software metric (measurement) used in computer science to indicate the complexity of a program, developed by McCabe (1976). Mathematically, this approach is based on the graph theory and it can be easy adapt to verify the complexity of the UML application schema. The final results of these research activities will be presented in the following article about the complexity of UML and GML application schemas, focusing on a number of selected application schemas prepared in the Head Office of Geodesy and Cartography in Poland within the INSPIRE Directive implementation works, as well as application schemas from INSPIRE data specifications.

Another important point that must be looked at is the correlation between the quality of application schema and the quality of spatial data sets and services. This raises a question as to whether the faulty application schema results in defective data; in other words, whether the error propagation occurs.