Domain Expert User Development: the Smartgov Approach

End-user development aims to give users the tools they need to implement their own software. Domain expert-user development can thus be viewed as a special case of end-user development and domain experts as a special case of end users with the necessary knowledge of how software operates, along with the tasks it must perform, the business rules it must enforce, and the validation checks it must perform. In some cases domain experts do not use the software directly, even though it indirectly supports their work; an example is software developed by tax officers, the domain experts in tax-collecting government agencies, for use by taxpayers, the actual end users, to simplify the work of the agencies through minimization of errors and population of electronic data repositories.

Using a user-centered software engineering paradigm [8], domain-expert users work with software developers to create specifications for the software to be implemented by the developers. This process is usually iterative. Domain experts are typically questioned by developers, developers design a first prototype, the domain experts ask for changes, and the developers address them and return with an altered prototype. Since both user groups are usually involved in other assignments as well, this process is likely to take time. Impedance mismatch problems [7], or problems in the communication between the domain experts and the IT staff due to their different backgrounds, perspectives, and terminology result in additional delays during this phase of system development. An alternative would be for the organization to help domain experts create the software with minimal or even no involvement of IT personnel. This approach was pursued by the two-year SmartGov project, which was partially funded by the European Union and successfully completed in 2004.¹ In the framework of SmartGov, a knowledge-based platform was developed to assist public-sector employees with suitable domain expertise to generate online transaction services by simplifying service development, maintenance, and integration with installed IT systems.

According to the European Commission [2], “transaction services (such as e-forms), though perceived as the future of e-government, have not realized their full potential.” E-forms play a significant role in e-government, being viewed as the basis for realizing most of the 20 public services [1] all European Union member states must provide their citizens and businesses. SmartGov delivers an intelligent e-forms development and maintenance environment and associated framework for e-government services.

Domain experts play a central role in the development of e-government services. They know what data must be collected and what data must be presented, the laws and regulations that govern e-service, and, most important, the business rules (including validation checks) that pertain to individual data items or groups of data items. Unfortunately, in most cases domain experts remain passive mediators of knowledge instead of more active players in the development process.

The role played by domain experts can be enhanced through a development environment that assists them in offering their knowledge, as well as in taking a more active role in the process. Such an environment must provide the domain experts the necessary functionality to create e-services for ordinary citizens in a friendly, intuitive manner. For that, the user groups involved in e-service development and maintenance must be identified and their needs recorded.

Analysis carried out in 2002 by participating local authorities during the SmartGov project [4] identified four main categories of public-sector e-service stakeholders:

• Manager. The public-sector managers who take a strategic view of the provision of services and obtain high-level metrics (such as statistics and performance indexes) through e-forms;
• Expert. The public-sector domain experts who require e-forms to support their work (directly or indirectly) and have the necessary domain knowledge needed to develop e-forms; they usually collaborate with IT staff to create e-forms;
• IT personnel. The public-sector employees (such as IT staff, clerical workers, and, in some cases, external consultants) who set up and support the infrastructure (such as Web servers and databases) necessary for e-forms; and
• End user. End users (citizens and businesses or other public-sector employees) required to fill out e-forms and interact with e-services.

These stakeholders confront significant problems implementing e-services. Apart from organizational and cultural barriers [6], obstacles also stem from the following sources:

• Limited view of domain experts. Difficulty involving domain experts in the e-service life cycle is due to senior government managers viewing them as simple users and not providing incentives in the form of acknowledging their role and expertise;
• Lost information. Valuable information can be lost during the transfer of knowledge from domain experts to IT personnel when that knowledge is supposed to be translated into software specifications. Though necessary for building the software, they reflect a low level of abstraction and cannot be used to help end users, create documentation for the service, or even facilitate communication among domain experts, while the original form of the knowledge would serve all these purposes;
• Complexity of creating e-forms. Most important, the process of encapsulating domain expertise in these forms is difficult; and
• Inadequate access. User friendliness may be lacking for end users in the form of online help, domain-specific information, external references, examples, and support for multiple access devices.

Though the approach employed in SmartGov tackled all these issues, here we focus on domain experts and the functionality needed to transform the whole process of creating e-services. Following the SmartGov approach, development of a new e-service starts with the definition of the data the service requires, the rules for verifying user input, the back-end processing that must take place, and any possible output. Domain expert users play a central role in all these tasks.

Domain experts know what data must be collected for the execution and efficient operation of the service. This knowledge is derived from the laws and directives that underlie the service and govern its operation. In the SmartGov approach, this implicit knowledge can be made explicit in the form of documentation and data-validation rules. Domain experts are, however, usually reluctant to provide it, fearing they’ll lose its associated power. Requirement analysis in SmartGov identified incentives that if employed would likely encourage domain experts to participate and share their knowledge. Since the main reward for public servants is enhanced status and prestige that would advance their careers, any approach aiming to make domain knowledge explicit must support this aim. To this end, domain experts must be acknowledged as authors of their specialized knowledge, and the management of public administrations must be able to identify them among the population of public servants. Creating a centralized knowledge base helps do this while also allowing the association of validation rules with pertinent e-service objects; it also offers context-sensitive documentation for developers and end users, as well as a means to retrieve objects (such as input fields, verification rules, and examples) affected by changes in laws and directives. Creating a centralized knowledge base that permits identification of knowledge contributors is supported by a platform that provides the tools domain experts and developers need to implement e-services.

Meanwhile, the approach in [5] to end-user development, along with a number of other articles, sought to capture the benefits of that approach, as well as the broad range of issues that must be tackled for end-user development to succeed.

Back to Top

SmartGov Platform

The SmartGov platform offers functionality for managing knowledge and validation rules, creating objects, and designing and deploying forms and services. The central concept in the SmartGov platform is transaction service elements, or TSEs, which are effectively widgets that may be used and reused as e-service components. However, unlike user interface widgets, they contain metadata and domain knowledge, along with visual design. Metadata may be the object’s type, range of values, validation checks, multilingual labels, and online help; domain knowledge includes information about the relationship of the object to other elements, legislation information, documentation, and more.

The platform contains a toolbox of predefined TSEs (such as value-added tax, Social Security number, and generic integer) that can be used by domain experts as a basis for implementing e-services. Widgets or parts from existing services, along with associated knowledge and rules, can be reused in new services.

Domain experts can identify and copy parts of existing e-services for similar or even for the same purpose. Identification of such parts is not based so much on the similarity of visual appearance but on semantic analogy. The domain expert can change anything, ranging from visual design to associated knowledge and validation checks. Common validation check types were modeled and included as templates in the SmartGov platform.

This approach to e-service development (see Figure 1), along with an easy-to-use interface (see Figure 2), helps address the obstacles in traditional approaches. Figure 2 shows the TSE editor screen accessible after logging onto the SmartGov platform. In it, the platform displays the most recent tasks carried out on the platform, as well as the services the active user’s work group has undertaken. The user uses the main menu on the left to access basic functionality for editing services and service elements, as well as for knowledge management.

The effort required to create e-forms is reduced by exploiting the reusability inherent in the development of e-services. Once a domain expert creates a TSE, it can be used (instantiated) as many times as needed, with each instance inheriting all linked knowledge and validation checks. To enhance reusability, the user can create groups of TSEs, along with cross-validation checks, then copy them (through the instantiation paradigm) as an integral unit from one e-service to another. One such example is that of the personal details field group. Almost all e-services include a form for entering user details (such as name, surname, postal address, email address, telephone number, and fax number. These fields can also include checks for testing validity and knowledge units (such as help).

During the pilot application phase of SmartGov, we developed a first pilot service for the Greek Secretariat for Information Systems of the Ministry of Finance in Athens that displayed the tax-clearance results of a citizen’s tax-declaration form. A second pilot service was later developed for use by freelance workers and companies trading goods and services within the European Union to declare the amount of their cross-country trade. Although the two services cater to different user groups and functions, each uses an identical form, namely personal-details declaration. Using the TSE group editor option in the SmartGov platform, developers of the second service were able to instantiate a copy of the personal-details part of the first service, including all linked knowledge units and validation checks, reusing it without further modification in the second service.

Since most services within the same public administration share common parts in both their front- and back-ends, the notion of reusability helps minimize the development effort and promote uniformity across a variety of services.

One especially valuable aspect of this development process is that it actively involves and acknowledges domain experts as knowledge contributors. The SmartGov platform enables domain experts to directly enter their knowledge for use in e-services; the fact that their contribution is recorded and acknowledged (they are publicly identified as authors of the TSEs they created) helps them overcome their initial reluctance to provide that knowledge.

The implementation of the SmartGov platform in the framework of the SmartGov project was based on open standards and technologies. Figure 3 outlines the architecture of the SmartGov development and deployment environments, along with underlying technologies. These environments are based on Apache Tomcat, a Web container developed at the Apache Software Foundation, and various open technologies, including JSP, Java, and XML. The development environment includes an authorization and personalization component, management modules for all e-service ingredients, and an integrator component that compiles and deploys the service (service logic and default authorization). Each service in the deployment environment has its own authorization component that enables end users to logon. Deployed services communicate with back-end systems through the SmartGov information interchange gateway [4].

We evaluated the SmartGov platform at the sites of the two participating user partners—the General Secretariat for Information Systems at the Greek Ministry of Finance in Athens and the City of Edinburgh Council in Edinburgh, Scotland. In the first, the domain expert users were accountants who also had computer experience; in the second, the domain expert users were occupational therapists who had only limited computer experience. We categorized the evaluation of the platform along three main axes:

Technical acceptability. We used a usability engineering technique to assess the SmartGov platform against the user requirements we collected during the initial phases of the SmartGov project. We found through conformance to user requirements analysis that the overall total of user requirements met either fully or partially was 78%, with more than 80% of the compulsory requirements being met.

Usefulness. We were concerned with SmartGov utility and usability and how it might be perceived by public-employee users in their native environments. Taking a holistic view of the platform in-situ at the two pilot sites in Athens and Edinburgh, we detailed the use of the development environment by the Public Authority Staff to create the online pilot services. The evaluation of platform usefulness was generally positive, meeting the majority of both participating agencies’ success criteria. However, the sites also reflect a disparity reflected in the separate results, due to the fact that the service developers in the European Commission had limited IT experience, whereas in the General Secretary of Information Systems, the service developers had a high level of IT expertise. Overall, however, the learning curve for occupational therapists was greater than for accountants, both user groups were positive in their comments regarding the usefulness of the platform as a means for creating e-services and for sharing knowledge with both peers and end users.

Social/organizational acceptability. We focused on a cost-benefit analysis of the platform in order to estimate SmartGov added value. Carried out by consortium partner Archetypon, it suggested that the platform provides significant added value in terms of development time, maintenance effort, and running costs.

A detailed report of the of SmartGov trials and evaluation results can be found in [3].

Back to Top

Conclusion

The SmartGov platform implements an innovative approach to developing transactional e-services in public administration. Regarding the applicability of this approach elsewhere, it is most efficient in contexts where the e-service is based on complex principles (such as laws and directives) and where many services exist. The presence of a complex background increases considerably the effort needed to implement and maintain e-service if a traditional development approach is followed. Such environments are usually encountered in public administrations, which base their operation (and therefore their services) on laws and directives that are subject to frequent modification. In certain environments (such as e-business), the implementation of the transactional service may not be based to the same extent on domain expertise or on laws and directives, perhaps rendering traditional development approaches as efficient as the proposed SmartGov approach.

SmartGov’s most important contribution in the development procedure of e-services is the active involvement of domain expert users. Domain experts are a special case of users—those with deep knowledge of a certain domain. To this end, it is imperative they be activated during the development of an e-service. The SmartGov approach showed that, with appropriate incentives and tools, their role in the design and implementation of the service can be enhanced, and they can become in effect developers of their own services.

Back to Top

Back to Top

Back to Top

Back to Top

Back to top