Impact

Our proposition does not only stem from conceptual inquiries but is also grounded in the requirements of actual PA workplaces. Our work draws on 3 cases of PA organizations located in three different countries, Hungary, Poland, and Slovakia. Due to their recent endorsement by the European Union (EU), these countries are still struggling to meet the quality criteria of EU in several aspects of public service. This task becomes harder considering the PAs’ heterogeneity within the boundaries of a single country, let alone at a pan-European level. This heterogeneity is increased due to the complexity of activities of the PAs, the continuous change as well as their scope (regional, national, international). Amidst these conditions, we were called to support three PAs on three different areas:

1. LATA (Mestska cast Kosice – Sidlisko Tahanovce) pilot user has chosen the involvement of the public into the process of making local legal regulations by annotation of the city ward general binding regulation.
2. MEC (Ministry of Education and Culture): pilot process is the higher education portfolio alignment with world of labour needs.
3. UMC (Urząd Miasta Czestochowy) pilot process is the management of education institutions’ material resources.

All three cases concern processes in the PA environment addressed to citizens, businesses, or PA-internal groups. The steps of these processes are knowledge- and information - intensive as:

• They depend on a number of applicable laws, regulations, court decisions, etc.
• They typically process a huge amount of information and documents from different sources.
• They require continuous monitoring of the content of the employed knowledge resources as they frequently become subject of change. 
• The application and interpretation of law and regulation is often not trivial, ambiguous, such that it requires experience in the field but also constantly updated knowledge.
• Content items logically belong together (e.g. a regulation interprets a law in a more or less binding manner, while example cases illustrate the borders of applicability) and are only valid in a certain, complex application context (which region, which timing of events, which kind of citizen or decision,etc).

SAKE’s intervention in PAs has two main goals:

• The empowerment of the public administrator.
• The improvement of Decision Making.

Through our endeavour to have SAKE deployed and used by the project’s pilot partners, we elicited useful insights regarding the achievement of the above goals. SAKE unquestionably achieved the 1st goal of the empowerment of public administrator. SAKE system can change PAs working experience by enabling both routine and non-routine aspects of their work. SAKE’s choice of services i.e. AMS, CMS, GWS, in combination with their strong integration proved adequate to address and support the knowledge-intensive context of public administrators. Nevertheless, SAKE still needs improvement. SAKE is a prototype yet and cannot be compared with any commercial system. However, it is the underlying concepts and novelties that make SAKE appealing and interesting. Public administrators limited in many cases IT literacy, their mis-perception of Knowledge Management but primarily their reluctance to change have been major hindrances in SAKE’s intervention. All these factors can limit or even erase the potential of any IT intervention. Regardless of what the origins of these negative perceptions are, SAKE should try to rectify any problematic aspects of the software and communicate its manifested empowerments through clear and simple messages.

The 2nd goal of SAKE which was the improvement of decision making has only been achieved with a modest advantage. Several intrinsic characteristics of decision making are responsible for this result. It is well known that improvements in decision making become effective and obvious after a long time. This could not be captured in the time frame of SAKE application in the three PAs. There is another facet of this problem which apparently affected people’s perception and judgement. SAKE makes process execution and task delivery very transparent. Pending tasks for PAs are queuing in SAKE workflow and stay there until completion. Under normal working circumstances the PAs might have several requests on a daily basis. The feeling of an “endless” queue is to a certain extent demotivating making the PA feel less efficient in his/her work. It can become even more terrifying by thinking that the progress of the tasks is being monitored by managers. SAKE’s intervention is deep and requires a lot of time before it becomes beneficial to the PAs and their work. We therefore believe that SAKE’s facilitation was perceived as a burden causing frustration and mixed feeling for its role in improving decision making.
In addition to the above basic goals, SAKE contribute to the achievement of the following benefits:

• Improvement of work to higher quality.
• Simplification of work.
• Acceleration and facilitation of the creation of outputs with improved quality.
• Better transparency of process.
• Enable easier opinion making and using previously gained knowledge.
• Cooperation of many various departments.
• Make decisions much faster.
• Make process more strict and formalized.
• Improve process efficiency.
• Initiate processes on line.
• Execute process faster.
• Bring many opinions to an integrated platform.
• Joy from using new technologies.

Lessons learnt

Although the research agenda for SAKE was huge, there are some research challenges which remain open. We are mentioning three of them:

1. Ontology management.
2. Workflow management.
3. Change management.

Ontology management is a quite developed research area, especially in the scope of Semantic Web research. For example, the ontology management life cycle process is well defined and commonly accepted. However, some more efficient methods in the some phases of that cycle are missing. Especially in:

• The development of ontologies:
  • This is still a very difficult engineering task (knowledge acquisition bottleneck) and the quality of a ontology-based approach depends very much of the methodology used for the development.  
  • A possible research approach can be collaborative ontology development.

• The maintenance (evolution) of ontologies:
  • Ontologies are live artifacts and they change in time, either because there are some changes in the environment and they must adapt themselves to new conditions, or their internal structure must be changed due to, e.g. non-efficient usage of some parts of the ontology. Especially important and difficult is to discover a need for changing an ontology.
  • A  possible research approach is to use historical (usage data) to mine some anomalies, i.e. patterns.

Workflow management is a mature discipline. However it was always seen as a very rigid system, that tries to automate of the service execution. On the other side, there is a gap in:

• Involving people in the orchestrated services:
  • However it was always seen as the automation of the processes and in that context all research has been done by assuming that there will be available services which can be orchestrated in a WfM engine. This view is especially common for the web services community.
  • A possible research approach is to include people as first-class citizens, integrating human-oriented tasks.

• Ad-hoc changes in the workflow:
  • A WfM engine is able to load and execute a process instance, but not to change it on-the-fly, i.e. during the execution of that instances. 
  • A possible research approach is the new, logic-based definition of the workflows and using formal methods for perform ad-hoc changes.

Change management is an always actual topic since the changes are immanent part of our life and businesses. However, there are two topics which deserves special attention.

• Methods for discovering changes:
  • Nowadays systems are dealing with static changes which cannot be updated in real-time.
  • An approach is the extension of the current data mining methods for real-time adaptation of patterns.

• Methods for maintaining patterns for changes:
  • The methodologies for systematic management of patterns for change are missing.
  • The method from maintain databases/ontologies or more complex structure can be used as a starting point.