Most parliamentary institutions already have several computer systems in place for document management, data storage, electronic voting, teleconferencing, the handing of parliamentary procedures (e-parliament systems), and many more. This plurality creates frequent obstacles in seamless data exchange among heterogeneous systems and the smooth integration of new ones with different technology stack.

This interoperability problem has been tackled by the Hellenic OCR Team, https://hellenicOCRteam.gr, a scientific crowdsourcing initiative for the processing and analysis of governance and parliamentary data. The Hellenic OCR Team, which recently partnered with Bússola Tech, https://bussola-tech.co/, is an open research platform that links experts from academia and the public and private sectors who jointly tackle administrative and operational issues, as well as strategic challenges faced by representative and governance institutions. 

Systemic interoperability is paramount for the digital future of parliamentary organizations as it promotes transparency, accountability, and, ultimately, trust between citizens and their parliaments. The Hellenic OCR Team demonstrated a mechanism to link different systems using so-called Enterprise Integration Patterns, “a technology-independent vocabulary and a visual notation to design and document integration solutions”, https://www.enterpriseintegrationpatterns.com/.

The use case included a binary system to demonstrate this approach. On the one hand, an authoring tool for legislative text, Legislation Editing Open Software (LEOS), https://joinup.ec.europa.eu/node/701276, was utilized. This was linked to a repository with a set of verified parliamentary control data and metadata from the Hellenic Parliament, https://www.hellenicparliament.gr/. The approach to interlink these two independent systems, an online processing tool and a database, made use of a set of system-specific patterns to achieve the desired interoperability between the two distinct systems. To do so, a set of customizable software agents has been created, one for each system. As both systems handle different types of data, direct communication is not possible. This is because LEOS works with Akoma Ntoso, http://www.akomantoso.org/, a XML-based standard for legislative documents, while the repository contains a different XML representation of parliamentary control questions. 

Each agent includes an adapter which integrates with the host system based on the logic of the adapter pattern. The adapter “translates” system-specific data structures into communication patterns following a common messaging protocol. A messaging mechanism is in place to process, queue, send and receive messages from and to other decentralized systems. The equivalent example from the physical world is that of an international conference (interoperability use case) where participants (systems) only speak their native languages (data). The only way to communicate (interoperability) is to use interpretation (interpreters = agents) from and to a reference language, e.g. English (communication protocol). 

The integration of the two systems, even only for demonstration purposes, proved successful. This not only highlighted the tool’s versatility, but also opened up new grounds in interoperability between, formerly unrelated, legal systems and data sources. Moreover, the bi-directional behavior of the editor, enhanced with supporting agents having the ability to consume heterogeneous data and produce canonicalized output, may contribute to the creation of homogenous linked datasets. The whole process, that is the proposed platform architecture and the analytic solution design, are presented with great detail into a recent open access scientific article by Sotiris Leventis, Fotis Fitsilis and Vasileios Anastasiou, https://doi.org/10.3390/bdcc5030045.   

Reference: Leventis, S., Fitsilis, F. and Anastasiou, V., 2021. Diversification of Legislation Editing Open Software (LEOS) Using Software Agents—Transforming Parliamentary Control of the Hellenic Parliament into Big Open Legal Data. Big Data and Cognitive Computing, 5(3), p.45.

Permalink: https://doi.org/10.3390/bdcc5030045