Diavgeia Redefined is a Google Summer of Code (GSoC) project which aims to solve the existing problems of Diavgeia.

Solutions

Observe legality and good administration in an automated manner

Problem: Government institutions upload their decisions, as .pdf files that follow no specific document structure. Moreover, Diavgeia currently hosts over 25.2 million decisions and that makes it obvious that citizens have no way to observe legality or good administration in a automated manner. Diavgeia applies some unreliable OCR techniques, relying solely on the honesty of government institutions to upload decisions in a machine readable format (but that is not always the case).

Solution: Diavgeia Redefined implements a RDF Schema, which covers almost every possible decision type. That said, Diavgeia should stop supporting the upload of .pdf files and introduce a new decision-format, compliant to the RDF schema. In this project, we adopted the Notation3 (N3) syntax and please feel free to browse the RDF samples. Through these samples, you can explore and understand the RDF schema easily and realise that the .pdf decisions of Diavgeia can be equivalently expressed as .n3 files.

As a result, these decisions can be stored to a triplestore and thus citizens can examine government's transparency, by making interesting SPARQL queries.


Unifying decision with its metadata

Problem: In the current implementation of Diavgeia, each decision has metadata information that can be accessed through Diavgeia's OpenData API. The metadata information is filled by the government institutions and often contain important data (i.e. financial transactions). However, there is no guarantee of consistency between the decision's pdf text and its corresponding metadata. That means, a decision may contain 5 financial transactions in the pdf text, but the government institution may have filled only 2 financial transactions in the metadata. That means, that Diavgeia's OpenData API can not be used by the citizens in a reliable way to examine the transparency of the government.

Solution: By adopting the RDF, the concept of metadata does not exist anymore. Each .n3 decision has unified the decision body with its metadata.


Linking to greek legislation

Problem: The majority of decisions, follow the model of taking into consideration the law X - we decide. We want to make sure, that these laws exist and link each decision with the laws that has taken into consideration.

Solution: Diavgeia Redefined uses ΝΟΜΟΘΕΣΙ@. ΝΟΜΟΘΕΣΙ@ is a project which has expressed the greek legislation as RDF Data. Thus, government institutions can link their decisions to the greek legislation or to prior decisions of Diavgeia. Government institutions are provided with the option to not link their decision with other entities, as there are decisions which have not yet been expressed as rdf.


Saving Disk Space

Problem: The vast amount of pdf decisions, leads to disk space scalability issues.

Solution: By adopting the RDF, Diavgeia stores only small text files instead of the corresponding PDFs. Using some simple @prefix and @base techniques, as well as a bzip2 compression, a x100 disk space reduction was achieved.


Immutability of decisions over time

Problem: Diavgeia offers no guarantee that decisions will remain immutable over time. Government may modify or completely remove a decision from Diavgeia and thus citizens and government institutions that have not downloaded this specific decision on their personal computers, have no way to prove its existence. A mechanism to ensure decisions' integrity and immutability over time, will introduce new levels of transparency among citizens, government and government institutions.

Solution: This is achieved by commiting decisions as Proof of Burn transactions to the Bitcoin Network. For further details, please read the Blockchain section.