Brief overview of the paper and its main findings

The article introduces an architecture and implementation for gathering, transforming, validating and publishing data catalogues, specifically for governmental open data. The authors compare their implementation Piveau with CKAN and uData by comparing features and find that Piveau in most areas is as good or (slightly) better than the others.

Major and minor points

Although the abstract and conclusion mention that the source code is available on GitLab, only some of the code appears publicly available. A direct link to a file in GitLab on page 652 (page 5 of the pdf) is broken, presumably because that code is not public. It appears contradictory to the claim that the code is available, which I understand to include all code and supporting files to allow reuse. Neither the publicly accessible repositories nor the Piveau homepage provide instructions for running the system or links to existing installations. This makes it very difficult to verify that features exist and compare the system to other solutions for myself.

A minor point is that the authors seem to have missed LinkedPipes ETL in their overview of existing ETL and Linked Data tools. Of course it is impossible to provide a list of all ETL tools in related work, but LinkedPipes ETL is ETL aimed at Linked Data and was around at the time of writing.

Conflicts of interest

No conflict of interest.