In this paper, we propose a hybrid approach to making LOD accessible to the community used to CSV files. In both of those ways, the publisher is the one fixating the schema of the CSV file with no possibility of customization on the consumer's side. Another way of publishing the data is as a set of normalized CSV files mimicking the structure of a normalized relational database. One of those ways is publishing the data as one big redundant CSV file, usable e.g. However, this will, quite naturally, again fail to satisfy all users of the data, because there are multiple options of publishing a dataset in CSV. It seems that a solution to this conflict is to publish data both as LOD and e.g. This in turn is a source of conflict with the more LOD focused community. Members of the wider IT expert community then try to persuade the public administrations not to publish data as LOD and to revert to a more traditional format such as JSON or CSV. Note that the RDF Schema vocabulary is used for building other vocabularies, not for describing the structure of the published LOD, which typically reuses a number of different vocabularies. A common complaint against RDF and SPARQL is that the users miss a data schema, an equivalent to XML Schema for XML, JSON Schema for JSON and CSV on the Web descriptor for CSV files. It is then difficult to explain its advantages, including better data discovery thanks to the linkable nature of the data. In this wider IT expert community, RDF is considered unnecessarily complex, hard to understand and impractical to process, especially when transferring open data in bulk. However, it is this wider IT expert community which is the one supposed to publish data produced in the public administration as open data and to build consumer-facing applications on top of the published data.
While nowadays, it is not hard to find an understanding of the advantages of publishing data as LOD in the computer science, biochemistry and academia communities, this is unfortunately not the case in the wider IT expert community. It is backed by a family of Web standards such RDF for data representation and SPARQL for data querying. In the last decade, linked open data (LOD) became the de-facto highest standard of publishing data on the Web, a.k.a 5-star open data. In addition, Simplod configurations can be stored in and shared via Solid pods. We propose a tool called Simplod focused on rapid, straightforward and customizable formulation of a SPARQL SELECT query intended for customizable bulk transformation of LOD published in SPARQL endpoints into CSV files.
In this paper, we propose an approach to making LOD easily accessible to the community used to CSV files. However, it is this wider IT expert community which is the one supposed to both publish data produced in the public administration as open data, and to build consumer-facing applications on top of the data. Unfortunately, in the wider IT expert community, RDF and SPARQL is considered unnecessarily complex, hard to understand and hard to process, especially when transferring open data in bulk. One of the advantages of LOD is better data discovery thanks to the linkable nature of the data. In the last decade, linked open data (LOD) became the de-facto highest standard of publishing data on the Web, a.k.a.
0 kommentar(er)
