Proceedings of the ESWC2015 Developers Workshop                            54




demo
Consuming RDF data the OOP-way
Tomasz Pluskiewicz, Karol Szczepański

Introduction
Object-oriented programming (OOP) paradigm is a prevalent method for designing computer
software. In OOP data structures are represented as classes, which declare a set of named properties.
This is especially important in statically typed languages, where this structure is fairly rigid and
usually does not change at runtime. That structure causes friction when data in other representations
needs to be accessed as objects. A well-known manifestation of this issue is the object-relational
impedance mismatch, which occurs when data from relational database management systems is used
in an OOP program. Developers face similar issues when using RDF data, which, even though
seemingly object-oriented, has a number of properties different from typical objects in OOP sense. I’d
like to show two possible ways of overcoming this problem that can be used in programs written for
the .NET framework.


ORM-like solution for RDF
In a past project we have created a full-featured ORM (object-RDF mapping) library called Romantic
Web as an object-oriented abstraction over RDF data. We have designed a library, which offers many
features commonly found in relational ORM solutions, including .NET’s Language Integrated Queries
(LINQ), handling inheritance and complex data structures. It uses Rob Vesse’s dotNetRDF library to
access actual data backend. However our design enforces a certain way for designing the RDF store
and also resulted in a fairly complex tool both in terms of usage and implementation.

More details are available at http://romanticweb.net


Using JSON-LD in place of an ORM
In some cases it may not be necessary to introduce a complex and limiting library to access RDF data.
Still, manipulating graphs at triple level is not ideal in an object-oriented program and will likely result
in unreadable code. A simple solution to that problem may be JSON-LD. Plain JSON has proved an
invaluable tool for transmitting and manipulating data in many languages on all platforms. By
leveraging JSON-LD @context, developers are empowered to create a simple bi-directional
mapping between objects and RDF graphs. With minimal amount of code I have created a simple tool
which uses a well-established JSON manipulation library Newtonsoft.Json to transform JSON-LD to
and from RDF. It does not offer any advanced features but can be a lightweight alternative to the
aforementioned ORM or manual work with RDF triples/quads.

Code and sample are available at https://github.com/wikibus/JsonLD.Entities




                                      Copyright held by the paper authors