Some of the LoD Cloud shortcomings identified above can be resolved by providing a systematic and formal descrip- tion of the LoD Cloud. There is an apparent lack of an ontology which formalizes and systematically captures the information contained in LoD Cloud datasets. Such an on- tology would bring multiple benefits with respect to the use of the LoD Cloud by providing systematic descriptions of the domains captured by the datasets, schema level linking of the datasets, additional schema-level axioms, and hence also better reasoning capabilities. Typically, such an inte- gration would make use of an upper level ontology. [...] For applications, these ontologies have been integrated with domain specific ontologies (de Melo, Suchanek, and Pease 2008; Oberle et al. 2007) to provide advantages such as bet- ter knowledge discovery, reasoning, or consistency verification.

The LoD Cloud datasets lack schema level mappings and do not convey relationships between concepts of different datasets at the schema level. To exemplify, a feature in the Geonames schema can serve as a venue for an event, e.g. the current model identifies “Atlanta in Georgia was the venue of 1996 Olympics” at the instance level. This creates significant limitations with respect to the reasoning potential which knowledge on the schema level would provide.

...the user has to be familiar with multiple datasets, and has to express the precise relationships between concepts in the RDF triple pattern, which even in trivial scenarios implies browsing at least two to three datasets. In our previous work (Jain et al. 2009) we made progress towards alleviating this obstacle. [...] This is perfectly fine from a knowledge engineering perspective, but it makes the querying of the cloud difficult as it requires users to understand the various heterogeneous schemas.

The LoD Cloud is of very shallow expressivity as a knowledge base and thus hardly allows to make use of underlying formal semantics through reasoning. The LoD Cloud primarily consists of ground level RDF triples, and hence does not utilize rich expres- sive features provided by OWL or RDF Schema. [...] Although DBpedia references Geonames using the owl:sameAs property, from the perspective of querying this makes it difficult as it might confuse the user as to which is the best source to answer the query. This problem gets even more compounded when contradictory facts are reported for the same entity by different datasets. For example, DBpedia quotes the population of Barcelona as 1,615,908, whereas according to Geonames it is 1,581,595. One can argue this might be because of difference in the notion of the city of Barcelona. But that leads to another interesting ques- tion: Is the owl:sameAs property misused in the LoD Cloud? This issue is partly related to Lack of expressiv- ity since there is no mechanism to perform verification of facts. Additionally, the LoD methodology prohibits reifi- cation of statements, thus disallowing assignment of con- text to statements

The interlinking of these diverse datasets promises a “Web of Data” that will enable users to easily navigate be- tween these datasets in a manner analogous to how users currently navigate from one webpage to another in the “Web of Documents.” Moreover, the LoD Cloud can significantly benefit both the AI and Semantic Web communities by en- abling new classes of applications and enhancing existing tasks such as querying, reasoning, and knowledge discov- ery.