Representation of semantic data is a computationally expensive process with a lot of embedded metadata for building semantically contextual graphs. However, such representation also comes at a storage and processing cost. XML standard has always been a more complete representation option on basis of which other standards have been developed. However, the introduction of JSON-LD provides further options for flexibility. Unfortunately, flexibility of semantic data processing also comes at a cost from loss in fidelity. Representing JSON-LD maybe a plausible option. But, storing the raw form of RDF in XML compatible native form is always favorable. This loss in fidelity may arise during content negotiation and during conversion. But, RDF is quite a memory intensive representation format which requires a separate processing requirements. Even viewing RDF from property graph perspective may not be sufficient. And, utilizing triple stores and even quad stores have always been the best option even of today, while such options still provide issues with vendor lock-in at times. Although, RDF and semantic web have come along way, there is still a lot that can be done both in terms of standardization as well as better distributed semantic graph storage. Semantic integration is again a core aspect of Linked Data requirements which is another aspect that requires more standardization and advancement. JSON-LD appears to be a useful option for front-end client processing in a lightweight integration. It also has some fundamental limitations in comparison to RDF. A question arises as to why the W3C gave up on the idea of RDF/JSON standardization. However, this is a case of what is more important in the semantic web community and for an application context, whether the representation should be in computer readable or human readable form. Nonetheless, the core representation format of semantic web for storage, in most domain contexts, should really be maintained in the native form of RDF/XML and associated derivatives for obvious reasons.
