Current issues with Social Network Representations

In the social network interoperability space, several organizations have made great strides in forging agreements on interaction protocols among applications and between applications and their host containers. Less attention has been paid to another critical aspect of application interoperability, i.e. the shared representations of social network data. The importance of a shared syntax and semantics can not be overestimated: in practice the mappings among (versions of) data formats and vocabularies are too difficult to infer automatically and result in a serious loss of information. Semantic technologies are likely to offer a way out, but do not provide an adequate solution at the current state of the art, largely because of a lack of coordination among related efforts. In this paper, we will illustrate some of the practical problems we have encountered when helping large publishers to annotate their content with microformats and existing RDF vocabularies (ontologies) in the context of SearchMonkey, Yahoo’s Semantic Web product. We will also discuss potential solutions. While some of these issues may seem trivial at first, a lack of agreement on their solution still means that often applications developed by our users cannot fully rely on the data they receive.