Semantic tags

From Master Projects
Jump to: navigation, search

About Semantic tags



The RDF triples make use of URIs which are both used as a way to name entities and get information about these entities. This later usage, which consists in de-referencing the URI to get a description about it implies having access to the Internet infrastructure to make sens of the URI. But this mechanism is not necessary to the use of the data. One may create triples by using a URI even if that one is not de-referencable and two data sources containing triples describing the same entity can already share information without having access to Internet.

However, semantic data are essentially aggregated on central web servers which need to be accessed through the Internet. For instance, geo-location services featuring about information touristic places would work around a server which fetch the position of the user, match the relevant data and send back the result. Apart from the need to have a connection to that central server, some users may see privacy issues in sending their information to a tier.

An other approach, applied to information sharing, is that of devices like the Poken USB key. Poken users can put their business card data on a small portable device and easily share that information with similar devices. No centralisation is needed, making thus poken a totally free sharing mechanism. It also enables users to better control the diffusion of their data.


The goal of this project is to extend the ideas of devices like Poken in order to create autonomous devices making use of Semantic Web standards. Such a device would essentially consist in a triple store extend with communication mechanisms and minimal reasoning tools. The (upcoming) SPARQL protocol will be used as a mean to exchange RDF triples over wireless ad-hoc connections established among devices, eventually through a meshed network. The entire application will be either implemented as a service for mobile OS powering phones and tablets or on specific devices used only for the purpose of sharing information. This can be seen as a semantic extension of RFID tags.

Use cases

  • Touristic information points

A tourist visiting a city currently as two electronic guides options (when available). The first one is a tourist guide with pre-loaded information that can be carried around. This service is read-only and may contain a limited amount of information coming from a single source. The second one is to use online services which can provide more contextual information. This later case impose having an Internet connection available while touring which, despite being increasingly popular, is still a constraint. The alternative using a semantic data tag will be to put in front of touristic places a semantic tag containing structured information about the thing displayed. In a context of a place with several places to describe, it will be possible to build a mesh network allowing tags to act as data relays. A tourist coming next to a tag could simply read it directly with a compatible (i.e. SPARQL/RDF/...) device and, if he wants to, store part of the information. Then, this user would be invited to share his impressions on the guest book contained in the tag.

  • Intelligent devices

The semantic tag are used as a device extending over devices with meta-data level capabilities. By talking to each other in SPARQL, a mobile and a beamer can negotiate on the features one can do and find a way to finally project an image on a wall. Only using direct communication. If needed, the beamer may also indicate the name of an other resource able to play music providing that, as some point, such an information came to its attention. The combination of mDNS+Description of end points vocabularies+IPv6 flexibility of addressing will allow any semantic-enabled device to find other semantic-enabled devices to talk with on the network it is currently connected to (this network being connected to Internet or not).

  • Enhanced word-of-mouth

As word of mouth consists in passing information from person to person, typically by talking, an extension of this process can be made using semantic tags. Instead of considering two persons talking to each other it is two tags that start chatting when they get in contact range. Once in range, they can exchange and share information. This communication mean can be considered as a way to spread information in a context where centralised knowledge-bases are difficult to use/access. The advantage of semantic web standards is to provide a well defined way to describe things and communicate these descriptions.