The presence of presence as part of unified communications deployments increasingly is being felt by enterprise users. It’s paying off in better productivity, but it’s when presence gets semantic that the real magic can start to happen.
That’s something that DERI and Cisco, as part of the multi-party Lion2: Enabling Networked Knowledge project funded by Science Foundation Ireland, are exploring. The idea is to “bring more granularity and more meaning to presence than state,” says Cisco lead architect Keith Griffin – that is, an indication that you are available or away. After all, just because you’re sitting at your desk doesn’t necessarily mean you are fully or even partially available, and just because you are on the phone doesn’t necessarily mean you aren’t free to respond to instant messages. “The Semantic Web provides more intelligence to that.”
How? The general idea is to apply Semantic Web and Linked Data technologies to the concepts and protocols – SIP and XMPP – that are most common for presence and instant messaging in unified communications environments. It would work by adding semantic extensions to presence standards and using them in conjunction with ontologies such as SIOC, FOAF, Dublin Core and Online Presence, to map to fields about who is sending what message and when to whom. Policy- and rules-based scenarios can help determine interaction procedures. So, a linked data system could put together enough information to say what call or Webex session you’re on, who you are speaking with based on policy rules and other input, and then determine the presence state for you and present that to world.
“Once mapping is achieved as a first step we’re able to prototype further, to take data out that is associated with IM conversations into a graph-based form that gives you flexibility to take into multiple directions,” says Griffin.
Continues DERI’s researcher Alexandre Passant, “As soon as you do that, imagine all the mash-ups, geolocation and faceted browsing you can do. That’s not provided directly by XMPP but once we know the topic and the time stamp we can aggregate in different directions and link with other presence information.” As an example, it could become possible for someone to leave an IM meeting request at the next geographical location that a currently busy worker is expected to be at, for delivery upon his arrival – avoiding interrupting him in the middle of a critical task but also ensuring the message gets his attention as soon as the appropriate opportunity presents itself.
Given that the Semantic Web is still in an emerging state at most enterprises, their work focuses on capturing existing information, say, in LDAP directories, into ontologies like FOAF that can become part of the presence system. If someone had a profile in a given format, it could be linked to natively. Either way, the result can be “some rich information in graphics that can be used by the applications in the organization, so the user can see who they interacted with and if that person should be part of their social network and not just an ad hoc contact, for example,” Griffin says. Organizations that closely cooperate with each other could arrange a federated presence to enable the same wealth of interactions across corporate boundaries, and such a presence could even extend into business-to-consumer interaction.
Privacy considerations matter here, of course. “It’s important that if you model data in the enterprise we can still apply privacy policies in terms of when to share presence information for inside and outside presence,” says Passant. “That’s even easier when data is in RDF because you get more context – what project I am on, what group, and policy depending on that, make presence available only to people from my team but not to people from the other team, based on certain hours and certain other conditions.”
The project is about halfway through at this point, having accomplished defining its problem domain but also defining issues around ontology mappings, policy rules, and coming up with something like an IM analysis engine. Now it’s focusing on what results of this might be ripe for standardization. “If presence is going to be something that is ubiquitous, it makes sense if everyone takes a similar approach if not the same approach,” says Griffith.
Passant adds that the development of a full presence system based on Linked Data and semantic web technologies, including NLP and entity extraction for more fine-grained management of messaging, also can have payoff on the integration cost front as it relates to the entire bigger picture of collaboration and communications systems. New prototypes or clients can be accommodated just by mapping to them from existing system.
Says Griffin, “Presence is a hugely valuable feature but it is one of many features in a communications or collaboration system. So the adoption of semantic web or linked data technologies and techniques is about the bigger system – about the social semantic enterprise, about having social graphs and mapping the enterprise and creating social networks between people and organization,” he says. Ultimately, “prescence is additive. …It’s an example of one valuable feature of many that we will need to create an overall experience.”
- Automatic Hashtags & Machine Learning: The New Google+
- Gmail, Meet JSON-LD
- Cambridge Semantics Wins SIIA Software CODiE Award
- ADmantX Named "Cool Vendor" in E-Commerce by Gartner