I had an interesting conversation with Arnaud Rayrole (@arayrole) from USEO about wikis, enterprise social networking tools and the future of collaboration platforms. We met while at the Solutions Intranet et Travail Collaboratif conference.
Me in full action at the conference
Our conversation revolved mostly around the opposition and complementarity between people-centered tools versus content-centered tools. Below are the most interesting tidbits I recollected and some observations.
His take is that companies (especially large companies) are looking for tools that will allow them to find experts about any subject domain more easily. We both agreed on the fact that a declarative model (I say what I'm an expert in in my profile) would not be sufficient. The main issue with a declarative model is that most people will not take the time and effort needed to keep the information about themselves up to date.
His key point was that people actually show what they're experts of through conversations with one another. Conversations come naturally between people, whether through email, instant messaging or comments on a wiki. It's from those interactions that expertise in a given domain can be identified and assessed.
What would then be needed is therefore a way to excerpt meaning from those conversations in order to identify who's an expert of what. A social layer could be added on top of this in order to sort the relevance of content based on whether it comes from a close member of one's network. A tool such as Stack Exchange (which powers the Stack Overflow website that recently received Series A funding) could actually fit that description pretty well. The tool makes a great job of turning a conversation into a coherent piece of content and helping identifying experts in a given domain. Maybe I should give Joel Spolsky a call to suggest a potential business model? (I actually tweeted him, not expecting much of an answer though...)
However switching from a declarative model (I opt-in to share specific information) to an implicit model (I have to opt-out of sharing information) can have its drawbacks too (see the so-called Google Buzz Debacle for an example). I'm not sure how enterprise will respond to new applications that broadcast implicit (Albert is an expert about Swiss cheese) and explicit (Robert purchased French cheese) information to an internal public. While from an organizational standpoint this could bring a lot of benefits to the company, it may be a tough sell for individuals.
That's what tools such as Xobni have started doing but from what I understand of the service it keeps implicit information related to a given user available only to that user (who am I interacting most with? when am I most likely to get answers to the emails I send?). It's a bit different from the solution Arnaud was envisioning, where information would be collected and used at all levels of the platform but still very valuable.
The question of access rights is also interesting to look at. For instance, should I be allowed to see that such or such person is an expert in a given area if her "expertise degree calculation" is based on data I am not allowed to see?
Talking with Jean-Patrice from Personall
As for where XWiki stands with regards to this, I actually think we're not that far from the ideal tool he depicts (obviously...). If you take a look at the XWiki Incubator you can see an experimental dashboard that @jvelo put together. It brings together the latest comments and annotations that users have been adding to pages and displays them front and center on the page. Along with the list of pages people have been editing, it brings a lively and relevant view of what people are up to on the wiki. Conversations are displayed prominently on the page and allow users to get a quick grasp of who's talking about what.
I guess a next step for Jérôme's experiment would be some kind of faceted search interface (similar to Google's new look or the Facebook dashboard maybe?) that would allow one to filter the activity based on specific criteria (such as modifications taking place in specific spaces / from users outside of one's network for instance).
I'll stop here before the article gets so long that nobody ever reads it in its entirety, but special thanks to @arayrole for the stimulating discussion!