Mopsos

Just had a meeting this morning with Jennifer Dahan-Templier from Droit-in-Situ.

Droit-in-Situ is an amazing little company who develops learning modules including audio, video and text in a very powerful and cost-effective way. The secret lies in a patented technology and process allowing sequences of audio/video files to be time-stamped and linked with other reference material. In substance each binom (file F, time-stamp T) becomes a URL. A very simple concept, and a brilliant idea, which allows the creation of learning modules like this one (take the time to browse through it, event if you don't understand French or if you don't like the graphics - look at the details, look at the way speech sequences are precisely linked with reference material - this is great). Droit-in-Situ applies this technology to process recorded seminars, events, lectures, and more generally any event where recognized experts and gurus talk to an audience. Jennifer believes that these are "moments of truth" where deep knowledge is actually being transferred. Thus this is where technology must be used effectively to disseminate this knowledge even further, so that in can be used in real-time, when needed.

During our conversation, we talked, among other things about tags. Jennifer was very cautious about introducing tags in her framework. She believes that tags are subjective ways of classifying information that introduce a strong bias in favor of the tagger's personal vision, whereas Droit-in-Situ consider that the only relevant classification they can "ethically" use to link content is what has been selected:

• by the expert herself (e.g. topics and reference material)


• by the organizers of the event (e.g. schedule of the event).

She does however sometimes introduce a new classification scheme of her own, but in all cases, it must be objective (e.g. dates when the seminar is about History for example)

This made me think about the use of tags. To be honest, my personal "tag cloud" on del.icio.us is a mess, and only I can -sometimes- make sense of it. Each time I stumble into something interesting on the Internet which I cannot process immediately, I bookmark it and tag it using my own keywords of the moment. After a few months, I cannot really remember whether "Schneider" means the link is from Schneider, about Schneider, relevant for my projects at Schneider or to be forwarded to Schneider people. But still, I can use those tags more or less, because they are meaningful to me anyway, even if I am not completely satisfied with my tagging practice.

This also made me think about group classification. What happens when an organized group of people, say a project team or a community, wants to jointly keep track of important learning content in an effective way that truly capitalizes on each member's knowledge? What they typically do is sit down and compare the tags and keywords each member of the group uses to characterize the common domain, from which they negotiate and derive some group taxonomy or ontology that every group member will be asked to use in the context of the group's activity. So far, the best simple tool I have found to do this was our former knexsis platform (ID=guest password=guest then click on the "search" tab) , which has evolved into a commercial product called Jalios. It's about creating a shared taxonomy on-the-fly, and allow it to evolve over time, as new concepts are added. On knexsis/jalios, any collected piece of information can be tagged using an official and organized list of tags/keywords, common to the group, and each group member can add new tags if she doesn't find her way in the official list. The bottom line is a trade-off between total freedom (subjective tags on a personal KM tool) which may be meaningful to me but not to others, and more constraints (nodes on a group taxonomy/ontology) which I might find more cumbersome but which allows for better sharing within a group. At CNES, the French space agency, expert communities design reference ontologies of their common knowledge domain which are used separately to index a single company-wide repertoire of knowledge (Arisem technology), thus foregoing the all to frequent illusion of designing one single "company-wide taxonomy".

So maybe the next step for Droit-in-Situ is to allow for external tools like del.icio.us or Jalios or Arisem not only to link to specific html pages of DIS content, but also to specific time-stamped sequences of expert speech embedded in those html pages.

I wonder how this can be done technically…

No-one has commented on this entry (yet).

Thanks for signing in, . Now you can comment. (sign out)

URL:
Remember me? YesNo

Comments: