Tuesday morning, the InnovationWell discussions included several product demonstrations, including two from pharma companies who had done some of their own work.
Johnson & Johnson talked about a new electronic lab notebook (ELN) system they built in-house on top of a technical backbone they already owned. Probably the most interesting aspect of this was that the development time was six months: they used an existing platform and they had many experiences and discussions with other ELN systems to guide their hand. They've also transitioned all their researchers (in a specific group) to using ELN's instead of their paper systems.
Jeff Spitzner of Rescentris talked about their most recent release of their Collaborative Electronic Research Framework (CERF). It sounds like they have learned a lot of lessons from their interaction with CENSA and their clients. I liked hearing the idea that the organization of the data is separate from the review and approval process. (Said this way, it doesn't sound so interesting.) Rescentris were also fairly active in the discussions throughout the two days: they have clearly been thinking about what the technical end of a CERF should be.
Alex Heifetz of Delta L Training talked about their work to use Second Life for training, including providing some live examples they have from their island in SL. In at least one example, I saw an attractive interface to an expert system. One (claimed) benefit over traditional flash-style training animations is that one can get data back from SL about the interactions, even if a user stops in the middle of a planned activity. Another example showed the benefits of doing training together in an area where a group of people could collaborate on an activity.
Frank Guerino of TraverseIT showed how their KnowIT takes a different tack on the question of "data integration." He suggests that their system is a Yahoo-like portal for the enterprise, with modules related to 60+ typical operational aspects of business (project management, budgeting, etc). The data for those modules can come from legacy systems, or from direct data entry into the modules. The value for organizations is in having everything in one place with a common interface across all the business applications. With their overlaying models, it becomes possible to navigate across many relationships, whether it is people or the documents they work on or the data they've created.
One interesting point of this talk was the suggestion that applications create silos within because the relationships between the systems are not known, so the users of the two systems cannot see how the one affects another. The claim is that if the relationships are known and browsable, the silos can come down.
Schering-Plough talked about an ongoing wiki project for use by discovery scientists. The speaker also mentioned the Pfizer study I pointed to earlier, which appears to be going better than the SP implementation. One suggestion is that the cultural aspects of getting wikis to work are important, but it wasn't clear which ones were at play in these two examples. I had the thought that maybe there were simply not enough people operating as big contributors, or that there wasn't enough "process" around using the wiki, even if the wiki is considered a "subversive" activity. Maybe the image of subversion is enough to keep it off the radar screens?