Attending the Tuesday lecture series of the ACS CINF division was a very educational experience, that provided me with some key insights into the present state of scientific collaboration. I observed that almost all people who presented were focused on different techniques that improved the way scientists could get information they need from sources that were not human, i.e., publications, databases, etc.
Almost no one was focused on getting information directly from other humans using direct human to human communication that could be allowed by such technologies as instant messaging and VoIP, among others, that could be incorporated into a social network site. I find this a bit ironic if you take for granted the perception that it is generally the "older crowd" who may see all these new forms of human communication and interaction, such as instant messaging, blogging, or even engaging each other as avatars in SL, as "cold" as in perhaps a bit less than human. Yet at the same time, these same people would rather skip all these forms of human interaction on the web, and go with computerized data mining techniques instead of just talking to someone out there who might know what they want to know on a wiki or a blog, etc.
This suggests that there is a perception in the scientific community that the only way to "validly" communicate information is by making it available in a publication, and that getting the information directly from the source, as in the authors involved in that publication, seems generally out of the question. One of the reasons why this could be is due to the fact that the perceived worth of a scientist is determined directly from the number of publications he or she has had, how many citations they have gotten from each, etc. This perceived worth oddly enough has become hardly even a subjective statement at all, as this worth can be calculated using various methods resulting in hard concrete "objective" numbers. Using such bogus means to calculate an individual scientist's worth based solely off of data obtained from his or her publication statistics causes scientists to "keep their mouths sealed" until whatever it is they have been researching gets published in order that they might increase their value by a fractional amount.
Thus, there is a huge effort being done to mine that already published data, when it may in fact be more efficient just to "consult your neighbor." To use a (paraphrased) quote from the character Ripley played by Sigourney Weaver in the film "Aliens", "I don't know which species is worse, us or them. At least we don't see them screwing each other over for a percentage."
I think there is a standing question that needs to be thoroughly researched, if it has not been already, that can be paraphrased as the following: is it faster to get the information you need by scrutinizing publications, or by asking someone who knows? Of course this question is a lot more complicated, full of its own nuances and subtleties, than the paraphrasing readily makes obvious. One such subtlety is the difference between being presented with the key information and actually understanding that key information. There is also the great task of even knowing that certain information being presented to you is the key information.
It is probably not a black and white matter that data mining is always more efficient that asking someone else or vice versa, but due to the complexities involved in the human cognitive process of understanding information presented on different mediums, in some cases asking someone would be better, and searching the web of publications using various methods would be better in others.
One reason that asking someone else who knows may not be as efficient as searching the available publications, is because at present the lack of flow involved in contacting and engaging in conversation with one of the key human beings somewhere in the world who has the information and understanding needed using a computer is just too much for people to deal with. However the increasing popularity of interaction between online will help resolve this problem by increasing the chances of making contact with one of those key persons instantly online, since more and more people will be available online.
There are still many obstacles to be overcome but, in time, I feel that the right combination of online human-to-human interaction with data mining techniques will be the best way for scientists to get the information and understanding they need that will continue the accelerating trend of mankind's growth in knowledge and technology that has been observed since our speciation