Some wonderful comments. Let me touch on a few things. EIES was a self contained garden because it had to be as most things were not easy to get otherwise. However, besides the Well, Planet by harry stevens, and some other systems were directly influenced by EIES and built by people who had used EIES directly even though some never admitted it. We did, for example, build for example a complete user information marketplace and i wanted to turn it on for real money but NJIT ruled we would have to pay state taxes and they did not wnat to get involved so we turned it on as grand monopoly game to sell any piece of information for play money where the provider was a person and he or she could set the price for the joke, poem, recommendation or what have you. However, anyone who paid the price had the right to attach an opinion on whether it was worth the money and the seller could not remove that from the original ad used for selling. It was great fun and still does not exist because of the lack of micro money. At the time everyone was assuming information providers were companies and not people. To some extent the investment business still has that blinder on. Collaboration still has a long way to go. The complete design does appear in the literature. It was a fore runner to ebay and recommender systems in a very real sense. Peter and Trudy Johnsen Lenz who built the "topics" system on EIES, using the language that was a virtual browser between our users and the back office system, wrote a wonderful article on the "rhythms of conferences" where they pointed out that participation is a series of peaks and valleys rather than a stable average. People shift around among different topics in different conferences and a critical mass at the right time triggers a major peak. Topics by the way supported 200 users exchanging information on scientific questions facing state legislatures and had a flow of over 200 comments a week where questions were limited to 5 lines and answers to one page and you could choose which questions you wanted to track answers to. When we turned it on everyone was asking questions and only professionals representing professional societies were answering. We created a membership list which tracked for all the users how many questions each user was asking and how many answers they had made and all of a sudden most people started to answer questions. The exchanges were archived in a database for future use if someone became interested in the questions and their answers later. Questions ranged form things like How do you define and antique? to things like Standards for burying power lines? There was usually about 15 such questions every week. Having this ability to tailor for any individual or group led to a lot of interesting things. We even had a fun feature where every user could tailor their own interface by replacing all the prompts. We had users tailor "English butler" "insults" etc. Not having the web we could meausre the time and what people did direclty and learned a lot of general things about user behavior summarized in part in a couple of CACM articles by Roxanne Hiltz including the one first major article on information overload in this technology. One of our most popular conferences, which you had to have permission to join, was one where everyone liked to insult each other. I do urge some of you to look at hte NLM report on my website on information overload. Ignore the body and just go read the 60 page appendix of the raw comments on the 34 professionals in Emergency Management on their problems on gathering and using information on the web. This is the most significant part of the report with respect to meaningful evidence. While they do not state it directly it becomes clear that they want a user controlled recommender systems even though they do not realize such a thing can be done. But clearly they as a "community" want to nominate the material to be included and to evaluate it collectively. There is alot of "communities" they have already built with volunteer newsletters and websites and about a dozen are described in our report. These are places where a lot of professionals are putting in a lot of effort for free to beneift one another using only simple tools like websites and editors. They list about 300 links they are using and the sites are active in the text of the appendix. There biggest problem is dealing with the "gray literature" which is what NLM calls non journals, but is what the practitioners want the most. An awful lot is happening among communities of practice and i think the word "communities" is still a good one. Distinguished Professor Emeritus Information Systems, NJIT homepage: http://is.njit.edu/turoff