In EMSL we noted a very quick pace of technology changes. There are interesting temporal issues with regards to development. How long to develop something before there is some replacement available elsewhere?
Middleware development issues: Industry might come up with new products/services. There are middleware initiatives because OSs have been developed with potential middleware services in mind. Maybe we should have a workshop regarding these issues. We should examine the middleware layer more closely. How middleware supports collaborative applications and what is required in the operating system. Have a workshop that comes at it from both angles. Alternatively, include people from industry in the workshop. How do you find out all this information? Is there a common place to go for technical information?
We should look ahead to an analysis of people's use of application sharing. Do characteristics of use differ based on domains? How long are messages. This analysis could help middleware. Do a detailed examination of logs-what have they used, what would they like to use? A broad user needs analysis could accompany this.
There is an interest in the history of collaboratories and collaboration. How do features become stabilized. This analysis might also look into archiving, use of archives, etc. What are the characteristics that determine when a collaboratory evolves from a collaboration?
Another issues has to do with whether people should choose off the shelf vs. customized software. Open source could better provide opportunity for customization (e.g cheaper). Open source appeals to Africa, where cost is a major barrier.
What is the difference between an Engineering Collaboratory and a Science Collaboratory? For example, in an engineering one, product development is different. There are minute-to-minute adjustments based on what your collaborators are doing. We could examine the different types of collaboration that go on. Collaboration in a production environment is different with respect to collaboration. Academic engineering might be different than in corporations/firms. However it is still a different type of research community to look at and they might differ from other scientists also.
Success might depend on the type (e.g. Shared Instrument, Distributed Research Center, etc.) There are also different measures of success for individual, collaboratory, and field levels. How might a domain be revolutionized? Identify what could be better in then future?
There is too much anecdotal research in this field. There are serious problems with seeing them as truths. Empirical work often "debunks" anecdotes. Perhaps we can retroactively build control groups by determining who was/wasn't involved. At least studying science means that there is a record of who did what.
Funding allocation: What are the criteria that determine investment in collaboration? What are the characteristics about the technology that influence money? For example, NIH is giving distributed center grants only. Funding agents determine criteria, but "we" could propose successful approaches and go to the funding agents and saying "ask 'this' about a collaboratory". Funding agents are determining what they think is important when they haven't thought about what it really takes to get people to work together. We hope that this project will define some of the strategies so that RFPs will use these strategies, What is mandated and what is left for people to figure out? What are the RFPs based on? Often funding agencies do not specify what should be done? (CFAR). But for NEES, people have already decided what it should look like and punish people that don't conform to this. In NSF and DOE, panelists decide everything. We should get information out to them on what successes are.
What are the incentives to participate and incentives NOT to participate? Certain problems REQUIRE collaboratories. Certain projects would not be able to be done without collaboration. This is repeated in every part of NSF.
How do we get collaboratories to communicate with one another? Could we establish a publication format or forum where people could talk about policy, new organizational forms, transfer issues. Workshops are not going to answer all questions. How do we convey to other people what needs to go on? Do we need another journal to serve this purpose (information exchange)? Possibly we could use some forum with a reputation system.
We could look at shared systems in education/online support systems for teaching. This arena provides a larger N, more places of contact. Could look at long term effects of human capitol development and its context of use.
We should look at more dual use collaboratory, which supports research and teaching.
Why doesn't software reuse happen: Need to customize? There is no funding emphasizing that.
How does both technology building and research happen at the same time?
back to the Final Summary