Table of Contents 
Previous Chapter
John M. Carroll & Chris Connolly
In this section we discuss recommendations for tools, which if created and deployed would enable greater collaboration among NSF PIs and among scientific and other communities in general. The World-Wide Web can facilitate asynchronous and synchronous collaboration among researchers in a variety of ways. As suggested in the Short-Term Recommendations section, it allows the creation of standard PI Home Pages with links from project abstracts posted by NSF to PI Home Pages, and it enables on-line submission, review and distribution of NSF project proposals and reports. It can support the development of a world wide distributed digital library of scientific and technical information, as discussed in the strategy section, including development of link databases to facilitate querying document links. It allows the use of reply-to and public annotation capabilities for technical information posted on the Web, as well as interactive collaboration tools (e.g., a World-Wide Web Interactive Talk (WIT; see
<URL:http://info.cern.ch/hypertext/WWW/WIT/User/Overview.html>) capability, and object-oriented multi-user discussion environments (MOOs)).
Several of our recommendations will require active involvement by NSF Web users. Hence, as also discussed in the section on short term recommendations, adequate infrastructure, training, and personnel support at NSF will be needed.
An important aspect of collaboration is knowing who has linked to posted information, or what other documents point to a posted document. Link databases associated with posted documents would allow the investigator who authored a document to notify those who have linked to his or her documents of revisions or follow-on work. They would allow researchers to access the set of documents that refer to the current document. Such an architecture would make documents on the Web part of a real network of knowledge. A version of this capability is already implemented in Hyper-G [Kappe 93; Kappe 94], and should be made available on the Web (e.g., [Pitkow 95]).
Encourage development and oversee use of reply-to and public annotation capabilities for technical information posted on the Web. Investigators should be able to communicate directly concerning posted information, with one another and with sponsoring NSF program managers. For example, it should be easy to reply to a currently viewed document, by generating an e-mail message to the PI or the program manager. It should be possible to publicly annotate posted documents (though this may need to be managed by NSF). Creating such direct links would encourage direct collaboration among investigators.
Many Web facilities are being developed specifically for collaboration, both synchronous and asynchronous. For example, there are a variety of debate forums, based for the most part on WIT. WIT uses the forms capability of HTML to support debates structured in ways similar to Rittel's classic IBIS (Issue-Based Information System [Rittel 73]). Users state contrasting positions on various issues and adduce evidence and arguments to make their cases. Two current limitations with this capability are (1) that in many cases its actual use has devolved into serving merely as a better interface for general newsgroups, and (2) that it only supports linking flat text objects. More focused WIT collaborations have been created by incorporate management schemes, for example in the Virginia Tech computer ethics debates
<URL:http://info.cs.vt.edu:8000/debates/html/Debates.html>. Research on approaches to moderating these interactions could increase their potential utility as tools for collaboration within the scientific community. Extending the WIT framework to support linking documents and bibliographies as backing for positions taken would also help make this a more suitable tool for scientific collaboration.
The most common tools for synchronous collaboration are multi-user discussions, such as BioMOO, a MOO for collaboration among biologists
<URL:http://bioinformatics.weizmann.ac.il:70/1s/biomoo>. BioMOO hosts informal discussions, presentations, and even journal clubs. All of the participants in BioMOO enter remotely, and at any one time are likely to be scattered throughout the globe. The result is a rapid exchange of ideas and results which heretofore would only occur at annual meetings or the occasional workshop. Currently, most MOOs are exclusively text-based and accessed via telnet. BioMOO is an example of a MOO that can exploit Web browsers, however the activities it supports that go beyond text-based interaction (for example, the 24-hour posters), are all asynchronous. An important avenue of exploration and development is the support of higher bandwidth synchronous collaboration.
A pioneering example is the USC Mercury Project, which allows users to remotely command an excavating robot
<URL:http://www.usc.edu/dept/raiders/>. In general, the Web can be used to allow researchers to remotely access exotic equipment at other sites. For example, researchers wishing access to range data could conceivably access sites which have available range sensors, which would then capture images and transfer them to the requesting researcher. As another example, remote access to anthropomorphic hands or other robotic devices could enhance the development of algorithms for dextrous manipulation. This remote access capability should be useful in any discipline which requires data collection using exotic or expensive hardware. An important direction for development is the support of greater interactivity; collaboration in the Mercury Project is implemented via fairly cumbersome turn-taking.
The actual use of Web-based collaboration tools is not well-understood: What are the tools used for now? How effective are they? What are the routine problems that users encounter? Studies are needed of scientific collaborations such as those currently supported by BioMOO and the USC Mercury Project. These collaborative communities are on a different scale than typical computer-support cooperative work (CSCW) projects, and it is likely that their fundamental objectives and concerns are just different. Unless we focus systematic effort at understanding these communities, developments in support for scientific collaboration using the Web will be no better than evolutionary.
Kappe F., & Maurer, H. (1993) Hyper-G: A Large Universal Hypermedia System and some Spin-offs. Computer Graphics (experimental on-line issue).
Kappe F., Andrews K., Faschingbauer J., Gaisbauer M., Pichler M., & Schipflinger J. (1994) Hyper-G: A Network Tool for Distributed Hypermedia, (on-line documentation)
<URL:ftp://iicm.tu-graz.ac.at/pub/Hyper-G/doc/>.
Pitkow J., & Jones R. K. (1995) Towards an Intelligent Publishing Environment, Computer Networks and ISDN Systems, 27, North-Holland (in press).
Rittel H., & Webber M. (1973) Dilemmas in a General Theory of Planning. Policy Science, 4, pp.155-169.
Table of Contents 
Next Chapter