Derivative "Collaboratories"

Many publicly accessible collaboratories are K-12 oriented web-based information clearinghouses that present some sort of interaction between students and remote scientists, educators, or datasets. For example, the CoVis Collaboratory, which is funded in part by the National Science Foundation, is

the website of the Learning Through Collaborative Visualization Project at Northwestern University, comprising…a community of thousands of students, over one hundred teachers, and dozens of researchers, all working together to find new ways to think about and practice science in the classroom. (http://www.covis.nwu.edu/   )

Many collaboratories, such as the University of Michigan's Learning Collaboratory at http://databases.si.umich.edu/LearningCollaboratory/   are intercollegiate university computer labs that can connect to each other using point-to-point or multi-point audio or video transmissions, but do not host access to remote instrumentation. Some collaboratories, such as the nonprofit Research Collaboratory for Structural Bioinformatics (RCSB) at http://rcsb.nist.gov/   are shared data repositories but do not incorporate human-to-human communication functions or provide access to remote instrumentation.

Some collaboratories, such as the University of Michigan's Collaboratory for Research on Electronic Work (CREW) at http://crew.umich.edu/index.htm , and the MIS Collaboratory at the University of Texas at http://cism.bus.utexas.edu/collab.html , are metacollaboratories, are web sites for university-based centers of research about collaboratories or collaboratory work. These sites offer extensive data archives and digital library functions, but (in the case of CREW) no remote communication function, and in the case of both, no access to remote instrumentation.

Undoubtedly, some government agencies, such as NASA and branches of the U.S. military, have fully operational collaboratories that facilitate manipulation of remote instrumentation such as launch rockets, satellites, and space station instruments. But, these setups are "hidden" from public access and, for security purposes are more likely to co-locate participants in command centers than they are to co-locate instrumentation virtually. Likewise, there are undoubtedly private, proprietary implementations of the collaboratory environment, as in large corporations such as at IBM (1997) and Bellcore (Cooper 1993). Such sites are not pursued by this study.

The sites that are pursued in this study are publicly-accessible and identified and located either from references in the collaboratory literature or by searching the World Wide Web. The following descriptive studies were developed during prolonged immersion at each site. Prolonged immersion involved a complete exploration and evaluation of the information environments for content, context, design, and delivery, and for other aggregated things, conditions, and influences that constitute an information environment.

SPARC : The Space Physics and Aeronomy Research Collaboratory

The Upper Atmospheric Research Collaboratory (UARC) at the University of Michigan was undergoing major revision during the term of this research, and has been renamed the Space Physics and Aeronomy Research collaboratory (SPARC). SPARC is funded, in part, by the National Science Foundation under the Knowledge and Distributed Intelligence (KDI) initiative as part of NSF's Upper Atmospheric Facilities. SPARC maintains a primary public web interface on the University of Michigan's Collaboratory for Research in Electronic Work (CREW) server at http://www.crew.umich.edu/UARC/ . Buried deep in the SPARC site is a form for requesting guest access or a logon account for access to the password-protected "working" portion of the collaboratory. This access point is at http://sparc-1.si.umich.edu/sparc/central/edit/Login/REGISTER . SPARC also maintains a separate K-12 education-oriented web site with games, experiments, and other instructional material at http://www.windows.umich.edu/sparc/ . The education site uses Java technology and requires the most recent version of any cookie-enabled graphical web browser software for full functionality. There are no direct links between these many sites.

(The NSF's Upper Atmospheric Facilities include four large incoherent-scatter radar and optical facilities located along a longitudinal chain from Greenland to Peru. Along with the Sondrenstrom Facility which is accessed via SPARC, they include Millstone Hill Radar near Boston, Massachusetts, Arecibo Observatory in Puerto Rico, and Jicamarca Radio Observatory in Peru ( http://www.geo.nsf.gov/atm/uaf.htm#desc ).

Upper atmospheric space physics is concerned with the effects of the sun on activities in the Earth's upper atmosphere. SPARC's purpose is to design, develop, deploy, and evaluate Internet-based technology to help space scientists work together in collaborative studies of space and upper atmospheric science. The SPARC team includes an international community of space, computer, and behavioral scientists
( http://www.windows.umich.edu/sparc/   ).

The primary instruments to which SPARC provides access are located at the Sondrestrom Facility, part of a global network of incoherent scatter radars. The Sondrestrom Facility is located north of the Arctic Circle in western Greenland. Figure 21 is an aerial photo of the Sondrenstrom Facility that shows the transmitter and instrument building, power plant, and crew housing. SRI International, one of the world's largest not-for-profit independent research enterprises ( http://explorer.csc.com/C3TEI/sri.html ), operates Sondrestrom for a variety of universities and government labs.

Figure 21. The Sondrestrom Facility

The Sondrestrom Facility's incoherent scatter radar has a fully steerable 32-meter parabolic antenna to which a wide range of optical and radiowave instruments are attached. Prior to UARC/SPARC, scientists had to travel to the site to use the instruments ( http://128.18.44.75/iono/issfsond.html) .

While SPARC provides access to various archived and live datafeeds from the various Sondrestrom instruments, manipulation of the remote instrumentation is available only through specially configured workstations at five locations. Real-time manipulation of the instruments is not available via public Internet, and SPARC gatekeepers did not respond to the researcher's request to remotely access instrument controls.

As with the M2C Collaboratory's TPM web interface, the SPARC web interface is a sea of frames, a technique for dividing a web page into multiple, scrollable sections. The SPARC's main web page (Figure 22) includes nine frames, each with different options, and each less than an inch deep, making it necessary to scroll through the individual frames a line at a time, or open each frames in a new window. Even though the SPARC site does not transmit video, it does have a RAM-hogging refresh default, and quickly maximizes the desktop. Like the M2C Collaboratory's TPM web design, the site is awash in a cacophony of acronym-laden choices that may be difficult for users without a background in space physics to understand. Unlike the M2C interface, SPARC offers a chat window and a list of others who are online, providing a real sense of co-presence that is absent from the M2C. A SPARC session feels more like reading a book than it does like seeing a video, as with the M2C experience.

Figure 22. SPARC Website

DOE2000

The U.S. Department of Energy, though its DOE 2000 initiative, and its prior initiative, the Distributed Computing Experimental Environment, or DCEE, takes the lead in implementation of the collaboratory environment. Besides DOE's M2C TPM Collaboratory examined in the previous chapter, and NSF's SPARC as described above, this research found only two other functioning Collaboratories in the public domain that meet all the CIRAL criteria for inclusion, and both are funded by DOE.

(See http://www-itg.lbl.gov/DCEE/Overview.fm.html   for DCEE Program overview and final report.)

Successful collaboratory implementation depends on basic and applied research, much of which cannot be easily identified for reasons discussed in Phase One. Any successful implementation rests on the shoulders of this often unrecognized research. Such research and development areas include telecommunications, computer science, multimedia transfer protocols, and hardware and software development, and comes from virtually every quarter of the government as well as from industry and commerce. And, just as there has been no direct tracking of collaboratory-specific research, there has been no direct tracking of federal expenditure on research and development leading to implementation of the collaboratory. Nevertheless, for bringing the research together and establishing a functioning collaboratory environment that is open to the public realm, DOE owns the prize.

(For detailed historical tables of federal funding for research and development 1959-1999, see http://www.nsf.gov/sbe/srs/nsf99347/htmstart.htm )

Besides the M2C Collaboratory, two other DOE collaboratory environments have been successfully implemented since 1997. The first is the Remote Experimental Environment (REE) Collaboratory built around the DIII-D Tokamak fusion reactor at General Atomics Corporation in San Diego, California. The second is the Environmental Molecular Science Collaboratory (EMSL) at the Pacific Northwest National Laboratory (PNNL) in Washington State. In addition to the technical and architectural research they provide, the REE project and EMSL each also includes a sociological research component. The REE social research was contracted to Sara Bly, of Sara Bly Consulting, but results of these studies are intended for internal use and are not published in the scholarly literature. They are, however, available online. This chapter investigates the REE DIII-D Tokamak Collaboratory and the Bly Reports, as well as the ESML Collaboratory, and its sociological and psycho-social studies, and concludes with a general synthesis of the environment of the functioning collaboratory.

REE Collaboratory Testbed at the DIII-D Tokamak

The Remote Experimental Environment (REE) Collaboratory is a DOE-funded testbed built around the use of the DIII-D Tokamak at General Atomics Corporation. A tokamak is a

toroidal plasma confinement device invented in the 1950s by the Russians Tamm and Sakharov …The word "tokamak" is a contraction of the Russian words: "toroidalnaya", "kamera", and "magnitnaya", meaning "toroidal chamber-magnetic.
( http://fusioned.gat.com/webstuff/Tokamak.html )

More simply, the DIII-D Tokamak is a very large (about three quarters the size of a football field) donut-shaped (toroidal) chamber in which plasma, the fourth most common and the most abundant state of matter on Earth (along with solid, liquid, and gas) is magnetically suspended at very high temperatures, and into which hydrogen atoms (from water) are introduced, where they fuse, creating energy. Figure 23 shows the inside of the DIII-D Tokamak at General Atomics.

Figure 23. Inside the DIII-D Tokamak

The DIII-D Tokamak is one of serveral tokamaks in operation globally. Hydrogen atoms are injected into the high temperature magnetically-suspended plasma inside the tokamak, where they fuse, or come together, in a controlled thermonuclear reaction. The reaction creates helium energy the same way the sun and stars create energy. Unlike traditional fission nuclear reactors, which split apart heavy atoms such as uranium, and produce troublesome waste and other side effects, tokamak fusion reactors bring two light hydrogen atoms together. Unlike fission, the physics of fusion make it inherently safer: a fusion reactor cannot go through a meltdown. The waste generated by fusion is expected to be less radioactive and to have a much shorter half-life, and is thus easier to dispose of than fission waste. Fusion energy is also much cheaper to produce: 50 cups of sea water contain the same amount of energy as two tons of coal, or a thimble of the element created by the fusion process is equivalent to 20 tons of coal. If advancements continue at the present rate, it is expected that energy break-even with fusion reaction could be accomplished by the year 2010. Commercial power plants could then come on line just as the Earth's oil gauge becomes critically low, about the years 2050-2060. (The preceeding description is synthesized from an introductory slideshow available at http://fusioned.gat.com/Teachers/SlideShow/SlideMenu.html )

The DIII-D Tokamak at General Atomics is one of several major fusion programs worldwide. Collaborating organizations include Princeton Plasma Physics Laboratory (PPPL), Lawrence Livermore National Laboratory (LLNL), and Oak Ridge National Laboratory (ORNL), some of the very same labs involved in the M2C TelePresence Microscopy Collaboratory discussed in the previous chapter. As an international resource for plasma physics and fusion energy science research, scientists and researchers from around the world are also involved in REE Collaboratory sessions.

The complexity and cost of firing up a tokamak fusion reactor require a high degree of international collaboration. The energy released by the fusion reaction is abundant, and harvesting multiple very large datasets from multiple, simultaneous, sequential experiments is possible, and usual. There are a variety of instruments attached to the tokamak, and each is used for a different type of experiment, which take place simultaneously at various stations around the tokamak. Instruments can be added to, and removed from the tokamak as needed.

Unlike the TelePresence Microscopy Collaboratory, which involves two or more remotely-located scientists working together at a single microscope with a single sample, DIII-D Tokamak Collaboratory sessions involve hundreds of scientists and researchers. These scientists and researchers are either stationed at various instruments around the tokamak, in other buildings nearby, in labs elsewhere in the United States, or are logged on from positions around the globe. Scientist may be conducting a shared experiment, or may be working on individual experiments, all at the same time, but everyone must nevertheless, and to varying degrees, stay in touch with each other and with the local or remote control team, which coordinates the overall "firing" of the fusion reactor. Unlike the M2C Collaboratory, the tokamak equipment is not accessible to the public via the Internet. Rather, it is wired to an internal network to which password-protected remote logon must be arranged. Unlike the M2C Collaboratory, the REE DIII-D Tokamak Collaboratory is not always online and available; collaboratory sessions are orchestrated events.

Coordinating communication between and among the hundreds of researchers, scientists, and technicians who are present at or logged on to any given tokamak sessions was the prime concern of the first REE Collaboratory experiments in 1997. Rapid adjustments of instrument settings between the rapid series of shots that make up a tokamak session are often necessary. These fine-tuning adjustments rely on input from any of the people involved (Bly 1997). Communication is achieved via audio/video transmission from the DIII-D control room, and also includes the broadcast of meetings, use of inter-process communication software to post events to the network during a tokamak shot, the creation of a DCE (Distributed Computing Environment) cell for creating a common collaboratory environment, and distributed use of computer cycles, remote data access, and remote display of results (McHarg et. al. 1997).

Not everyone involved in a tokamak session needs to stay in touch with everyone else all the time, however, so communication channels are customized. Any given scientist might need to be communicating on more than one channel, to more than one subgroup of operators or scientists, at any one time during the experiment. Communication involves a very complicated, cacophonous, and continuous connection. Simultaneous communication between and operation of the instrument by as many as fifty scientists around the world is described by one participant as being very much like playing a musical machine:

…you're playing it. And you're there all the time. You're doing a dynamic experiment. You're continuously changing the experiment. You're watching the data as it comes out. You're adjusting things, you're measuring things, you're changing the experiment as you go along. (Bly 1997, 3)

In the preliminary report on the ways collaboratory technology was used, and the effects of those technologies on the work of scientists and technologists during the 1997 experiments, Bly et. al. (1997) find that

Cultural evolution in creating collaboratories may be as critical as computing technology development. The introduction of collaboratory technologies will influence the design and use of the technologies in ways that are different from many groupware situations today. Understanding the work activity as well as the technology is critical to the long-term success of these projects. (1)

At the beginning of her study, Bly (1997a) identified six aspects of scientific work common across the laboratories involved in the DCEE initiative:

1. Expensive and hard-to-duplicate equipment for data collection
2. Complex planning and coordination of the experimental design
3. Multiple person, multiple specialties needed for carrying out experiments
4. Rapidly iterating and changing experimental parameters
5. Collaborators who are geographically distributed
6. Shared analysis and results.

At the conclusion of her study, Bly (1997b) identifies seven characteristics of activity that took place during the 1997 DIII-D Tokamak collaboratory session:

1. Nothing is fixed.
2. Everything is interdependent.
3. There are multiple paths to and from multiple sources.
4. Activity is constant and busy, but not frantic.
5. There is strong dependence on recognized, habitual signals and activities.
6. The frustration level varies.
7. The pre-operations meetings are valuable and relatively straight forward.

The REE DIII-D Tokamak Collaboratory experiment was much more a communication- and media-oriented special event than are either the M2C or the SPARC projects, which maintain ever-ready and permanent presence on the Internet. Each REE Collaboratory session is a one-time event involving hundreds of people, and requires special configurations that rely on the coordinated firing of a very large instrument.

While collaboratory session in M2C and SPARC are also events, they occur within a stream. The M2C instruments are pretty much always available and require minimal preparation and coordination for special use, and the SPARC instruments are always online, supplying a steady stream of data in a one-to-many sort of mode. To continue the musical metaphor, M2C is like a CD player that can be turned on to listen to a specific piece of music. SPARC is like music in the background of a public elevator: it is always on unless something has gone wrong, and you don't get to change the channel. REE is like assembling an international orchestra for a command performance at the White House. Thus, we see, the type of instrument has much to do with the configuration of any given collaboratory, and so must be considered a major influence on the other factors of a collaboratory ecology.

EMSL : The Environmental Molecular Science Collaboratory

The Environmental Molecular Science Collaboratory (EMSL) at http://www.emsl.pnl.gov:2080/docs/collab/ is concerned with the Department of Energy's mission to develop new technologies to clean up the nation's hazardous waste sites. EMSL is specifically but not exclusively concerned with the Hanford Nuclear Reactor site in southeastern Washington State. The Hanford Site has approximately 1.4 cubic kilometers of hazardous and radioactive wastes, 150 square miles of contaminated aquifer, and 60 millions gallons of radioactive wastes (260 MCi) stored in underground storage tanks (of which more than one-third are believed to be leaking). The Hanford Site also has 270 tons of spent fuel, 9 inactive reactors, and 7 major inactive reprocessing plants. The site is the equivalent of nearly 1400 Superfund sites divided into 78 distinct groups sharing common traits and geographies (Kouzes, n.d.).

The William R. Wiley Environmental Molecular Sciences Laboratory hosts EMSL at Pacific Northwest National Laboratory (PNNL) in Richland, Washington. The EMSL "Virtual Scientific Facility" is operated by PNNL for the DOE Office of Biological and Environmental Research (OBER). The EMSL's philosophy of the synergy of the collaboratory is represented by its logo (Figure 24), which is Borromean Rings: three symmetric interwoven loops that cannot be pulled apart although no two rings are interlinked, and removing (breaking) one allows the other two to slide apart.

Figure 24. EMSL Logo: Borromean Rings

EMSL houses many unique facilities for basic scientific research, including the world's first commercial gigahertz Nuclear Magnetic Resonance (NMR) spectrometer, a scanning near field optical microscope, and the most powerful IBM parallel supercomputer yet built. Overall, EMSL is home to some 300 researchers with unique expertise, equipment, and software.

As part of its work, EMSL is developing software tools to support collaboratory operations, specifically CORE, the Collaborative Research Environment, a suite of software tools that provide general collaboration capabilities ranging from distributed file systems to videoconferencing. CORE supports real time computer display sharing, a Web based Electronic Notebook, secure direct acquisition control of the EMSL spectrometers, and the ability to consult with EMSL staff for training, as well as during setup to optimize the experiment, or at any time during analysis or other activities.

EMSL does not provide the public direct web access to any of its instruments, nor to active, ongoing collaboratory sessions. To participate in any of the EMSL activities, the CORE software must be downloaded and configured on the users' computer. CORE software may be downloaded from the EMSL main webpage at http://www.emsl.pnl.gov:2080/docs/collab/   The software was not downloaded for this study.

EMSL Collaboratory sessions are achieved using X-Windows technology, a cross-platform interface which allows remote scientists, regardless of the type of computer they're using, to telnet (or remote logon to the computer in the lab), open a graphical window on their computer, and participate fully in ongoing activities. Collaboratory sessions must be arranged ahead of time by submitting a proposal, negotiating the time for the experiment, acquiring passwords, and arranging for any necessary training. EMSL Collaboratory sessions include both short term, project specific work, and long term ongoing collaborations. An example of an ongoing EMSL Collaboratory project is the one between Kelly Keating at EMSL (who does nothing but collaboratory science) and a scientist at Lawrence Berkeley Lab. Their ongoing experiment involves investigating the breakdown of DNA after exposure to radioactive waste. (Keating 1999).

Payne and Myers (1996) provide technical specifications for the architecture of the CORE suite. How a typical EMSL Collaboratory session is set up between two remotely located scientists using the CORE suite is described in words and pictures at http://www.emsl.pnl.gov:2080/docs/collab/virtual/VNMRFScenario.html .

The CORE suite of software is designed to support four basic types of collaborations EMSL has identified:

Peer-to-peer collaborations between scientists from the same discipline who share the same vocabulary and focus, who need to share realtime manipulation of the instruments and have access to unanalyzed data, and who may be in competition;

Student-Mentor collaborations where there is unequal knowledge, the introduction of new vocabulary, the need for reference materials and lectures, and the need to observe student efforts;

Interdisciplinary collaborations between scientists from different fields, (which may be bi-directional student-mentor projects), in which there are shared concepts, goals, and samples, but limited common vocabulary, and where access to analyzed data and data summaries (rather than raw data) is needed, and communication about the meaning of results is necessary; and

Producer-Consumer collaborations where resident scientists perform experiments and interpret the data on behalf of others who prepare and send samples to the lab, and in which there is no competition, few shared concepts, and limited common vocabulary.
(http://www.emsl.pnl.gov:2080/docs/collab/
research/CollabSociology.html)

EMSL research maintains that these four types of collaborations are universal and apply to education, medicine, and business environments as well as to the hard sciences, and that when one type of collaboration is undertaken, it often spawns one of the other types (Payne and Myers 1996).

EMSL's sociological research further identifies the communication needs of each of these four types of collaboratory sessions. Figure 25 is an EMSL- produced slide that shows the different modes of communication required for the four types of collaboratory sessions. (http://www.emsl.pnl.gov:2080/docs/collab/presentations/talk3.96/pg11.gif ) . The darkest squares represent communication modes that are not necessary, the medium gray squares represent modes of communication that are most necessary, and the light gray squares represent modes of communication that are moderately necessary to support collaboratory functions.

Figure 25. Communication Modes of Collaboratory Types

The EMSL Collaboratory includes access to digital library services in the form of collected published and unpublished articles, including preprints, but does not support the peer-review process or allow annotation of papers. Digital archives include scientist and experiment Notebooks (password protected), software files, and system logs. EMSL also has an aggressive program of both public and formal education, and is developing interrelationships with several regional colleges and universities using its CORE technologies to improve undergraduate science education.

EMSL's Collaboratory for Undergraduate Research Education (CURE) connects regional colleges/universities with each other, and with the lab. Again, downloading and configuring the CORE proprietary software is necessary to participate in the program, as is training, and arranging for access to password-protected portions of the site. Figure 26 is a screenshot of a CURE web session during an introductory chemistry course at University of Washington. The CURE program is dedicated to the proposition that "learning science can be done by doing science: interacting with real data and professional scientists, working and learning collaboratively project oriented and interdisciplinary work" ( http://www.emsl.pnl.gov:2080/docs/collab/projects/CURE/index.html ).

EMSL's preliminary psycho-social research into the collaboratory is concerned with how the introduction of electronic communications might change and/or enable new types of collaborations, and how the synthetic environment of the collaboratory might be sustained. Findings of this psycho-social research are:

• A collaboratory is a social system…nothing less than the village square and campfire juxtaposed to the Information Age
• It relies on social mechanisms of cooperation in which collaborators must gain value from a relationship in order for it to be sustained
• It uses props and controls for social discourse
• Either people or work are the focal object
• It relies on precedence and ritual organized by spatial syntax and by ritualized norms of acting

• A sense of place is essential. (Wise n.d.)

Figure 26. EMSL CURE Session

Conclusion

Chapter Seven explores the Materials MicroCharacterization (M2C) Collaboratory's five TelePresence Microscopy (TPM) sites, and tests the usefulness of the CIRAL matrix of criteria for inclusion as a collaboratory developed in Chapter Six. Chapter Eight briefly explores "derivative collaboratories" that do not meet the CIRAL criteria for inclusion, and explores, through prolonged site visits, four additional Collaboratories that do meet the CIRAL criteria. Site visits to the Space Physics and Aeronomy Research Collaboratory (SPARC), the Remote Experimental Environment (REE) Collaboratory Experiment at the DIII-D Tokamak, and the Environmental Molecular Science Laboratory (EMSL) Collaboratory are described. Most of the thousands of online environments that use the word "collaboratory" fail to meet the criteria for inclusion because they do not provide access to and manipulation of remote instrumentation. There are undoubtedly other functioning collaboratories operated by business or the government which are outside the reach of this research because of privacy or security reasons.

Each of the four collaboratories explored in Phase Two is a unique information environment. The instrumentation to which each provides access is a major determining factor in the nature of those environments. Each of the four Collaboratories serves a different scientific purpose, and each deploys a different combinations of communication modes.

Comparison of Collaboratory Architectures

In preliminary sociological studies, EMSL identified four types of collaboratory: peer-to-peer, student-mentor, interdisciplinary, and producer-consumer, and the communication modes required for each. The M2C Collaboratory focuses on the simultaneous transmission of microscope and macroscope video views to and from solo or small groups of scientists working online with one or more of the five TPM sites. Participants have access to and can manipulate remote instrumentation, are supplied raw data in the form of microscope lens views, or lens view still shots captured in collaboratory notebooks. Digital archives, and limited digital library resources are available. There is no need for special software. M2C does not support audio, whiteboard, chat, or forum functions. M2C functions as each type of collaboratory, but does not provide the full range of communication functions described by ESML.

The SPARC Collaboratory is centered on the delivery of text and chart-type data that is constantly transmitted from remote instruments, and text-based communication among unlimited participants logged on via the Internet using publicly-available software. Participants have access to the data flow from but cannot control the remote instrumentation. Control of the instrumentation is limited to five specific sites that use specialized workstations. SPARC does not offer collaboratory notebook, whiteboard, audio, or shared work space functions, nor does it support lectures. SPARC functions as all but the Producer-Consumer type of collaboratory, but does not provide all the communication modes identified by EMSL.

The REE-DIII-D Tokamak Collaboratory is centered around the audio and video intercommunication among hundreds of scientists, researchers, and technicians who either have their hands directly on some aspect of a very large piece of equipment, or are logged on remotely to participate in experiments or control the tokamak. REE-DIII-D Tokamak Collaboratory sessions are events that require considerable preparation and coordination. Special software is required to participate, and all the communication modes identified by EMSL are supported. The REE-DIII-D Tokamak Collaboratory experiment functions as all four types of collaboratory.

The EMSL Collaboratory is focused on facilitating shared, project-oriented experiments in either one-time mode or on a continuing basis. Special software is required to participate. EMSL does not make its instruments available via public access to its website. EMSL functions as all types of collaboratory, and facilitates all the communication modes identified.

Comparison of Collaboratory Data Types

Table 8 summarizes the data types generated by each of the four collaboratories explored in this phase of the study. The primary data produced by the M2C Collaboratory are video taken from a microscope and macroscope cameras, with subsidiary data being the record of the video transmission from the labs, the archives of the site's distributed email list, and the content of researcher notebooks. The primary data produced by SPARC are multiple, continuous dataflows of chart and graph-type information, and the logs of the text-based chat sessions. The primary data produced by the REE DIII-D Tokamak Collaboratory experiment are multiple, large data sets created by multiple instruments attached to the tokamak, and the audio, video, and text records of the intercommunication between experiment participants both before and during the event. The primary data produced by EMSL are raw data from the instruments, and the contents of collaboratory notebooks. All the collaboratories also have data generated as session logs.

 Table 8. Comparison of Collaboratory Data Types