Although the work of the VMDL Project has been focussed around the use of information and communication technology, the case-study research reported and synthesised in Chapters 3, 4 and 5 did not place emphasis on the technological, particularly software, artefacts that underpin the research. In this chapter, a more specific focus is given to communication technologies themselves: the tools and environments distributed collaborators can use to connect with each other and with mutually accessible resources. Interspersed throughout the chapter will brief case descriptions of research communities using examples of particular technical configurations. And unlike the decision taken for the case studies around which Chapters 3, 4 and 5 revolve, a common-denominator approach to technical tools will not be taken, but instead an overview of both existing and emerging technical possibilities will be presented.
Section 6.2 provides a base discussion of the technologies required for broadband communication, emphasizing synchronous (two-way) video communication. The section forms a valuable background to Section 6.3, which shows the application of such multimedia communication technologies in a variety of different networks of distributed researchers and other collaborators. Section 6.4 shifts attention to software tools for groupwork--groupware--through placing those in the context of various tools and technologies for both synchronous and asynchronous communication. While possibilities such as audio and video conferencing are included in this broader context, the emphasis in Section 6.4 is groupware via data-communication networks. Section 6.5 completes the chapter by looking particularly at an emerging form of network-based "groupware": the so-called "MOO" environment and its emerging integration with World Wide Web functionalities. Again, an example is given of a distributed research community meeting its collaborative needs via such environments.
This section is concerned with providing an overview of the technologies required for broadband communication between people. The most demanding requirement is for the sychronous transmission (and reception) of visual images such as those broadcast through television channels. However, facilities for television transmission are expensive and alternative ways, using technology based on advanced digital networks, offer the opportunity for synchronous communication at a cost which can be afforded by distributed research communities.
To understand the basics of multimedia-based communication technology*, one has to distinguish between three main parts:
1. the different audio-visual components, such as camera, screen, loudspeakers, etc;
2. the communication channels;
3. the devices and processes which allow the output of the first to go through the second in large quantities and in a fast and reliable way; such processes include the coding-decoding and the compressing-decompressing of the digital signals.
The physical devices for creating digitised signals use a technology which is relatively stable and satisfactory, and is improving and becoming cheaper all the time. Little needs to be said about these elements but, as detailed in Section 6.2.2, there is much development to be done in order that the signals from such devices can be processed for transmission, not just point-to-point but between multiple sites. However, the first area considered is that of the digital communication channels themselves.
A great deal of research and development has been undertaken to improve the capacity of channels to carry digitalised analogic data, such as images and sounds. There are 2 main communication channels:
a. normal telephone lines;
b. digital ISDN* lines.
Digitalisation essentially means that there has to be a system, a device converting analogic data (coming from the source) into digital data, in order to be able to transmit data over a telecommunication line. At the other end, there should be a device (of the same type) to de-convert the digital data.
Associated with communication channels is a fundamental characteristic, the bandwidth, which determines how fast data can flow through the chosen channels. The bandwidth is defined by a rate which is expressed in bits/seconds. The more bits per second the better it is, but also the most expensive, sophisticated and rare is the technology. As audio, but even more so visual, data transmission requires many bytes, the bandwidth is a critical factor and is always the bottleneck in communication.
To take a specific example, a TV image displays (depending on the system) from 25 to 30 frames every second (25 to 30 sequential still pictures create a continuous movement for the human eyes). A single video frame contains 60 kilobytes of data and so (30 frames times 60) requires 1.8 megabytes/second to meet the television capacity. As will be discussed later, such a high rate may not be necessary for a video conference. For comparision, a normal telephone line has a very low bandwidth and if it were used to transmit one second of television image, it would take around 30 minutes. An ISDN communication channel, which usually guarantees transmission at 128 kilobits/secå, would provide only one frame every 4 seconds.
A higher bandwidth solution is provided by Ethernet connections. Ethernet is a protocol commonly used to network computers together in offices and universities. Typical Ethernet connections to the Internet allow transmission of about 1.5 megabits/sec and so it can offer 3 frames per second of high quality TV images transmission. The Ethernet solution is getting closer to the bandwidth required but it has a major drawback. An Ethernet connection shares its total bandwidth among all the users who want to commicate over the network and so it cannot guarantee a certain bandwidth. In addition, it does not control the route which the sounds and the images take to go from one place to another. Sounds and images (as well as any data) are sent as 'packets' and these just take the route which is the most available (that is how Internet is built). In some cases, two continous packets - the two parts of a sentence - may take different routes, and although this would not be a problem in itself, too many people are often using such networks. They become oversaturated and create delays and losses of packets, transforming video conferences into guessing sessions. What was gained by a high bandwidth (in theory) is lost by an unreliability which is increasing daily. The transmission is not guaranteed because the line is not dedicated.
To solve this problem, a new technology has emerged: the MBone (for Multicast backbone). To help and favour good transmission of the different packets, the idea of MBone was to create prioritary 'tunnels' to support the routing of the packets*. The MBone is a 'virtual network' layered on top of portions of the physical Internet. It uses the IP-multicast protocolså to provide multicast video, audio and shared whiteboard facilities across the Internet. Multicasting means that these data are addressed to anyone rather than to defined individuals. By making use of the multicasting processes, MBone provides multipoint connections, either one-to-many or few-to-few, while preserving Internet bandwidth. MBone worked well for some time, but it now suffers the same problems as the Ethernet solution; indeed, being freely available on Internet, it is already saturated.
To solve this new dilemma (for the time being), a new concept is emerging as a hybrid between the ISDN concept and the technical possibility of Internet. Whereas ISDN has a low bandwidth, it is on the other hand highly reliable. When an ISDN connection is established, a dedicated line, not polluted by other transmissions, is used. The entire available bandwidth is guaranteed by the provider of the service. The concept now appearing is based on this principle: to have a dedicated line between sites, but a line regulated by TCP/IP Protocol. Such a line was inaugurated in October 1995 between Europe and the US. It will be tested by teams of physicists collaborating through teleconference for a period of 15 months. The results of this testbed with real users will serve as a basis for the American state to offer such a service on a public basis from 1997. This new experimental solution shows how scientific collaboration will certainly have to move away from the saturated public communication channels, towards private solutions. Paradoxically enough, scientists provoked this state of affairs as CERN was at the origin basis of the World Wide Web which caused such a world craze for electronic communication.
Having considered the evolution of the communication channels, it is possible to turn to the third element of multimedia-based communication technologies: the devices and processes which allow any data to go through the communication channels in large quantities and in a fast and reliable way.
Two techniques are central to the processes behind multimedia-based communications.
Instead of keeping the demanding TV framerate of 25 to 30 frames per second, it is possible to work at a lower framerate by accepeting lower visual and audio quality, but with a corresponding reduction in costs. According to the type of video conference meeting (number of people involved, their movements, their activities, etc), users may choose a solution accepting a lower quality for a better price. The Swiss PTT presents their video conference services as:
"The low cost connections (128 to 256 kbits/sec) are perfectly adequate for short interviews between private individuals. Standard connections (384 to 1920 kbits/sec) are recommended for longer interactions. As to the high quality connections (896 to 1920 kbits/sec), they are ideal when the audio and visual quality is prime necessity. You just have to choose the system which is adapted to your needs, in relation with the number of participants, the time duration, the type of interaction and the movements."
Choosing a high or low framerate across a given communication channel is linked to data compression. Data compression is one of the most critical process in multimedia-based communication technologies. By compressing data, one can send more data and faster and/or use a lower bandwidth channel, more common and cheaper. But the best compression ratio demands a great deal of computation which takes time and computer resources.
Compressing/decompressing is a very active research field which is producing many compression schemes (in other words, compression algorithms). This plethora of schemes poses a big problem: interoperability. In order to communicate, each party needs to use the same algorithm, in other words, at each end of the communication process, the same standard needs to be used. Usage of different compression schemes is often the cause of unsuccessful video conference transmisssions.
There are a number of standards: H261 is an international standard created by ITU/CCITT and is one recommended for video conferencing. It gives an acceptable quality of 5 to 6 frames/sec at very low bandwidth. The MPEG (Moving Picture Experts Group) standard, defines a variable rate compresion algorithm for full motion video. The price for a real-time encoding video-board is still high but MPEG is a powerful standard.
This overview of the communication channels would not be complete if we were to keep silent about the ATM standard, which was THE buzzword at the major telecommunication fair, Telecom 1995 in Geneva. ATM, or Asynchronous Transfer Mode, is a new standard for a cell-switching networks* that defines the technology for handling heavy data loads and transmission speeds of 1.54 megabits to 1.2 gigabits per second. The ATM technology is based on a fixed (rather than variable) packet length, called "cell"å. It operates in connection orieted mode, the cells all following the same pathway defining a virtual connection between a transmitter and several receivers. The bandwidth which ATM will allow is finally meeting the TV transmission rate. From the normal telephone line to this new solution, there has been a thirtyfold improvement in bandwidth.
The key points about multimedia-based communcation are the following:
1. it is necessary to distinguish between the physical communication channels (analogic phone lines, ISDN digital lines and optical fiber) and the various devices and protocols built on top of them and which manage the data transmission;
2. a dedicated communication channel (ISDN or the new IP lines) ensures reliable communication, because the provider guarantees the bandwidth whereas the Ethernet, MBone and other LAN (Local Area Network) solutions cannot be considered reliable essentially because of saturation of the networks; the distinction is between dedicated and shared communication channels;
3. the higher the bandwidth, the larger the quantity of data (images, sound, text, etc) that can be transmitted, but the more expensive and less frequently available it is;
4. to use a low bandwidth but maintaining an acceptable visual and audio quality, the data must be compressed and coded at one end, decompressed and decoded at the other end;
5. there are essentially two main solutions: the network solution (also called packet-based system) available with almost any phone line and the CODEC solution* (also called circuit-based system) which requires the installation of a special ISDN line. The first is possible in almost any office, researchers' desk or home, the second is still reserved to special conference rooms.
This fifth point introduces a user's perspective of video confrencing. In concrete terms, for the user, what is a video conference setup?
There are essentially two video-conference configurations: either group (or meeting-room) based or individual (or workstation) based. The first almost only uses ISDN lines (though compressed video is well established in North America), the second the different communications channels (ISDN line, normal telephone and network solutions, such as Ethernet and Mbone).
The first relies rather heavily on equipment which is usually housed in special video conference rooms. For example, at CERN, there are two such rooms, one for a maximum of 15 participants, the other for 30. They are equipped with a main camera, a main and an auxiliary monitor and a document camera. The video conference system is Picture Tel with a H320-Codec capable of operating, on an ISDN link, at 56/64 kbps up to 384 kbps (other compression rates are included: H221, H261, G711, G722 and G728). To euip a room with basic Picture Tel equipment costs approximately 35,000 SwFr. plus extra costs for more bandwidth (e.g. 384 kbps) and other local equipment such as two screens. VTEL is another frequently used system, especially in the US.
The second configuration relies on the desktop video conference solution based on a computer. It is designed for personal users and is developing very rapidly, essentially because of its modest prices, its flexibility and availability. The users communicate through computers and can, not only see and hear each other, but also work on similar documents. Such an environment may be more or less sophisticated and may involve a large variety of tools.
A basic desktop video conference environment usually consists of a computer, a small video camera set on top of the computer screen, standard telephone lines and desktop video conference software. ProShare Personal Conferencing (Intel)å, Person-to-Person (IBM), Show Me (Sun) and CU-See-Me (Cornell Univ.) are frequently used destktop video conference software. The first two are commercial products often used by companies which may run either on a LAN (Local Area Network), on normal telephones lines or on ISDN. The last two run with an IP protocol to operate across Internet; these are most frequently found in universities and research labs. Some software, such as the ones used with the MBone solution, are sharewares; these are for example, Visual Audio Tool, Network Video, Whiteboard (shared window), Session Control.
As an example of a commercial desktop video conference, ProShare Video System 200 includes a small video camera (to be set on top of the computer screen), a video card, a phone card, a combined headset/microphone and the software. It enables one PC user to see and talk with another PC user at a remote site. Both users can call up the same file and work on it simultaneously. Each conferee's image appears in a movable section of a Windows screen; the remainder of the screen is devoted to the application, which can be edited, penned on, or pointed to from either desktop. The system is also available without video. In November 1995, the whole package costs approx. £1,550 plus a 33-megahertz 486 PC.
Whereas the meeting-room solution is (usually) restricted to groups, essentially because of its cost, the workstation-desktop solution is used more and more for small gatherings of a few researchers or employees. Because of price and flexibility, the desktop video conference solution can expect to see very rapid development. From being a group to group visual communication process, multimedia-based communication technology is becoming a more intimate process, where one or a few individuals share and collaborate on a similar project, program, document, video sequence, etc.
Another important feature which is often mentioned when one describes a video conference session is its capacity to handle two to many sites. One thus hears of point-to-point (2-way communication) or multipoint (many-way communciation). This characteristic (point to point or multipoint) is usually attached to the video conference links based on CODEC ISDN systems. Point to point means 2 sites communicate between each other without any intermediary, whereas the multipoint configuration requires a central unit which manages the communications. This unit is called a MCU, for Multiple Control Unit. Whereas network solutions may in theory allow an unspecified number of users, the number of sites linkable with a MCU is limited. As an indication, the statistics for the use of the ESnet* MCU in October 1995 show that only five video conferences involved more than five sites, seven involved five sites, 30 four sites, 45 three sites and 23 two sites.
The integration of seperately developed technological solutions to video conferencing are emerging. For example, technicians at the CERN are working on gateways to allow Internet and ISDN/telephone to communicate. The goal is to allow people at their respective workstations in different European labs (IP network connected) to participate to a working session held at the CERN video conference rooms (ISDN connected).
Another mixed configuration used at the CERN is a desktop linked to a TV monitor. This allows a larger attendee to follow the conference and to interact if required. The variations of mixed configurations are numerous and they open the door to many forms of scientific and other collaborations. Here are some possibilities envisaged by Galvez (1995):
a. video conferencing using Conference Room equipment and collaborative work using Desktop (Conference Room used for transmitting video and audio over ISDN and Desktop for applications or documents sharing over the IP network);
b. audio using Conference Room and Video and/or collaborative work using Desktop or video using Conference Room and Audio and/or collaborative work using Desktop (mixed way of transmitting audio and video, depending on problems occuring during connection or to save money on ISDN link);
c. a gateway to join Desktop and Conference Room world (as mentioned above);
d. the possibility to follow a Conference Room conference from the office, using the workstation (video and audio signals is taken from the Conference Room and transmitted over the Local IP network to be received by remote workstations);
e. the possibility to record Conference Room conference on the workstation hard disk for later vision on a television screen or by another host using internal or external IP network;
f. the possibility to use the workstation for playing video and audio clips and transmit them through Conference Room (video and audio feed are taken from the workstation and plugged into the Conference Room)
g. the possibility to play recorded conferences using VCR (Conference Room) and to broadcast them through the network;
h. the possibility to use television screen and audio capabilities to follow a MBone conference if multicast possible (a comfortable conference room is taken advantaged of with big television screens and good audio quality to follow a MBone conference received over IP at a workstation).
It can be seen that the frontiers between the different systems are blurring away. The technology is becoming more and more capable of uniting systems, techniques which at one point were far away. For the users, it will mean easier systems (because more compatible), but mostly more possiblities to work together, more flexibly, more cheaply and with a better sound and image quality. Users will just need imagination.
This section provides a series of examples of the multimedia-based communication technologies described in Section 6.2 being utilized by distributed research communities These examples include five academic communities and two business-based networks. The examples are described in Sections 6.3.1- 6.3.6 and include:
a) the Recto-Verso network (Section 6.3.1), a distributed research network using (so far) room-based compressed video conferencing system between small or large groups of researchers; the description includes the type of activities during these sessions, the rationale for using such media and the other electronic media which this group is planning to use to collaborate;
b) the ERVN physicists network (Section 6.3.2) is a larger network using a room-based compressed video conferencing solution with a MCU (supporting multipoint video conferences). The logistics which such configuration requires are discussed;
c) the medical labs network working on the Genome Project (Section 6.3.3), which uses a desktop video-conferencing solution between individuals or small teams, a solution which is spontaneous, flexible but intrusive in some cases;
d) the ATLAS and the CMS physicists networks (Section 6.3.4), which use the Mbone technology , experimenting with a mixed configuration: room-based (compressed via ISDN) and desktop (via the Internet TCP/IP protocol);
e) two businesses (Section 6.3.5), which illustrate how both individuals and large groups use video conferencing in business, their rationale and types of uses (when and for what);
f) the Télé-Université of Montréal (Section 6.3.6), which is experimenting with a sophisticated multimedia video conference environment.
These sections of the report will focus, as much as possible and wherever the data are available, on users' perceptions of multimedia-based communication technologies. The major contribution comes from members of the Recto-Verso project who have been using multimedia-based communication technologies for some years within and outside the Recto-Verso network. Section 6.3.7 concludes these descriptions with an overall analysis of the application of multimedia-communication technologies for distributed research communities.
Recto-Verso is a distributed (human) network and project which involves researchers from various sites in France and in Québec. It is the follow-up of a project called "Procédés et outils de Téléformation" which started in 1993. This preliminary project set up the collaborative process between the French and the Québec teams and was, according to one of its members, indispensable for the present collaboration. Even though the collaboration stayed at that time minimal, this preliminary project helped them to become familiar with one another.
The research object of Recto-Verso is the use of multimedia communication technologies to deliver distance and open learning activities; it is also to set a model for distance collaboration. In 1995, the project entered a more operational phase after the Québec government, through its "Fonds de l'Autoroute de l'Information - FAI", allocated it an important budget. In June 1995, the network members held a major face-to-face meeting in Lille (France) where the operational framework for the project and its upcoming activities were discussed and the communication media chosen. During this event, a video conference was used to allow Québec partners and official representatives to participate to the meeting.
Even though 1995 marked an important time for the network, the partners had already been collaborating for more than two years through video conferences. About twelve two-hour video conferences have been held, at intervals of about two to three months. Except twice when three sites were involved (once Montréal, Lyon and Paris in March 1995, the second time, Lille, Montréal and Québec in June 1995), the meetings were point-to-point video conferences, with the French partners all participating from one site, in Paris.
In the official Recto-Verso document (1995), an overview of the planned video conferences is given:
"The video conferences are to be held between Montreal and Paris. Four two-hour meetings for coordination (February, April, August and November 1995), one four-hour plenary session during the Lille seminar in June 1995 and four two-hour sessions (February, April, August and September 1995) about the following themes: setup of the network; reengineering of existing activities; setting of the generic model; design of new activities."
As an example of the issues typically discussed during the Recto-Verso video conferences, here is an example of the agenda of the March 1995 video conference:
Visioconférence 24 mars 1995:
10h 00 - 12h00 Montréal 16h 00 - 18h 00 Paris et Lyon
1. Etat de la situation au sujet de RECTO: - au Quebec; - en France
2. Besoins de communication et lien électronique entre les chercheurs
3. Besoins et mécanismes de coordination du projet
4. Séminaire de juin 1995: - objectifs; - programme; - participation; - dates et lieu
5. Varia (Les autoroutes de l'information - Projet VERSO, Janvier 1995)
As another example of the organization of a project meeting, an multi-point conference was held in June 1995. This one was an official, formal event during which people in the Quebec province from different ministries and administrations (in other words, the financing entities) were presented with the state of the project. Participating in the event was one site in France (Lille) with 6 active participants and a silent crowd in the background and 2 sites in Quebec, one in Montreal and one in Quebec with approximatively 10 to 15 participants each. The video conference system was Picture Tel (with a bandwith of 128 kbits/sec (i.e. two 64 kbits/sec channels), a rate of compression of H320 and a framerate of 15 frames/sec).
The meeting was structured in the following way. Everybody from the three sites introduced themselves; then, alternatively the six participants in Lille synthetised the major decisions of the Lille workshop, essentially the planned activities and the future outcomes for the project. From time to time, the Québec participants asked questions of clarification. The interactivity was very low at the beginning but then increased; the usual pattern was one question followed by an answer; no real debate was started (in other words, there were never more than five interactions on the same issue). The major objective of the meeting was for the people in Lille to inform the decision-makers in Quebec. It was not a working session, but rather an administrative and political event.
The Recto-Verso researchers have so far used video conferences essentially to tackle administrative, logistic and even political matters, but not to work collaboratively on their research activities. The fact that the main collaboration focused on preparing the project proposal to get the project financed may explain the type of video conferences held so far. According to one of its members, the collaborative activities have not really started yet because it needs time to set a proper environment to work collaboratively at a distance, especially when people are from different cultures. However, two of the video conferences (which were to be held in June 1995 but which did not take place for technical reasons) were much more content- and research-activity oriented than purely logistic in focus.
Interviewed about the rationale for using multimedia-based communication technologies (in and outside their network), the Recto-Verso partners mentioned the following aspects.
For one partner, video conference is the only possible media for a team of researchers so physically distributed; it is also good, because it brings people together at the same time:
"For me, it has 6 major functions in our network:
1. to know where the project stands at each end
2. to plan the next activities
3. to convey our respective conceptions and to try to develop
a common understanding of some concepts in
order to develop a common reference framework
4. to discuss about the content of the project itself
5. to set the next video conference meeting
6. to maintain motivation.
The video conference is a tangible manifestation of the collaborative work and the fact to meet is the occasion for the colleagues to show the others that work has been done and that it needs to go on. The video conference has a certain emulating effect. Our video conferences may be seen as temporal marks for the project. We handled administrative and logistic affairs as well as matters related to the content of the project."
For another partner, video conferencing in Recto-Verso is before all functional and economic:
"In fact, we do the same work than in presential meetings but as we are so far away, the video conference is the only way we can meet. It allows to maintain regular bridges between the teams, it is good for flying over, for general issues, for questions of organisation. But video conference stays a very formal thing, it is hard to grasp the people, what emanates from them is cut. There is no social weaving, you dont really belong to it. It is frustrating in a way. Audioconference is better, especially for two or three persons who know each other. But the ideal, for me, is to have a continuous video conference link to allow for the spontaneous gathering of the researchers as their needs arise. It thus allows to work for short periods on precise and sharp issues. It is by far less formal and is closer to the life of a non-distributed research team. Concerning the formal video conference, the ideal number is 6 persons at the different sites. Be careful to keep similar size at each site; as duration is concerned, it should not last more than 2 hours."
For a third partner, video conferencing is good for taking stock, for communicating a lot of information, for taking decisions:
"In comparison with presential meetings, the time of reaction is not the same, you dont react as quickly, I cant explain why; there is something mystical, which makes the people more rigid, more formal; it intimidates me a little; in addition, I see myself. As to the frequency of our meetings, once every three months would be a good rhythm for the large video conference; to work in the small groups, once a month would be ideal."
For another partner from Montreal, video conferencing is good for a wide range of purposes:
"Our video conference room is used daily and constantly and for any type of activities: discussions, seminars, workshops , competition jury, etc."
In Recto-Verso, alternative media such as telephone, fax and email, have not been used much so far. Audio-conferencing has also not been used and the reason had to do with opportunities rather than with the relevance of the technology. Indeed, one of the objectives of the project was to test the relevance of various communication technologies to support collaboration. However, some small telephone conferences had taken place between Quebec and the main French partner, "in order to take counsel together as to certain points"*. Surprisingly, electronic mail, had not been used much:
"We tried to resort to electronic mail, but it did not work. I think that the French colleagues were not much used to working with this technology. And after quite a few unanswered emails, I renounced."
An important part of the Lille meeting in June 1995 was devoted to choosing the technologies to be used by the Recto-Verso partners. The discussion is a good example of the different aspects and resources taken into account by researchers wanting to collaborate at a distance.
Regarding electronic mail, it was agreed to use it regularly as a privileged communication media. The following criteria were set: to use a system available on PC and Mac; to use one which respects the special characters of the French language; to consider carefully the question of attached documents and the file transfer format and to organize quickly the distribution lists.
When discussing electronic conferences, the major discussion was about using an open or a closed system. The group agreed to make no compromise about a commercialised stable product with a good support. The importance of managing and moderating the electronic conferences was stressed.
In discussing video conferences, the following aspects were raised: will it be a system handling small or large groups of people? Will it be a point-to-point or multipoint video conference? It was agreed for any system being considered, to first check the availability of the system at each site, its interoperability and the standards in the various countries. The second task will be to see how to migrate from using the ISDN link to using the two national scientific electronic networks, RISQ (Réseau Interfacultaire Scientifique du Québec) and RENATER. In order to support multimedia-based communications, new functionalities will be added little by little. The ultimate objective is to move towards, or at least, to test an ATM link, whenever this technology is available. The objective is also to move from point-to-point to multi-point video conferencing.
Deciding on common tools to produce documents was also a major issue in the discussion. It was agreed to try to use the same word processor and to define common styles; Word 6 was recommended because of its appended graphical editor, but it was not available at all the sites. The issue was between choosing software with limited functionality (but used by almost everybody) or more-sophisticated software which was not yet available to many. The issue of presentation and publication was tackled briefly and used by some as an argument for choosing the latest word processors, as some offer these options. To manage the project, a tool such as Excel was suggested. The participants agreed that, for the moment, data-base tools were marginal in the project. The question of collaborative working tools was left opened but agreed to be further investigated.
For public dissemination, the participants agreed on the creation of pages on the World Wide Web and to choose the Netscape software. Finally the location and the types of server to be used were discussed.
Recto-Verso partners' discussions show five important elements of multimedia-based communication technologies:
1. Being able to 'meet' participants, all and at the same time (synchronous conferences) which was offered by video conferences reinforces the feeling of belonging to a group and of working together. In a way, it is a visible sign that the project exists.
2. By comparison with electronic mail and electronic conferences, the synchronicity given by video conferences seems to create a more collaborative action dimension. As one member said:
". . . with electronic mail, you never know who has read your messages, you don't know who knows what; whereas a video conference is a way to have the people in front of you and to verify that everybody has the documents and understands the same thing."
3. Despite the inflexibility, complexity and cost of this technology (some French partners had to travel a few hours to participate in the video conferences), it seems to have led partners not used to electronic communication to accept the complex technology (video conferences) rather than the more flexible one (electronic mail). Synchronicity seems more important than ease of access. The fact that the project was not fully established may be an explanation; at this stage of the project, punctual meetings may be more important than continuous email interactions; the latter need more management and rigour than a meeting once every few months.
However, in the new phase of the Recto-Verso project, after using video conferences more than 10 times, the partners have finally decided on a clear electronic mail and electronic-conference strategy for transcontinental collaboration.
Interestingly enough, the same pattern of media introduction (first the complex media, than the easier one) is found at the Télé-University of the Québec province. For their daily working activities, they have been using video conferencing for at least 10 years, but email for only 5 years.
4. Recto-Verso usages of video conferences so far, indocate that this technology seems appropriate for administrative and logistic affairs but less for matters regarding the research content. According to one participant:
" . . . decisions were taken at each video conference meeting. These decisions bore either on the content (understanding of the main concepts, of the project methodology, of its problematic and of the anchoring of everybody's efforts and activities) or on administrative and logistic affairs. To create a common framework is the most difficult thing with video conference, because we belong to different universes."
5. Video conference is usually not used alone; it is most often supported by other media such as fax, telephone and email. For example, the video conference agenda is often faxed a few hours or days before the virtual meeting starts.
The adoption of video conferencing by the members of the ERVN physicists network is representative of an increasing number of scientific collaborations. ERVN (The Energy Research Video Network) is the result of "the increasingly extended nature of high-energy physics collaborations". An interesting account of how video conference was adopted by this group of national laboratories and universities around the world is given in Science:
"Much of the impetus came in the late 1980s from Stu Loken, director of the Information and Computing Sciences Division of Lawrence Berkeley Laboratory (LBL). Loken thought video conferencing might prove a valuable tool for high-energy physics collaborations, which involve many researchers in widely separated locations. Video conferencing had been restricted to television studios and wealthy corporations because it required expensive dedicated lines. But Loken saw an opportunity to apply it to physics in January 1990, when the Department of Energy upgraded the bandwidth of ESnet, a nationwide computer network serving the research needs of DOE's energy research community. The upgrade gave ESnet enough capacity to handle the 378 kilobits per second required by existing video equipment.
Loken set up a link between LBL and colleagues at Fermi National Accelerator Laboratory, the hub of two large top-quark searches, DO and CDF, that drew on hundreds of collaborators at outside laboratories. The system was rudimentary. Each conference room was equipped with a monitor and a camera that took in a table with no clear view of any one person; the image was transmitted to the monitor at the other end of the link. A separate camera and a monitor were used for presentation material. The setup was not much different in principle from the ones used for years to link TV studios, notes Loken, but "instead of a single talking head in a studio, you were taking in a round-table discussion with lots of people talking and many overhead presentations."
(Crease, 1994)
More than 20 laboratories around the world now belong to this network and it remains in growth. The central node is still at the ESNET Lawrence Livermore National Lab (USA) which holds and manages the Multipoint Control Unit (MCU). An ISDN connection now links the CERN with the LLNL. Japan and other countries are connected. From the CERN side, the most frequent ERVN video conference meetings host between 5 to 10 people, but sometimes 30 participants are involved. The main activities held during these meetings consist of: project follow-up; general coordination and administration; planning future activities. Participants also frequently use email, but telephone and fax less frequently.
The ERVN network shows how video conferencing is allowing collaborations which were never thought of before because of travel costs. In addition, they ensure continuity. Concerning the advantages of videconferencing, one physicist says:
"It's great for presentations, better than just faxing visual materials, because you get explanation and amplification of the documents."
Another US physicist speaks of being able to "stay in touch with experiments planned at the CERN" and "of sharing transparencies, drawing on whiteboards and quarrelling with speakers". Another mentions that it will "open projects to input from poorer institutions and countries whose voices might not be heard otherwise" and that it will improve the relations between faculty supervisors and graduate students working at distant facilities: "as any adviser knows, students tend to drift, and video conferencing makes close supervision possible without pulling the student back to campus and interrupting work."
Technical implementation is however a drawback as it still requires expensive equipment. In addition, the communication cost, especially in Europe where the market is not yet deregulated, is a second drawback. For example, the official Swiss rate for 1 hour video conferencing from CERN to the US is 270 Sw.F. (230 US$). Another important aspect is the management and planning of the video-conference sessions; they cannot be improvised. The MCU (Multiple Communication Unit) from LLNL is managing the booking, the setting and the tuning of the sessions. Not only do such room-based compressed video configurations (using the MCU) require logistical management, but in addition the number of sites involved is limited. As mentioned briefly earlier, 161 of the video conferences managed by the ESnet in October 1995 involved 2 to 3 sites and only 12 video oconferences involved 5 or more sites.
There is finally a psychological drawback: the formality drawback of video conferencing:
" . . . while the existing conference-room-based scheme works well for formal presentations, when it's a free, brainstorming discussion that pops around from one person to another, it works much less well. The pictures are kind of jerky, and it's difficult to see who's speaking. Informal discussions may go more smoothly with a new, workstation-based technology for video conferencing.
In these systems, camera and monitor are installed at the user's desk, with both the audio and video packets going through the Internet rather than across hardwired lines. . . . A desktop video conference is a system that will make it possible to join and leave a video conference during the normal flow of work and to analyze data together with colleagues at remote workstations."
By comparison with Recto-Verso, ERVN is a much larger network with more sites and thus more users. It is also a real multipoint configuration, which thus requires a MCU, the Multiple Communication Unit. The necessity of the MCU and of the LLNL as the organizer shows how video conferencing needs, at this level, very a well defined infrastructure and planning. Logistically, it can be said to be complex. Video-conference sessions are formal events rather than spontaneous gatherings of researchers collaboratively working on projects.
The third case shows a spontaneous and flexible solution but one in which intrusiveness can be a drawback. The network involved uses a desktop video-conferencing configuration.
The director of the University of Minnesota Medical School's Computational Biology Center holds video conferences with people located in a variety of sites: on the next floor; in the next building; at Michigan State University; at the Centers for Disease Control; and even halfway across the world at the Center for Scientific Computing . Video conferencing is seen as valuable because the joint research between the University of Minnesota and Michigan State University (for the Genome project):
" . . . requires close collaboration between the sites. Telephone calls alone won't suffice. The highly technical nature of the work makes it essential that we don't have any misunderstandings or miscommunications."
Each conference participant appears in a separate window on the computer screen, along with his or her name and network address. As many as seven people have been hooked up to a desktop video-conferencing session over an Ethernet network running at 10 megabits per second. The economic factor is an important justification for the promoters:
"we've spent less on the technology than we would have on a single meeting . . . it's already paid for itself. The price per workstation amounted to only US$1,200 more per workstation than the base hardware and software."
Apparently the media allows the scientists to work very closely, on a continuous basis and very flexibly, up to the point that it sometimes becomes intrusive. Indeed, without warning, a window may pop open on the screen, and someone would be there watching you, "I found people invading my life", says one of the participant. But many systems allow either to set a warning signal so that video conferences cannot begin without the user's permission, "much like choosing whether or not to pick up your telephone".
It is interesting to note that the media seems to make the participants overly concerned about their physical appearance. According to two researchers working on this group, "people tend to be extraordinarily concerned about how they look to others in the video-conferencing session." Because video-conferencing systems typically allow participants to view themselves as well as others, "most people tend to watch themselves, comb their hair and otherwise monitor their own behavior more than they would in a face-to-face meeting," according to a member of the Network. One of the participants admits he uses the self-monitoring feature, keeping a window open on his screen that displays how he appears to others in the meeting. This, however, isn't just vanity, he insists: "I didn't realize until I started using this how expressive my face is, and how much I give away. These visual clues are absolutely essential when talking about technically complex or politically sensitive issues." Apparently and according to this researcher, desktop video conferencing seems to provide many visual cues.
This experience of video conferencing points to two interesting aspects. On the one hand, the desktop solution described seems to solve the formality drawback pointed out in the ERVN example. Flexibility, spontaneity, simplicity, adaptability and economical reasons characterise it. If managed properly (so as not to become intrusive), the media seems adequate for spontaneous gatherings of researchers. It is however interesting to see in this case that the image component seems to take on much importance for the users. It is as if one has to learn to manage one's relation to the image. Whereas in face-to-face meetings, one does not have an image of oneself and cannot control what one sees, in desktop video conference, it may be the reverse. This point, the role and importance of image in multimedia-based communication technologies, will be re-examined later.
Desktop video conferencing with Internet connection is the option being experimented with by two other networks in which CERN physicists participate: ATLAS and CMS.
The ATLAS collaboration consists of 1500 physicists from about 150 institutes worldwide involved with detectors and test-beam activities. In order to allow people, especially from USA and Japan, to take part in the CERN activities and meetings without having to travel, video conferences via MBone technology have been experimented with for the last two years.
The initial goal was to broadcast the plenary sessions of regular week-long software meetings to outside institutes. Matters such as finding the optimum detectors and the site in which to locate them are discussed, results presented, decisions taken, decisions such as the number of layers in the detectors. According to one of its members:
"The media is good enough for people to make their opinion known."
But it may also be more. For the last event:
"the important point is that it allowed people from the USA (who would normally have missed the meeting since they could not travel) to participate and take part in the decision making process on the building of the ATLAS detector."
From their standard Unix workstations, members are able to follow the meeting held in a conference room at CERN. The equipment of such room includes a Sun SPARC 10 with 17" screen, a wall-mounted camera to view the projection, a screen connected to the Sun's video input card, four wall-mounted speakers and an amplifier, four suspended microphones and one microphone on a stand*. It is thus a mixed configuration: video conference room-based and desktop, both through MBone. Public domain softwareå through MBone, rather than the expensive and complex ISDN and compressed-video solution, has been chosen.
Initially between only two sites, CERN in Geneva and Amsterdam, the events now involve more sites. Multipoint connections require the multicast capability, the possibility to offer multimedia on Internet to anyone who has the necessary software on their workstations. Every six weeks, a full-week meeting is held at CERN in Geneva. The plenary sessions are broadcast on the MBone. In small sessions, the desktop participants can intervene and ask questions to the room-based people. So far, from 3 to 50 desktop participants have been participating in the room-based meetings. Many of the larger meetings are announced on a world-wide distribution list, and the schedules are available on the Webí.
To minimize bandwidth, the relevance of the audio and the video is carefully considered.
"Currently, we do not project the computer screen image but will do so soon with a LCD projector. If the participant sends video, it appears on the computer screen. However to minimize bandwidth, we try to avoid sending video in the other direction unless someone wishes to show a diagram or plot. We even had someone give a talk from Amsterdam but only the audio was transmitted, the transparencies were printed from WWW before the meeting and put on the overhead projector by someone in the room at the speaker's request."
(Chris Onions (CERN), personal communication)
Apart from the mixed configuration of room-based and desktop, some ATLAS members use the desktop configuration exclusively, calling up sessions whenever they feel the need to collaborate. Apparently it is very easy to do. One participant with the Session Directory software calls different participants. At the other end, the computer bips and the person accepts or declines the invitation. If he/she accepts, his/her name is added to the list on the Session Directory screen. If the Network Video software is available and open, his/her image may also appear (providing he/she has a camera on his/her computer). When everybody has been called, the session can start. The Whiteboard software allows the participants to write, draw and work on a piece of text, on programmes, etc. Each participant has his/her own colour for a pointer and can freely intervene on the shared whiteboard. The ATLAS members are just starting to use this media in this way. Approximately 20 people have been involved so far, but more are interested in the solution.
Whereas in the room-based compressed-video solution, the multipoint connections required the MCU device, the LLN logistics and money for the communication (more than 200 Swiss francs for a one-hour video conference), here the multicast capability is much lighter, logistically speaking. It is even very flexible and spontaneous and does not directly cost anything for the users.
However, the technology is not easy to use: it requires from users sophisticated technical skills, to 'patch (i.e. to modify) their computers'. In other words, it is still a solution for computer-skilled people. Quite a lot of tuning has to be done before a new member can be properly connected and the lack of standards (compression especially) is a major problem. The system is also suffering from the saturated Internet bandwidth which causes many packet losses. As far as losses are concerned, and according to one of the ATLAS members from the CERN, the audio link is by far the most critical and most important component. If the packet loss is more than 10%, it is impossible to follow correctly the event. If the audio quality becomes too irregular, the alternative is to use audio devices plugged to the telephone line. The solution is reliable and of good quality but more expensive, as it uses telephone lines. With the image, packet loss is not so critical, a 20% packet loss is still manageable. According to this CERN member and because of the saturated Internet bandwidth, using the video possibility to broadcast peoples' images is not used extensively (although it is sometimes good to see people's reactions) as it is a waste of bandwidth. The video possibility is used above all to transmit diagrams. Sometimes the diagrams are also made available on the World Wide Web.
Until MBone comes on ATM with dedicated lines, this technology will stay possible but unreliable. However, the MBone solution for ATLAS makes it possibile for the scientists to attend more meetings. It saves on travel costs, but it does not replace face-to-face meetings.
The CMS network is also a large physicists network with more than 1200 members from 135 countries. They use the two video-conference solutions: the compressed video and the MBone desktop and/or room configuration. They are also experimenting with gateways between the compressed video and the Ethernet worlds.
The compressed-video solution is usually used for meetings linked to managerial issues; financial aspects as well as planning and coordinating the network activities are their main objectives. The video conferences are held at a maximum of four sites. As with ERVN, the CMS network also uses the MCU of ESnet. Betwen ten to fifteen people at each site attend the meetings. The same complaints mentioned by the ERVN participants are also made by the CMS users. One of the CMS members from the CERN underlines the lack of compression standards which make connections with new sites an adventure. He also mentions the importance of having a certain discipline to participate to these events. According to him, new users should receive a small briefing to understand how to behave in video-conference sessions. This is essentially motivated by the two major limits of the media: the time delay which does not allow a smooth interactivity and the fact that the compressed-video solution reacts to the noisiest site but not necessarily to the relevant one. Indeed, the compressed -video configuration used by CMS allows only one site at a time to be broadcast.
According to CMS members, some of the recommendations for new video conference users are the following:
To ensure a smooth transmission:
· one person should speak at a time;
· little side remarks which punctuate so many face-to-face interactions should be avoided;
· side-site discussions should be ruled out;
· there should be a chairman of the videoconference and the sessions should be prepared in advance and managed;
· about slides, speakers should use large fonts and not move the slides in front of the camera;
· an ideal session (for the CMS group) lasts between two to four hours.
Other recommendations from CMS members are even more specific:
One should of course always bear in mind that a good setup is needed in order to get a satisfying result! For good audio quality, the room should be quiet (no fans, air-conditioning, etc. and no public room like terminal cluster, computer room) and a good external microphone and loudspeakers should be used. Often the attached microphones are too sensitive and of bad quality. The speaker should speak clearly and slowly, and the other participants be quiet.
If transparencies are to be shown, a stable image support should be available. Take the image either from a projection screen or a stand, don't 'hold' it just in front of the camera.
A 'session operator' might be useful for covering a bigger auditorium (change camera-view, control microphones, set program parameters, etc).*
With world collaboration, a last problem mentioned is the time difference between the sites (especially between the CERN and California); this makes the timing of the sessions sometimes complicated and also restricts some collaborations.
In the case of the desktop MBone video-conference solution, CMS participants are usually more computer skilled than those using the compressed-video solution. Following is text from the MBone Conference Announce Procedure on the CMS WWW page (see footnote). It can be seen that no basic user can easily get through the procedure:
"To get things moving, we would like users to set-up (some..) workstations with the public domain software (mmcc, sd, nv, vat and wb) for the packet-based solution. These programs allow already to receive video and audio in a point-to-point way. If in addition the workstation is put on mbone (see the faq for this!), multi-point connections are possible. Those who also would like to transmit will have to purchase a suitable camera and interface card. To allow us to register and make available all such workstations, we request you to send us your setup parameters (name of laboratory, name of responsible person and email, names of users whom one can connectto on this workstation, system and domain name (with the Internet address),which software and hardware is available and the .mmccrc entries)."
The Mbone solution is used individually afor small-team gatherings. In this case, a large screen is used and lately the sessions have been recorded and made available on the Web; as it takes a lot of bytes, the new video-conference session replaces the previous session every two weeks. Without allowing the people to interact and really participate to the meeting, it allows them to attend the session at their leisure; for sites with different times, it is a satisfactory solution.
Up to seven sites have been connected so far but as with the compressed-video multisite meetings, as soon as there are more than two sites connected, there should be some management of the interactive process. One recommendation is to make the turn taking very explicit, to have regular "tour de parole" allowing each site to intervene. Another recommendation is to speak distinctively ('stuttering people are excluded') and in a continuous way to ensure the packets' continuity. It is also important to mark clearly the beginning and the end of an intervention and not to drift along.
To conclude with the CMS use of video conference, both desktop and compressed-video room-based solutions are seen as complementary, fulfilling different needs and suiting different people. compressed-video demands complex logistics but is easy to use; MBone is light and flexible but more difficult to use, restricted to computer-knowledgable people. Also Internet is saturated, whereas the compressed-video solution is reliable. As in the ATLAS network, much hope is put in a dedicated transatlantic line regulated by the TCP/IP protocol and in the ATM technology.
The two following cases illustrate how and why both individuals and large groups use video conferences in businesses. They provide interesting illustrations which may be relevant to research communities.
For Blue Cross and Blue Shield of Florida (LaPlante, 1993) choosing video conferencing as their way to meet has been done primarily for economic reasons. One of the company's executive vice-presidents uses it mostly for quick interactions with his immediate colleagues out in the field. He thinks that:
"it's extremely effective for one-on-one sessions, where you can be fairly close in on the camera; . . . despite the inevitable time lag in audio and visual response, it's 75 percent as good as being there. Video conferencing is also effective for widespread information sharing. Four or five locations can be hooked up in order to disseminate a report that is broadcast from Blue Cross' Jacksonville headquarters. Straight stand-up kind of reporting also works well. . . [It] is less well suited for interactive problem-solving meetings. You have the time-delay problem that interrupts the natural flow of two-way dialogue. Then you don't see everyone at once in a multiple session. You miss a lot of the visual cues."
This executive vice-president appreciates video conferences because he "wastes less time traveling from site to site and thus feels infinitely more productive." However, one of the drawback is that:
"the sheer volume of meetings has actually increased as a result of video conferencing technology. Because we can pull people together more easily, we have meetings that we wouldn't have had before. This has actually increased, not decreased, the percentage of my time I spend in meetings."
In another company (LaPlante, 1993) video conferencing means reduction of travel time, increase of productivity because of less time out of the office, short and frequent efficient meetings, easy document transfer and ease of recording meetings. But videoconferencing will never replace face-to-face interaction for a sales call, for persuading and establishing a relationship; face-to-face communication is still the prefered method.
The last example is of an experimental environment tested at the Montréal Télé-University which includes a wide variety of communications tools. In contrast to the other cases reported, it was not a naturalistic setting but it points to the kind of environment with which researchers may be working in the near future. It is basically a multimedia video-conferencing environment.
The hardware is quite classical and involves PCs (386 and 486) each with its own camera, microphone, modem and network card. In addition, the following tools and resources were part of the environment:
1. multipoint video conference;
2. electronic conference (synchronous - asynchronous);
3. electronic mail;
4. editing functions to allow the design, implementation and sharing of documents with text, sound, graphics and video sequences;
5. annotation functions for multimedia documents;
6. access to a multimedia document databank for browsing, extracting and adding new segments;
7. screen sharing (multi-windows);
8. multimedia file transfer;
9. team logbook;
10. private notebook.
The first experimentation involved three students (in Montreal and in Quebec City) and was training oriented. The three students had to collaboratively develop a critical reflection about telecommunication and its applications (this activity called 'flash critique' was part of a distance course about computer sciences). Not only did they have to discuss, but also to produce, synthesize and present multimedia documents. They also had to plan their activities. To make the collaboration easier, the three participants were assigned different roles from the start: the secretary, the animator and the editor/speaker.
Although the main objective of the exercise for the participants was to assimilate a certain number of technical concepts, they were working on concrete cases like any team of scientists. The participants organised their work and used the different resources in the following way:
1. Organisation:
· video conference (video/audio presence);
2. Individual preparation and design of the 'critical flash':
· documents surfing, data banks, file transfer, wordprocessing and editing, electronic mail;
3. collaborative work:
· annotation, email;
4. collective synthesis:
· data banks, screen sharing, editing, email;
5. final editing of the 'flash':
· data banks, shared file, editing, email;
6. final presentation:
· video conference (video/audio presence)
The schedule above shows that video conferencing is used at the beginning and at the end of the team's work; at the beginning, to get to know each other, to determine the different roles and to plan the future activities, while at the end, to present the product of the team's work to external people and to get their comments. It is interesting to see that a video conference seems to be used for organizational rather than for operational matters, in other words, for setting and managing a project (or activity) rather than for developing and designing it, for the form of the project rather than for its content. This trend is confirmed even in the last video conference, which could have been more content oriented; however, "la presentation (du produit) n'a porte que tres peu sur le contenu reel, mais plutot sur ce qu'on aurait aime faire ou voir dans ce document." (Rigault et al, 1994, p. 17). It could be that the complexity of the technology creates a formality which is more associated with managerial issues than with discussion of content.
For interacting during collaboration, the synchronous electronic application-sharing conference was used little and was felt to be 'excessively laborious'. The reason was essentially technical; indeed, as far as turn taking was concerned, the least movement of the mouse blocked any interactions. As to the asynchronous electronic conference, participants used it and found it to be of critical importance and the same applied for electronic mail which they used for private interactions. Thus it seems that for working collaboratively on the design of their product, the participants looked for the most spontaneous and most flexible media: reliable and available when the needs arose. For people who work on a similar task, these parameters seem very important.
According to the evaluators of the experiment, participants lamented the technical problems which biased the results*. They found the textual conference an adequate medium for interaction when working on a project, whereas synchronous electronic conferences were considered to be technically cumbersome. The notebook was rated as essential "to ensure the cohesion of the team". The evaluation provdes the following comments on the audio and the video channels:
Each participant agreed that the audio link was more profitable than the visual link to solve problems. Each would have liked to have it at each meeting because it allowed flexible and quick interactions which made the understanding of current events easier. On the other hand, the possibility to see oneself was useful at the beginning, then it became superfluous and even boring. The participants would have liked to better control the content of their screen, for example to suppress and minimize the image of the other participants so as to have more space to do things, such as to discuss a text or a graphic displayed on the screen.
To summarise, multimedia-based communication technologies are adequate (more or less, according to the different types of technologies), for the following activities, if the members of a team already know each other:
· to exchange information;
· to disseminate reports;
· to keep up-to-date;
· to build productive momentum between face-to-face meetings and to extend the interaction of the group that cannot meet face-to-face, usually for financial and logistic reasons;
· to keep the link between a central unit (a lab or the headquarters of a company) and its sub-units;
· to create a memory of the meetings (by recording them);
· for a technical discussion (about data or objects which can be visualized).
On the other hand, they are not adequate for:
· problem solving;
· for making and taking complex decisions;
· for establishing relationships;
· for a first meeting of a team.
To assess or/and choose one of the multimedia-based communication technologies, one has to take into account the following aspects:
· the logistics involved;
· how easy or complex it is for the participants to use the technology;
· the management of the interactive process;
· the reliability of the communication channel;
· its availability;
· the cost for a session.
A summary of how each of the different technologies behave (according to the different cases already discussed), appears in Table 6.1:
Table 6.1 Comparisons of multimedia communication technologies
indicated in the examples in Section 6.3
Compressed Comp. video Mbone Mbone Desktop Desktop TCP/IP
Video Multipoint room-based desktop via ISDN via LAN (Internet)
1. medium heavy medium light medium light ?
2. easy easy complex complex easy easy ?
3. medium complex complex easy medium medium ?
4. reliable reliable unreliable unreliable reliable not 100% reliable
if Internet if Internet reliable
saturated saturated
5. requires requires available very requires very experi-
special special if Internet available special available mental
line line link if Internet line if LAN
6. expensive expensive free free expensive free ?
To draw a general picture of researchers using multimedia-based communicaton technologies, it seems possible to say that they usually want a flexible and spontaneous environment to allow them to interact whenever they have the need. Collaborative scientific research cannot always be planned many weeks ahead. It is often a spontaneous process. Researchers are not ready to put up with a poor and unreliable audio quality, but they may make compromises on the video quality, especially for peoples' images. Image is important for the visualisation of data and to work collaboratively on data, programs, etc. Researchers also look for easy-to-use technology (although the most technically experienced people are ready to spend time setting up their environments, providing there are other advantages). As far as the communication channel is concerned, researchers go for the solution which does not require the installation of special lines. Financially speaking they prefer the freely available Internet pTCP/IP protocol solution, especially in Europe where the telecom market is not yet deregulated, but they need it to be more reliable. They are not ready to abandon face-to-face interactions but they see multimedia-communication technologies as a way to meet more frequently and at less cost by saving on travel.
Through the different accounts of video-conference usage, it is noticable that two issues or arguments occur frequently:
· the question of interactivity, the social dimension and formality;
· the respective importance of audio and video.
For many users, even the ones who have used video conferences for a long time, a social dimension is lacking with multimedia-based communication technologies, especially with the compressed-video solution. It seems too rigid and formal, making it difficult to create the collective dimension of presential interactions, or to use Shrage's (1990) expression, a "shared space". It is a false belief to think that the image of the people makes it up for the lack of informality. Lacoste (1992) in her analysis, mentions various aspects of a video conference which give some flesh to this lack of social dimension:
"Taking turns to speak is not natural in a video conference. There is almost no overlap of interactions, not like in face-to-face interactions. On the contrary, they are carefully juxtaposed. This creates the rigidity which many users have underlined. In addition, when one contributes in video conference, one does not know where and who to look at, as well as who looks at him/her and how. In any face-to-face meeting, there is always a feedback, be it verbal, kinetic, audio, etc. The seven factors* involved in face-to-face interactions are, according to Duncan and Fiske (1977), either limited or absent in video conference, reducing the channels available for interaction."
Other rules are at play in video-conference interactions, rules which are still not yet very explicit and thus not understood and accepted by the participants. The fact that some people think new users need to be briefed and slightly trained behaviourly indicates the unusualness of the communication created by multimedia-based communication technologies:
"Video does not provide the same visual and aural information that face to face communication provides: information availability changes and there are differences in the salience of visual cues. . . It is not simply that video interaction conveys less information than face-to-face (although that may be true). But rather, people who interact only via video apparently use different information in forming impressions of their peers than do people who interact face-to-face. They apparently rely less upon task competence information and more upon communication competence information."
(Storck and Sproull, 1995, p. 5 & 12)
Concerning the lack of social dimension, the same authors mention the interpersonal distance which in a video environment either remains relatively constant or becomes uncomfortably too close or too far as a result of zooming the camera in and out. In face-to-face meetings, this interpersonal distance is easily and mutually adjusted. As a result of the increased formality which the media generates, they underline that the undertaken tasks through the medium are less commiting socially speaking.
"The increased formality leads to greater focus in tasks that are restricted to information exchange, i.e. tasks which do not involve decision-making or problem-solving."
(Storck & Sproull, 1995, p. 6)
The interlocutory silences are also more frequent in a video conference than in face-to-face interactions. Apparently, for Lacoste (19??), the flow of speakers is not as natural, there is some ambiguity as to the designation of the next speaker.
The last aspect mentioned by Lacoste is the presence of many more little verbal contributions of the type ". . yes . ., . . no . , . . well . . ., . . . ok . . . It is as if these rhetorical marks try to compensate for the lack of the usual feedback which in face-to-face gatherings is usually ensured by non-verbal cues. It is interesting to underline that one of the informants in the CMS case in his recommendations to new video conference users, suggests to avoid these little verbal contributions, as they disturb the overwhole meeting by overcharging the bandwidth.
The visual dimension in multimedia-based communication technologies is apparently not the most important feature although it is used over and over to sell this media. According to some, the image is more of a burden, "costing too much bandwidth" to be worth it. There may be a technical explanation: indeed, with multimedia-based communication technologies, the quality of peoples' images are still too jerky to be a good conveyor of the rich visual cues which exist in face-to-face meetings. The media does not reveal what people expect, so people prefer not to have it.
Many agree however, that the image is relevant in case of an information-driven video conference or at the beginning of a team collaboration, when the people do not know each other. For a collaborative session, people are less enthusiastic, real-time video does not add much, even more so when people already know each other. One has to distinguish between two different visual objects: the image of the people and the image of texts, objects, etc.
Although the visual component is not important to visualise people, it is however important to show the objects and the documents which are being discuseed. People using desktop video conferencing often reduce the window with the interlocutors to a very small size, giving more space to the collaborative activties. After a while, only the audio and the computerized activities prevail.
Finally about seeing oneself, this option is disturbing to many people; thus, many systems offer the choice not to have it. If the visual element is less attractive than one might have thought, everyone seems to agree that the audio must be excellent in order to follow a meeting properly. Whereas the image can survive a bad connection, the audio does not:
"You can do with a bad image of the people, but not with a bad sound".
In conclusion and to come back to research networks, an editorial in the Economist (1994), puts computer-mediated collaborative within teams of scientists into question, arguing that the quality of scientific work is at stake with such media (he is not only talking about video conferencing). Indeed for him, conferencing promotes agreements and homogeneity which are
"no friend to scientific creativity. . . .As faster exchanges of data make it possible to construct dispersed 'collaboratories'. . . the participants will need to put a lot more effort into finding healthy ways to disagree."
Not only scientific creativity is at stake but also scientific rigour. Indeed, a consequence of larger groups is that more and more scientists, because of overspecialization, ignore the origin of their data. It could be a dead-end of a consensus built on partially understood data. This argument is important and is a reminder that technology needs to be put into perspective rather than being blindly applied.
This section describe some characteristics of different styles of groupware and provide pointers to other more detailed descriptions elsewhere. It is not intended to be a recommendation or rejection any of the following approaches for particular schemes of group working. The applications mentioned here are included only as common examples of the different styles. Selection of software for group working (groupware) is obviously dependent on the common computing environment available, on the types of group tasks to be undertaken and on personal preferences.
The following sub-sections outline software approaches for communication (Section 6.4.1), resource sharing (Section 6.4.2) and other special group activities (Section 6.4.3) based largely on a taxonomy after Heeren (1996). For a discussion of standards issues and possible roles for various styles of groupware, particularly with regard to education and training see Collins and Pratt, 19??.
More detailed information is available on the World Wide Web, for example at:
Team-IT, a UK Government sponsored forum for Computer Supported Collaborative Working: http://www.csc.liv.ac.uk/users/team-it
CSCW web pages, http://www.consensus.com/groupware/
Tools for Open Electronic Networking, http://www.mtanet.co.uk/mta_oen/oentools.htm
The medium of communication ranges from plain text through audio to moving pictures with sound. Synchronous communication means simultaneous participation is required, therefore it can intrude into the normal work routine or must be scheduled to suit everyone concerned (like face-to-face meetings). Unscheduled speculative communications are also intrusive (like telephone calls). Synchronous groupware may be more suited to making decisions and be more likely to produce the feeling of belonging to a group than does asynchronous communication.
Internet Relay Chat - IRC- is a means of textual communication where people from all around the Internet can join in with real time 'conversations'. IRC is a multi-user system, where people convene on "channels" (virtual places, associated with different topics of conversation) to converse in public or private groups. Each channel has one or more operators who exercise control. Users have personal 'handles' (names) analogous to those used in Citizens Band radio which are used to prefix statements displayed on client displays to indicate who said what.
Connection to IRC servers requires special IRC client software. For more information, see:
http://www.yahoo.com/text/Computers_and_Internet/Internet/Chatting/IRC/
http://lcweb.loc.gov/global/internet/irc.html
A MUD (Multiple User Dimension / Dungeon / Dialogue, see also Section 6.5) is a virtual environment which many people can connect to and explore simultaneously. Each user takes control of a virtual character (or Avatar) which has a name, often different to its owner's real name and which can move, chat and solve puzzles.
Interaction is principally textual. Users control the software with pseudo-natural language syntax based on commands (for example, look, say) typed in response to program-generated information and prompts.
A MOO is a MUD based on an Object Oriented database such that all the virtual people, objects and places in the MOO have alterable properties that can retain their state between sessions. Users can sometimes create new locations and objects that become part of the shared virtual environment.
Connect to MOOs using Telnet or specialised Telnet-based client software, possibly integrated with a Web browser. In a MOO 'players' interact in real time principally through typed 'chat'. There are elements of asynchronous interaction since shared interactions can be recorded and a player's state is retained between sessions.
http://minerva.cis.yale.edu/~consult/mhelp/realtime/mud.html
http://www.dataphone.se/~doodles/mud.html
http://tecfa.unige.ch/tecfamoo.html
http://biomoo.weizmann.ac.il:8888
Virtual Reality or 3-D conferencing can be thought of as a graphical version
of MUD. Examples include:
Worlds Chat, Alpha World (Worlds Net)
Worlds Away (Compuserve)
MSN Vchat
See :http://www.superscape.com/
Audio conferencing can involve point-to-point connections like a normal telephone conversation but via the Internet or session recording, and multicast playback in which audio from a speaker is broadcast to the listeners. Example include:
Real Audio
Stadium
Direct TalkMail http://www.hursley.ibm.com/dtmail
http://http2.brunel.ac.uk:8080/multicast/vat_nv_record.html
Video conferencing (see also Sections 6.2 and 6.3) can be between a small number of (normally two to six) groups of people in expensively equipped rooms connected at comparatively high speed. Conferencing is also possible between standard multimedia personal computers enhanced with digital vision and sound input (camera and microphone) and normally connected at lower speeds. This personal form of video conferencing is sometimes called 'desktop conferencing' and the software may provide complimentary communication tools such as application sharing and shared workspaces (whiteboards).
A 'video' conference implies that the speaker at least is visible to the listeners ( show-and-tell'), sometimes listeners are passive participants in that they are not visible to each other or to the speaker.
Currently video conferences over the Internet may suffer lack of bandwidth causing jerky displays and lost or unsynchronised audio. Connections by MBone over the Internet also suffer when the Internet is congested (it usually is). Networking developments may increase bandwidth sufficiently to alleviate these problems in the future. Examples include:
CU-SEEME, ftp://cu-seeme.cornell.edu/pub/cu-seeme
Proshare, http://www.intel.com/comm-net/proshare/index.html
Fujitsu ICL TeamVision +44(0)1344-472000
PictureTel +44(0)1753-637000
Focus PC, GPT +44(0)800-622111
Typically users are spatially and temporarily remote. Messages in some systems must consist solely of plain text though increasingly systems permit messages to consist of formatted text with various typefaces and so on. Attachment of computer files (document files, sound samples etc.) to messages is usually possible and certain media types (graphics particularly) may appear directly in the body of the message.
There is much overlap of features in this area, for example automatic filtering of messages is sometimes available in mail and conference systems.
An asynchronous style of communication gives users more time for considered reactions and usually provides potent ways to review earlier interactions (for example indexed text searches and automatic organisation of contributions). In a busy environment not all users will have time to learn how to review past material efficiently.
Electronic mail consists of a central mail server or host that receives and accumulates messages for those individuals who use the central server until they connect to it and process those messages in some way. 'Processing' may involve moving or copying a message to a client computer (the user's personal computer), storing the message on the mail server in some arbitrary way or deleting it. The server holds messages created for onward delivery to local users, sent on to other mail servers or to gateways to other mail systems. Messages can be plain text and in modern systems comprise 'rich' text with various typefaces and emphasis. Graphics, sound and other types of file can usually be included in messages though their usefulness to the recipient may depend on the recipient having appropriate software available to handle them.
Electronic mail requires individually managed storage of messages and address books, so communication between large groups tends to impose a heavy burden on individuals. This is because the mechanism for distribution and reference of earlier material may not be universal to all users in the group and everyone has to develop and a system for managing the communi-cation themselves. Messages from all sources and on all topics normally arrive in one place, which can make organising messages consistently a difficult and time-consuming task.
Electronic mail can impose a system burden too, particularly as it typically involves storage of a separate copy of each message for each recipient, even when all recipients use a single central mail system. More efficient systems are becoming available in which only one local copy of each message is stored, no matter how many recipients there are.
Electronic mail is so popular already that users are quite likely to resist suggestions to change to a different mail application without very convincing reasons.
'Gateways' that allow messages to be sent between users of different electronic mail applications are not always part of the standard package though at least they are usually available. Interaction at the conference level (a common shared conference) between users of different types of computer conference system is only rarely available.
Most modern mail systems are designed with consideration for those users without a permanent connection to their mail host computer, so that users can work on their mail without actually being connected. Effective built-in, off-line working can be a valuable feature of asynchronous communication because it reduces the cost of connection. It also hides the complexities of file transfer that were so often a major barrier to potent, off-line use of older mail systems. For more information:
cc:Mail, http://www.lotus.com
Microsoft Mail, Microsoft Exchange (soon) http://www.microsoft.com
X.400 http://www.isi.edu:80/in-notes/rtr/rtr6.txt
http://www.isi.edu:80/in-notes/rtr/rtr10.txt
Mailing lists aim to assist group communication by electronic mail by maintaining a central address list and sometimes a central message archive. This means individuals no longer need to know and maintain all other participants mail addresses or store their own copy of all messages.
All interactions are by mail but some of the organisational tasks occur centrally in software though sometimes with human assistance.
Mailing lists are controlled by sending a limited number of commands by electronic mail. The syntax of the command language varies between servers, users may also be required to use more than one address for each list (often one for commands and another for communication). Users are not always 100% successful in their use of mail server commands and this often leads to a frustration on the part of other participants. Examples include:
MajorDomo
Mailbase, http://www.mailbase.ac.uk/
ListProc
ListServ http://sjcpl.lib.in.us/homepage/Reference/Listservs.html
Conferences are characterised by central storage and over time, the automatic organisation of messages from a number of users, for example by topic, time of creation and author's name. Besides assuming organisation of the messages, conference applications also provide tools for enhanced access (e.g., indexed searches). They can also usually restrict access by those not eligible and sometimes provide information about other users use of the system (e.g., who has read a particular message). In contrast with electronic mail, messages from all participants are stored centrally (possibly locally too) and organised by the conference system forming a record of the discussion that looks the same for everyone. Conferences typically store messages an a limited range of topics, reducing the confusion inherent in electronic mail where messages on all topics arrive and must be manually categorised either by each recipient or automatically using rules devised by each recipient. Modern conference systems frequently require special (client) software though some can be used with general purpose telnet or Web clients.
Lotus Notes, http://www.lotus.com/
FirstClass, http://www.softarc.com/
Usenet NEWS
Commercial Services: CompuServe, America Online, Microsoft Network, CIX
Screen-sharing applications permit a user to see and possibly control the screen of another user. This can be particularly useful to provide instruction.
Shared workspaces / whiteboards allow interactive use of common data (documents, files), often used in conjunction with a voice link or typed communication via the network link. For an example, see:
Proshare Premier (http://www.intel.com/comm-net/index.html)
Joint-editing programs allow co-authoring without the need to explicitly exchange files, possibly also with an audit trail of changes made and with the possibility of annotations separate and distinct from the document content. Functionalities are available for consolidating revisions, annotating changes (often with different colours), and routing documents. Examples can be found within:
WordPro
Microsoft Word 7.0
File transfer involves exchanging files such as word-processed documents over computer networks or by telephone using special purpose file transfer software. Common difficulties include the technical procedure of connecting computers together, requiring permission and passwords to access a remote computer, corruption of files by the transfer process and file-format incompatibilities between the sending and receiving systems. Examples include:
File format translators (Inso Corporation -
phone +44(0)181-9471122, Fax +44(0)181-9471810, email saleschi@inso.com).
FTP
Laplink
Kermit
Examples include:
Oracle
Lotus Notes
Microsoft SQL Server
The Knowledge Tree (http://ibis.nott.ac.uk/software/kt-about.html), an on-line question-asking-and-answering system that combines elements of a threaded Usenet style bulletin board with database functionality.
Examples include:
Group Decision Support System
Ventana Group Systems, phone +44(0)1494-680650, fax +44(0)1494-677840,
email 100334,103@compuserve.com
Appointment Scheduling
Microsoft Schedule Plus
Lotus Organizer
Project management software helps focus attention on the original objectives and the relationships between different strands of a project.
Workflow is a process in which information, often a document, is created and modified by users. Information is routed under software control (according to the purpose of the process) between users to be studied, altered or approved. Users receive their work in a software in-tray or mailbox. Workflow of this type is an example of rules-based messaging. Rules are programmed into an application and determine the path of the information, for example, which documents are sent to whom and what changes to the information are allowed at various stages in the process. Workflow often involves automated links between various applications. Examples include:
Lotus Notes, http://www.lotus.com/
Staffware, staffware@cix.compulink.co.uk
Linkworks, http://www.digital.com/info/linkworks/
http://www.columbus.ues.com/