AIDS Conference Information on the Internet -- How to Present It

Communicating AIDS information by multimedia and computer is quickly becoming a current concern. For example, the Retroviruses conference later this month in Washington, responding to widespread anger that many hundreds of people have been excluded, is now planning to provide video of some of the major talks on the Internet. And the next International Conference on AIDS, in Geneva, plans "to make unprecedented use of computers during the conference"; although that meeting is a year and a half away, planning is happening now.

How can computer communication be most useful? This writer developed computer software in an earlier career, and has long paid attention to what would be needed and useful in AIDS. This article looks at how the information presented in large public conferences could be effectively communicated through the Internet. We have no proprietary interest in the design sketch below, and encourage others to borrow concepts. We have no time to write the software ourselves.

The Big Picture

Today many large conferences bring together hundreds or thousands of people, who often travel great distances. It is widely agreed that at these conferences, chance meetings in the hallways, or other informal private meetings, may be at least as valuable as what is presented in the formal public sessions.

Computer communication can deliver current, in-depth information to those who cannot attend physically, and it can have important advantages over hearing the same material in the lecture hall. Also, it can improve (not replace) the physical meetings, by delivering formal information in advance and shifting the emphasis toward well-informed, private, less formal working sessions at the conferences, to make the best possible use of peoples' valuable time together.

This paper outlines a software design for distributing the FORMAL content of AIDS conferences. We will address informal communication later.

What Happens at AIDS Conferences?

In all the AIDS conferences we have attended in ten years, only three communication formats have been significant: (1) Speech only, either talks by individuals who did not prepare slides, or panel discussions or question/answer sessions, where interaction is unpredictable so it is seldom practical to prepare slides or other visuals in advance; (2) Slide (or overhead) presentations, and (3) Posters. Other formats (such as showing videotapes to an audience, or projecting the contents of a computer screen) have been rare so far.

For speech only (the "talking head"), a transcript alone can be even more useful than being in the room, because reading is faster than talking, because a transcript allows review and scanning forward or backward to find information of interest -- and because it is less graceful to walk out on a speaker than to just start doing something else. Audio and video should be available as options, however -- for people who receive information better by ear than by eye, or for when tone of voice is important, or when the atmosphere of the meeting is significant.

Slide (or overhead projector) talks have presented the biggest problem for conveying information to those not in the room, because most conferences only sell audio tapes, not video, and much of the talk is likely to be lost if the slides are not available. (Some talks use text-only slides which repeat the speaker, or outline main points. These can be quite helpful to the live audience, since information is best conveyed by ear and eye together; but the transcript reader has less need for them. The slides that are vitally important are photographs, charts, graphs, etc., which present additional data not present in the text of the speaker's talk.)

Posters are text and graphics. They can be presented on the Internet with the same software used for talks with slides. This means that the slide presentation is the key challenge for distributing formal conference presentations by computer; text-only talks, and posters, can be handled as special cases of the talk-with-slides format.

Software for Slide Presentations

Video feed does work well for slide presentations -- especially if the camera crew pays attention and shows the slide or a close-up when appropriate, instead of the talking head. We and others have found that it is often possible to get a better view of speakers, panels, and slides from a video monitor set up in the press room or elsewhere in the conference facility, than from being in the session itself, where the seats available may be too far from the screen to read the slides. Another advantage of watching at a monitor is that it is easier to have side meetings or conversations -- the important informal interaction at the conference.

But simply placing video on the Internet is far from ideal. First, transmission usually takes a long time (although file compression software greatly helps).

Even with the "information highway" of the future, when most offices and homes are expected to have high-bandwidth computer communication, straight video feed will still have disadvantages, because of the difficulty in scanning or searching it to find points of interest. (Audio tape is even worse -- as illustrated by an incident in the controversy over the Nixon White House tapes, when the White House staff, after having made years of tape recordings, considered listening to them and realized that it would take years to do so.)

Here is our proposal for software to communicate a slide presentation (or talk only, or poster board) on the Internet. We believe it would be relatively easy to implement.

It seems best not to transmit in real time, but instead to transmit a multimedia file for each lecture or other presentation; the reader/listener (the person viewing or listening to the presentation) can browse the file after it has been transmitted, using the software described. (As mentioned below, there can also be real-time transmission for a "preview" mode, to help reader/listeners decide what presentations they want to view.) This multimedia file will be prepared by editors, who will work from a videotape of the presentation, a copy of the slides, and any prepared transcript which the presenter happens to have available. Therefore, there is no additional burden on the lecturer (usually a busy physician or scientist) who wants to use this format. They do not need to be computer-oriented or to use any equipment or software.

Basic Concept, the Four Windows: Text, Slides, Audio, Video

The software, running on the reader/listener's computer, which does not necessarily need to be online, will provide four viewing channels -- text, slides, audio, and video. All of these channels will be optional; the reader/listener can select which ones to use. Each channel will appear on a different window on the screen; the reader/listener can re-size or move the four windows as desired, or close any for channels which are not used for a presentation, or which the information receiver does not care to view.

* The text window will show a transcript of the talk.

* The slides window will show any slides used, or other still graphics which are part of the presentation. When the speaker uses a pointer to call attention to one or more parts of a slide, a border such as a thin bright line could be used to mark the areas.

* The audio window will control an audio recording of the talk.

* The video window will usually just have the "talking head," since the slides or other graphics are available separately in the slides window. In practice, the video window will usually be small -- perhaps a couple inches square -- to reduce the file size and transmission time, and also because the video channel is now the least important of the four channels, since it does not carry additional content, but mainly is used to give a sense of the atmosphere of the room.

The four channels must be synchronized. This means, for example, that a reader/listener can quickly browse through the slides -- using the down or up arrow keys to immediately show slides sequentially forward or backward, or using a scroll bar in the slide window to quickly move to any part of the slide sequence -- and then immediately be seeing the corresponding text and video in their respective windows, and hearing the speaker's voice describe the slide selected. Similarly, the reader/listener can jump backward or forward, a little or much, in any of the windows, and the other three windows will immediately be synchronized with the reader/listener's selection. And as the speaker talks, the slides will change automatically, just as they were changed in the original lecture (in the auditorium -- or in front of a video camera in the speaker's office or in a studio, if a lecture is to be provided in advance of the physical conference).

Similarly, a reader/listener could do a computer search for one or more keywords in the text window. Once a hit was found, all four windows would instantly synchronize to that point in the talk, with the speaker's voice saying the words, and the slide which was displayed in the auditorium at that time being shown.

Note that the basic four-windows organization is easy to learn. The user just selects a presentation to see, and automatically it starts going on the computer, with the windows set to reasonable defaults to show the speaker's voice, image, transcripts, and slides. A line at the bottom will tell which keys are active (mainly the up and down arrows to browse through slides or text); users can play with these keys, or with the mouse and scroll bars, and all the windows will immediately synchronize on whatever part of the presentation is selected. An optional control bar will be able to speed up or slow down everything, including the voice and the video.

Optional Search Features

An additional search feature could be a glossary to assist the reader/listener in doing a computer search. The basic glossary would include AIDS-related terminology -- and it would automatically enhance itself temporarily with a scan of the transcript of the talk being viewed. Then, if the reader/listener tries to search for one or more words which were not used at all by the speaker, or otherwise tries a poor search strategy, the text window will show suggested synonyms for each word, allowing the user to enhance the search by accepting some or all of the words suggested by the glossary software.

And of course the reader/listener can always ask for a search backward instead of forward from the current point in the lecture or presentation.

In some cases it will also be possible to search for text in the SLIDES, not just in the transcript. An example would be searching for a drug name used in labels on the slides.

The reader/listener should also be able to search not only within the current presentation, but in all the presentations in a conference -- or in all AIDS presentations in all conferences -- or in certain sections of a conference. The reader/listener should also be able to create a personal RELEVANCE LIST, directing which talks, authors, sections of a conference, or entire conferences, should be searched first. Reader/listeners can create and save any number of such relevance lists -- and can share them with other users, or with the public. Leading virologists, immunologists, or other specialists could contribute relevance lists for public use.

Since relevance lists (and, in fact, all computer searching) are strictly optional, reader/listeners can learn and use the basic software -- the four windows: text, slides, audio, and video -- without ever knowing about these features.

Optional History File

Another optional feature is a history file, which keeps track of everywhere that a reader/listener has been in a session -- by recording the arrival point from every search or scroll, plus the length of time the presentation was allowed to play uninterrupted until the next search. A separate key or mouse click can be used at any time to mark a point of particular interest. With the history file, reader/listeners can quickly return to something they noticed, even if they did not recognize its importance at that time. They can scroll or search through a history file just like they scroll or search through a presentation. History files will automatically be saved forever, always ready for use, unless the reader/listener chooses to obliterate a session, to destroy all record of what they had selected and seen.

Notice that new reader/listeners do not need to know about the history file, but their history will be saved for them anyway, should they ever want to go back and view it.

Another use of history files is to prepare educational modules for the public or others. Teachers or other experts will produce history files to provide guided tours through selected parts of multiple AIDS presentations, to give in-depth looks at particular topics. Teachers can add explanatory material simply by creating their own "presentation" -- a string of explanatory segments one after the other. Then they can thread a history file through their explanations, just as they select segments of any other AIDS talks available.

Ultimately the history files could allow branches, giving the reader/listener control of different options within a presentation.

Note that history files could be transmitted very rapidly to sites which already had the multimedia files for the presentations used. If the receiving site does not have all of the presentation files, the missing segments will be transmitted with the history file.

Implementation Notes

* The video channel, which usually would take the longest to transmit, can be sent much faster because the slide channel eliminates the need for high resolution. The video, now useful only to give a feel for the atmosphere of the presentation, could be displayed in a small "thumbnail" window in a corner of the screen. (Or it could be viewed as a fuzzy picture in a larger window, perhaps appropriate for display to a large lecture-hall audience from an average-size computer screen.)

* Text and slides are usually the most important channels by far; audio and video give the speaker's voice and the appearance of the speaker and the room, but little additional technical content. Since the least important channels now take the most bandwidth, transmission could be speeded significantly by choosing not to receive them.

* Instead of real-time transmission, this system assumes that, for each presentation, a file is sent for viewing later; otherwise it would be impossible to scroll or search forward, for example. But users should also be able to preview a presentation, to decide if they want to request the full file. Occasionally users might wish to use the preview mode to obtain the whole presentation.

* Slides can be transmitted in one of two ways -- either as photographs, or as text and data which can be regenerated into outlines, graphs, or charts when displayed. Text and data can be sent much faster than the photographic image -- and it has the advantage of being searchable. But sometimes photographic transmission is necessary, so this system needs to handle both cases.

If this software includes any optional facilities for presenters to create their slides, it should know about color combinations that are easily distinguishable -- including by those with different kinds of color blindness, for example. Also, presenters might be asked to estimate screen size and audience distance. Then a warning could be given before an illegible slide was produced.

* Synchronization of text and slides is easy, simply by having the text contain codes like '(slide #1)'. Synchronization of video and audio is less obvious.

* This system should probably be implemented in the Java programming language. Then talks could be viewed on most existing World Wide Web browsers, regardless of what computer platform was used, with little if any need for the reader/listener to obtain and install special software. The software and presentations would also be provided on disk or CD-ROM, for users without Internet access.

Summary

Talks, discussions, slide presentations, and posters could be viewed more effectively through multimedia than by being present in the meeting hall, or watching the presentation on videotape. With appropriate software, the same quality of presentation could be efficiently distributed by the Internet or otherwise, and could be available to those who cannot get to the conference and/or do not have Internet access. We propose a design for multimedia software based on four information windows: text, slides, audio, and video.

Distributing information by computer could also improve conferences, by shifting the focus away from large lectures where most people arrive, sit, and leave without speaking to anybody. Instead, they can get the lecture-type information in advance, and arrive prepared for focused working meetings with people they need to see.

Of course the Retroviruses conference should not wait for any new software, but should make video from the conference available through the Internet and otherwise, using the best technology now available. If you know of software to recommend, please contact this writer at AIDS TREATMENT NEWS.