Figure. 1 : Web-based asynchronous
lectures. (Fall98)
Note :
Click above screen to see the lecture.
LATCHMAN, Haniph, KIM, Jongmyon & TINGLING, Dave
Electrical and Computer Engineering Department, University of Florida, Gainesville, Florida 32611-6005, USA, Tel. (352) 392.49.50, Fax. (352) 392.00.44, H. Latchman : latchman@list.ufl.edu
Abstract: As more and more institutions begin to use the Internet to deliver online education to students who for various reasons cannot attend conventional time and space constrained classes, the question arises as to whether the online delivery methodology can deliver entire degree programs with a quality comparable to traditional on-campus classes. In this paper we describe ongoing work in the Electrical and Computer Engineering Department at the University of Florida as we offer an Online MS degree in Electrical Engineering and as we proceed with the development of major components of a BS degree in Electrical Engineering. In particular we describe the implementation of the "lectures on demand" method using multimedia streaming technologies with the now widely accepted Asynchronous Learning Network (ALN) model. We also describe the technical issues encountered in delivering real-time and asynchronous content via the Internet as well as the WWW and Instructional design issues encountered along the way. The presentation will also include live demonstrations of the "lectures on demand in ALN" methodology used in the MS and BS programs[1].
Keywords: Online Lectures, Online Degrees, Internet, Streaming Video, Distance Education
With such an outlook, there is ample motivation to apply technology itself to offer improvements to our traditional models of education. For example, who could question the power of synchronized moving pictures and sound and its ability to impress its audience? The intentional or unintentional application of this of technology to educate has had global repercussions, both positive and negative. Unquestionably, the use of technology to enhance education and learning has proven effective, but we must remain focussed on using technology as a tool to achieve well defined teaching and learning objectives.
Today, cyberage technologies permit interaction both real-time and asynchronously, between students and instructor. It is with such technologies that we have experimented, focusing on the exciting possibilities of asynchronous learning that might prove viable in academia, by utilizing emerging internet standards. We will herein describe the successes and challenges that we have experienced in developing BS and MS online Degrees using a Lectures on Demand approach.
The tremendous advantage of the proposed internet-based scheme for joining a class in real-time from remote locations is that the student does not require expensive equipment to enjoy meaningful classroom interaction, as is the case when using circuit switched video conferencing. A PC with a 28.8 kbps modem dialup link and a web browser brings the virtual classroom to the users at home or work. If the student is on campus with access to a Local Area Network (LAN) it is possible to get a higher bandwidth with even better performance. In addition, the system allows many students from various locations to join the synchronous lecture at any time.
With the model implemented at the University of Florida, live lectures are encoded digitally for broadcasting. At the same time, the audio and video streams are archived so that students can access the material from anywhere and at any time, and essentially view the material as it was broadcast live. In this case live interaction is replaced by asynchronous interactions via mailing lists, bulletin boards and WWW pages. In addition, we show that the asynchronous experience can be greatly enhanced by an incremental investment of time in synchronizing the lecture notes and other materials such as scanned handwritten documents of the instructor, student notes taken during the class, PowerPoint slides $$[], or even interactive Java applets and simulations. Effective synchronization causes the browser to automatically display appropriate lecture material in another window (or frame) as the lecturer discusses it.
Figures 1 and 2 show the various objects used to convey information and facilitate interaction among the students and with the instructor. The system is designed for use in real-time lectures as well as for asynchronous delivery. The key components are an audio video window, textual/graphic windows and links to chat rooms came from WebCT tools and mailing lists. In Figure 1 the PowerPoint lecture notes are synchronized in time with the activities in the video window so that the slides change automatically as the lecture progresses. The student maintains control of the lecture and can fast forward or reverse the video stream as well as scan the PowerPoint slides backwards and forwards at will. In addition the lecture is also classified into major topical areas (as shown in the MENU window in Figure 1) so that the student can go to a specific topic of interest.
In the approach which we adopted for developing the MS and BS programs, we aim to provide a low latency, low bit rate, video and audio stream of the instructor lecturing 'live' in the classroom, with an on-campus contingent of students. The audio quality should be at least "toll" quality and the video quality should be good enough for the written material on the board or the slides from overhead projector to be easily read by remote students.
It is of interest to note that including the traditional lecture as a key component of the online course content is, in our opinion, a good first step in using technology to augment traditional learning systems which have worked well for many generations. Many institutions (some well-reputed) provide online courses which consist mainly of a calendar of activities and readings or online textual and graphical material which students must complete, working typically on their own. In our approach, the student has available all those facilities in addition to the lecture discussion.
An alternative approach has been
to make audiovisual recordings of an instructor lecturing to an empty classroom,
with no students and no interactivity. While this approach clearly has
some merits in respect particularly of being able to complete more material
in a given time, it is nonetheless not adopted here for several reasons.
Our fundamental principle is that traditional classroom learning works
well and will be a major feature of all traditional learning institutions
into the foreseeable future. In addition, since the classroom learning
models which work best are those which feature collaborative, active learning
with meaningful interaction between students and with the instructor, we
believe that the interactive questions of the 'studio audience' mimic this
setting quite well. Indeed, many of the questions asked by the students
attending the lecture as it is being recorded are also questions which
may have spontaneously been asked by an asynchronous student if present.
Of course, the ALN mechanisms are still provided to allow email questions
at any time, even during lectures.
Figure. 1 : Web-based asynchronous
lectures. (Fall98)
Note :
Click above screen to see the lecture.
In what follows we describe the specifications for each of the major components in the Lectures on Demand interface.
The target data rate for the audio
and video stream is roughly about 20 kbps, although as we shall discuss
later, newer technologies are able to provide an adaptation to higher data
rates if end-to-end bandwidth is available. The 20 kbps target is a stringent
requirement considering the fact that this bandwidth will support useful
audio and video signals. However a 20 kbps stream on a typical 28.8-33.6
kbps link will leave the student with enough bandwidth for other communications
(such as lecture notes, WWW pages and feedback) in addition to the audio/video
streams. Of the two streams the audio will have to be at least 8 kbps for
the required level of quality with conventional codecs and thus the video
occupies only about 12 kbps. Since the audio and video streams are independently
generated, there should be a tight level of synchronization between the
two. If the latency (at source while compressing) is different for the
two streams, the audio and video will eventually fall out of sync during
the course of a lecture. While the information content of a talking head
is minimal, the fact that spontaneously written material is also transmitted
in the video window justifies the inclusion of the video facility. An alternative
for asynchronous consumption is to provide a choice of combined audio/video
streams at various rates as well as an audio only source so that the user
can select the stream most appropriate for his network connection [10].
Figure 3 below shows examples for developing stage for synchronized lectures
at University of Florida.
Figure. 2 : Web-based asynchronous
lectures by SMIL. (Spring 99)
Note :
Click above screen to see the lecture.
Examples for Developing stage for synchronized lectures.
|
|
|
|
 |
|
|
For real-time lectures the generated audio and video streams from the classroom are to be broadcast over the Internet to students synchronous in time but asynchronous in space. Since the IP network is the delivery medium, the content delivered suffers from the inherent disadvantages of the network, namely, variable and unpredictable end to end latencies and the lack of any guaranteed quality of service. To overcome this inadequacy some mechanism to maintain the live or real-time quality of the streams at all times should be employed. A primary objective of the audio/video stream is to provide the lecturer with some degree of spontaneity in lecturing style via a high quality audio and a reasonable quality video usable by the remote student. To compensate for the very small video image of the materials written in real-time on the board or on overheads, an auxiliary web camera or electronic whiteboard may also be used to periodically capture the written materials as a www image and transmit this information to the first data window. In archived lectures, one data window can be used to display individual high quality slides synchronized with the video window. Another data window is provided to allow flexibility in displaying other relevant materials such as interactive simulations.
An online WWW-based chat window is also included as shown in Figure 1.
This facility allows students at remote locations to interact with the
professor during live classes, for example by submitting questions or images
or even audio/video clips which can then be displayed or played to the
class and an appropriate response given. Moreover when a student is using
the asynchronous mode of access, the chat facility or mailing list link
can be used to submit questions or comments to the entire class. The contributions
are archived so that the thread of the discussion can be followed at a
later time. Figure 4 below presents demonstration of synchronized
lectures at UF.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In the scope of this discussion, when we refer to streaming media, we are referring to RealMedia [5] clips created by RealProducer, which are continuous video and audio presentations. The video and audio components are also known separately as Real Video and Real Audio, respectively. Before the advent of Real Media, one had to wait for a media file to finish downloading to a local hard drive it could be played. With a streaming clip (Real Media) the internet user can play it almost instantly. Furthermore, live media can be streamed, so that content is sent directly to the audience, in real-time.
When a clip is streamed, small packets of information are sent over the internet connection. At the other end, the user receives the packets and plays the media clip bit by bit. To the audience, this process is transparent, except for a small delay at the beginning of the clip as the packets are initially buffered.
When content is encoded using RealProducer, the ideal scenario is to achieve a total bit rate less than or equal to 20 kbps for each stream. If the SureStream feature of the software is selected (recommended), it should be specified that the target audience employs a 28.8 modem. With these settings, the user experiences a total buffer time of no more than 15 seconds. Under ideal network conditions, the buffer time is less than 10 seconds.
The video thus generated conforms to either a 4:3 aspect ratio, (for example, 320 x 240 pixels) or a 16:9 aspect ratio in the case of letter size video (for example, 192 x 112 pixels), such that each pixel dimension is a multiple of 16. This is adequate for our purposes, so 20 kbps is accepted as our general target data rate.
SureStream encoding should be used wherever possible, (generally for live broadcasts). Standard target options are available so that the stream can be optimized for specific types of target audiences. Media playback to the type of audience specified should be tested from different access points (or ISPs) at different times of day, and using different CPUs.
A positive feature of RealProducer is its SureStream technology. SureStream dynamically reduces the bandwidth of the stream to adjust to prevailing network conditions, so that a smooth and optimal stream is delivered at all times. The technology allows Real Audio files to scale dynamically from 14.4kbps to 56kbps rates. When SureStream is applied in production of live broadcasts, a substantial decrease in cost and complexity is achieved.
The equipment that has been used to successfully produce on-line course materals in the LIST (Laboratory for Information Systems and Telecommunication) at UF consists of :
|
|
|
|
| CPU | Pentium® 166 MHz with MMX | Pentium® 200 MHz with MMX |
| RAM | 32 MB | 64 MB |
| Hard Disk Space | 1 GB | 1 GB |
| Color Display | 16-bit | 24-bit (True Color) |
| Video Capture Card | Any native Video for Windows® capable capture card | |
| Sound Card | 16-bit sound card or better | 16-bit sound card or better |
|
|
|
|
| CPU | Pentium® 200 MHz with MMX | Pentium II® 400 MHz |
| RAM | 32 MB | 64 MB |
| Hard Disk Space | 1 GB | 1 GB |
| Color Display | 16-bit | 24-bit (True Color) |
| Video Capture Card | Any native Video for Windows® capable capture card | |
| Sound Card | 16-bit sound card or better | 16-bit sound card or better |
As discussed earlier, our emphasis in developing the virtual classroom is to effectively deliver audio and video content from the actual classroom. To accomplish this, a commercial software suite, the Real Audio and RealVideo system from RealNetworks [5] is used. The live delivery of audio and video using the RealNetworks tools roughly follows standard client server architecture (See 3.6, Architecture Overview). A schematic diagram of the system deployed at the University of Florida is depicted in Figure 6. Audio and Video digitizing is done in the classroom. Existing audiovisual equipment at the Florida Engineering Education Delivery System (FEEDS) studio provides audio and video signals for an existing in class TV network and records the lecture on conventional VCRs. For real-time Internet delivery, a video digitizer card and a sound card are installed on a 200MHz-pentium-machine running Windows 95. The digitizer is capable of real-time digitization at up to 30 video frames per second. The sound card is an industry standard16 bit card and the PC hosting the real audio encoding software and video digitizer is connected to a 100 BaseT LAN. At the beginning of each class the RealEncoder is used to make a connection to the RealServer located in another floor of the same building through the LAN. The audio and video is digitized and streamed in real-time at 20 kbps to the server over the LAN. The frame rate of the video, audio bandwidth and the total bandwidth is selected in such a way as to maintain a high level of synchronization between the streams and low overall latency. Bandwidth and latency considerations are described in later sections. At the server these streams are delivered upon request from web browsers, with the current installation supporting up to 60 simultaneous streams.
In an earlier UF implementation, a single machine hosted the RealServer as well as the WWW servers. However, as th more courses were offered and the client load increased, it became necessary to have several independent servers in order to cope reliably with the demand.
Figure 6. Audio and Video delivery.
The content delivery from the server to an end user client is as shown in figure 7.
Real-time interaction or feedback from the remote students is achieved via electronic conferencing software such as Web based Bulletin Board System (BBS) software and WebCT [14]. Using these forums, instructor as well as the students log on to the same chat room. The option to list all participants in a chat session is desirable so as to mimic the "virtual presence" in the chat room. In the case of archived lectures, BBS interaction is complimented by asynchronous interaction via mailing lists and WWW pages. All these resources can be accessed by any web browser from the course home page. Thus far, user feedback has shown that the asynchronous experience is greatly enhanced by the synchronized lecture notes -- pre-prepared PowerPoint slides, scanned hand written notes fromthe instructor, or notes taken by a student during the class.
Most of the remote students have a network connection from an ISP provider with a 28.8 kbps dialup link, although dial-up connections at 33.6 - 50 kbps are now becoming quite common. In the testbed a bandwidth of 8 kbps is used for voice and thus the ceiling of 20 kbps limits the video bandwidth to 12 kbps. Various compression techniques were tested to find the best tradeoff between bandwidth and quality of the video. Fractal compression is found to be very impressive above 56 kbps and higher. However, for live applications, fractal compression caused unacceptably high latency. An optimal frame rate scheme is presently being used for live lectures. In this scheme during a very active session characterized by fast moving images the frame rate increases and then decreases again during relatively inactive periods. This method provides the best quality video in the very limited bandwidth scenario. Initially a frame size of 240 X 180 was used and then changed to 160 X 120 because the former frame size produced a higher latency on a 28.8 kbps link. The effect was not noticeable over 10/100Mbps LANs. The color format used was originally RGB24 but this was changed to YUV9 because of the encoding delay and greater end-to-end latency. Over a 28.8kbps dialup link there was no noticeable packet drops. The maximum latency observed over a dialup link was less than 10 seconds. When observed from a LAN the latency was of the order of a couple of seconds. The actual frame rate over a LAN varied from 1 to 5 frames per second (fps) and over a dialup link the average frame rate is about 1/2 fps.
Some experimental data of the movie compression process is given below (Table 1).
The Lectures on Demand approach is being used to deliver complete MS degrees in Electrical and Computer Engineering (ECE) and in Computer and Information Sciences and Engineering (CISE). Some 15 courses are currently available online or are presently being developed. Students must satisfy all regular UF MS admission requirements and then they can take the online Lectures on Demand courses during the semester that the courses are being offered on campus. The online students have a well defined window (say 2 weeks) within which to complete segments of the course including assignments and reports. Students may also obtain certificates in certain specified areas such as in Networks or Database Systems or Operating Systems. In addition the Lectures on Demand approach is also being used to develop a Bachelor of Science in Electrical Engineering (upper division) and four courses are currently being put online.
In one of the ECE courses, Computer Communications (EEL5718), which is taken by first year graduates and graduating seniors, students were asked to attend the course synchronously in time but asynchronously in space both on and off-campus. In addition several students took the course entirely asynchronously and the results of the learning experience were satisfactory to most of the students. The on-campus students who watched the lectures in real time over the campus network, had a much better video quality due to large available bandwidth, while the off-campus students who followed the course over 28.8 kbps modems had a less perfect video quality but the audio stream was estimated by the users to be better than telephone quality. The learning process through the synchronized multimedia format was evaluated by all participants to be better than video tapes since notes and other auxiliary materials were displayed on the screen so more attention was paid to the lecturer. The entire set of lectures (about 40) of this course is now available on the Internet for virtually anyone around the world who wants to take the course. The streaming audio/video material each course occupies about 400 Mb of storage so that the complete set of course materials including, lecture notes, slides and assignments can be stored on a standard 600 MB CD. This CD could then be distributed to students taking the course and they would then have direct access to the streaming audio/video from the CD while using the Internet for accessing interactive facilities such as online chat, mailing lists and other frequently updated class materials.
Compared to the traditional way of teaching, this new method, which reaches students who otherwise would not have taken these courses, provides learning opportunities to a much larger and diversified audience, thanks to the absence of time and physical constraints. Students in different countries can now, without leaving their families or their companies, attend courses virtually anywhere in the world. In addition the fact that the archived lectures are available on the Internet as soon as they are taught while a video may take several days to reach the student constitutes a considerable advantage over common courier/mail delivery of videotapes since delays are minimized or eliminated entirely. In addition archived lectures allow students who are attending the course on-campus to view entire lectures or parts thereof as many times as necessary. In this aspect this method complements and enhances the traditional teaching. Another advantage of the Internet streaming video technology over videotapes is the protection against unlawful copying of the audio/video material. While there is no simple way to prevent videocassettes from being copied, the streaming technology prevents lectures from being downloaded. Furthermore by changing passwords we can make sure that the lectures are not being illegally accessed by unauthorized users over time. A great deal of educational and cognitive research has shown that the most effective learning environment involves interactive collaborative, "learning-by-doing" models. Clearly the traditional classroom setting facilitates spontaneous interaction among students and between students and instructors. However social preferences and learned habits often suppress these collaborative, active learning modes in most classrooms and many in-class learning experiences are simply one-way lectures. It takes a great deal of effort on the part of instructor and students to generate a collaborative learning environment in the classroom, in which students work in groups on problem-solving and share findings- discoveries - and insight with each other. This "ideal" in-class model may well be described as a "synchronous learning network" - a "network" of people learning from each other - with the instructor serving the role of facilitator and coordinator rather than the source of a one-way flow of canned information. Our efforts in developing the hybrid synchronous and asynchronous learning model seek to simultaneously enhance the learning experience of three (3) distinct groups, namely traditional on-campus students, students at a distance who join the class in real-time via the Internet and students who take the class completely asynchronously.
The first group is the in-class students who can make use of the online materials at anytime and from a location of choice to supplement and complement their in-class learning. The asynchronous interaction such as electronic conferencing, chat-rooms and mailing lists provides an alternative, and often, a more desirable forum for collaborative learning, since many students are more comfortable with contributing to a discussion on their own terms, with ample time for reflective responses, rather than to be put on the spot in the live classroom. It takes much skill and sensitivity on the part of an instructor to get a lively and informal in-class discussion going without causing embarrassment to some students. The asynchronous mechanisms such as mailing lists and electronic conferencing systems do help in this regard - though experience has also shown that the meaningful use of these facilities must be encouraged and even required initially, so that the real benefits of these interactions can be perceived as a means to overcome inhibitions of custom. An additional benefit for on-campus students is that they now have greater flexibility in scheduling, since they can take some classes entirely asynchronously or via a live on-line connection from a remote location. Moreover, when entire courses are available online, the options become almost infinite: self-paced course completion, access to pre-requisite classes, which are not offered by traditional classes in a given semester, as well as the ability to take a class when the traditional in-class sections are full. The list of potential benefits of the availability of high-quality asynchronous classes for even on-campus students is clearly significant and these open new possibilities and modes of learning.
The facilities described in this paper allow remote students anywhere on the Internet to join and actively participate in live on-line classes. These students would simply point their Web browsers to the URL for the live classroom lecture [1] to get a connection to the streaming video and audio with optional links to the chat-room and online class materials. In this mode the student can ask questions as the lecture is being given using the live chat window. The instructor either repeats the question to the class or can display the question on the class TV monitors, which is then sent via the streaming video to all participants. The investment in hardware and software on the part of the online student is simply a standard multimedia PC with at least a 28.8 kbps Internet connection. A full duplex sound card (or alternatively a second half-duplex card) would allow voice interaction as well as or instead of the live chat window. It takes some coordination effort to manage these interactions, such as the need for requesting permission to speak (say via the chat window) and being acknowledged by the instructor, being given permission to "speak" including access to the feedback audio system. However our experience is that for small groups of remote students the system works quite well, depending on Internet latencies. Of course, for live Internet classes, it does help to have the lecture materials online or sent a priori to the students, as well as frequent references during the lecture to tags or locations in the class materials to facilitate coordination. Our experience is that while the video window attempts to send what the in-class students see, some material is less intelligible than others. In particular, it appears to be quite difficult to find an optimal set of contrasts and colors in the video capture and encoding process, which deals uniformly well with computer screen projections and material written on the chalkboard. The use of multiway application-sharing programs such as NetMeeting provides a partial answer to this problem but further work is required in this area. A final point worth mentioning is that there is an unavoidable delay of several seconds between the live in-class lecture and the streaming video/audio at the remote student's location. However the delay is usually tolerable on non-congested Internet connections.
The "Virtual classroom" environment may offer significant support for education of the future. Complete asynchronous access to the "virtual classroom" environment described above allows access to course materials by students who could not otherwise take the desired courses. This realization offsets some of the real or perceived disadvantages of taking online courses via ALNs. On the other hand, the rich set of information sources embedded in the ALN design described in this paper provides an appealing environment with high quality slides and images of hand written notes synchronized with the streaming video and audio. The mailing list and electronic conferencing facility allow collaborative interaction with other class members as the lecture is being viewed and this captures the desirable benefits of a network or people - albeit asynchronous in time and space - learning together.
In this section we present examples of several lectures using encoding
rates ranging from 28.8 kbps to ISDN rates of 128 kbps. In general it is
observed that the audio quality is uniformly good for all data rates, but
the video quality improves somewhat at the higher data rates. Figure 8,
9, and 10 represent hyperlinked snapshots at 28.8kbps, 56 kbps
and 128 kbps using the G2 encoder. The G2 format includes a new technology
called SureStream which allows adaptation of the data stream to the available
bandwidth. In all cases clicking the images below will provide a link to
the RealServer at the given rate.
Note : You will need to have the
RealNetworks
G2 Player to view any of the example clips in this paper.
28.8kbps Encoding
56.6kbps
Encoding
Figure. 9: 56.6kbps Encoding
Note :
Click above screen to see the lecture.
Dual ISDN 128 kbps Encoding
The examples below are included to give a flavor for how different instructional content appears to the student at the remote site using the learning on demand approach. The courses are graduate courses in Computer Communication (EEL 5718) and Queueing Theory (EEL 6507). The last example in the table illustrates the use the SMIL protocol (Synchronized Multimedia Integrated Language) for providing synchronized streaming media in an integrated fashion using only the RealPlayer.
Figure 11 below
presents several clips encoded at various speeds from two courses
at the University of Florida - Computer Communications (EEL 5718) and Data
Communications and Queueing Theory (EEL6507). The final example included
is based on an interactive class taught from Florida State University (FSU)
in Tallahasse to a group of 50 students in a multimedia classroom
at the University of Florida in Gainesville Florida. This latter example
illustrates the capabilities of not just learning at anytime and
anyplace, but also teaching from any place as instructors also become mobile!
The connection between Tallahasse and Gainesville was via a standard Internet
connection using low cost H.323 compliant video codecs operating at 576
kbps. The recording was done at the University of Florida and then
encoded with RealNetworks format. The PowerPoint slides were transmitted
from Tallahasse to Gainesville using Microsoft's Netmeeting.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
 |
|
|
A very useful feature in the Real Networks product is the provision for closed caption with language negotiation capability. Using this feature, the text of the material being discussed (or appropriate explanations there of) scrolls at the bottom of the screen.
Language negotiation is based on the language preference set through RealPlayer
options menu and the language code defined in the SMIL file. RealPlayer
chooses a specific clip to play. This facility enables hearing impaired
or foreign users to access the online. Our purpose in the regard
is the ultimate globalization of our educational offering.
Figure. 12: Language Negotiation
Note :
Click above images to see the lecture.
A very useful feature of the G2 server is that it provides a web-based facility for monitoring
the status of the G2 server [2]. Figure 13 shows
the G2 Java Monitor. The system administrator can control and
monitor the G2 server with this interface. The information provided
includes CPU usage, memory usage, the number of players connected,
which files are are being viewed. Clearly the monitor provides very useful
information and can be used to identify demand profiles and and busy intervals.
Figure. 13 : Monitoring the
Utilization of the G2 Server
(Laboratory for Information
Systems and Telecommunications)
A client/server model is used to deliver
streaming content; in this case the client is the user's installed RealPlayer
application, and the server is the RealServer G2 software. The browser
merely works to establish communication between the two, in a manner described
below figure 14.
When a user clicks a URL link to Real content within his browser, The browser sends an HTTP request to the web server (1), which then responds by sending appropriate "metafiles" back to the browser (2). The metafile contents are interpreted by the browser so that the RealPlayer application is started as a "helper application." RealPlayer now reads the URL from the metafile (3), and subsequently interacts with the RealServer using RTSP (Real Time Streaming Protocol) to request the specified content. The RealServer then streams the specified clip directly to the RealPlayer in real time (5), without any further interaction with the users browser.
The development of a sophisticated Web-based educational environment is facilitated by a powerful tool called WebCT [14]. WebCT is so flexible that it can be used to facilitate the delivery of entire online courses, or simply to publish supplementary course material.
A significant advantage in using WebCT for online course development and presentation lies in its effective use of the browser. Not only does WebCT present online course material for browsers, but it also uses the web browser as the interface for its course building environment. Thus, even non-technical users can quickly develop and administer high quality online courses that might otherwise have required the services of web programmers.
WebCT offers a rich set of features and tools that can be optionally be incorporated into any web based course, to enhance the learning experience or the utility of the interface itself. Such features include a conferencing system, online-chats, student self-evaluation, electronic mail, course calendars, student homepages, and course content searches. Tools also benefit the course instructors or administrators, like auto-marked quizzes, student progress tracking, grade maintenance and distribution, and more.
An example of
how these tools can be implemented into course material is shown in the
web page above. This is the main student interface to the course. In this
example, relevant WebCT features for the student have been under into five
major links: "The Lounge," "Course Content," "Lectures," "Useful Utilities,"
and "FAQs."
The prototype system for hybrid synchronous and asynchronous learning environment described in this paper is still in the process of development and refinement as we gain more experience and assess our preliminary effort in terms of teaching and learning effectiveness.
From a cognitive point of view, the most important challenge is to bring together the educational content that students can access in both synchronous and asynchronous modes and to unify their respective pedagogical approaches. This merger is possible if an integrated educational approach is taken. In such an approach, the expositive teaching and active learning activities should be combined and balanced according with their respective pedagogical objectives, their didactic efficiencies, and technical constraints [7].
In a hybrid synchronous and asynchronous learning framework, the lecture can be "attended" live or later played from the archived audio/video stream. In either case, with much of the traditional class material already on-line, lecture segments could now be reduced to the more desirable length of about 15 minutes, which is consistent with the defined pedagogical objectives. The extra time can then be used for the other active learning modes such as interaction and collaboration, which are often neglected in traditional education. To enhance and to support additional personal readings, printed documents or textbooks can be converted to an electronic form. Documents in PDF or HTML formats can nowadays be produced easily with commercial software or converters (LaTeX to HTML for example) and new standard such as MathML for describing mathematical expressions in WWW pages are already beginning to emerge [8]. Electronic documents offer the advantage, that while keeping the sequential structure of printed documents; dynamic links and search capabilities can enrich them. However, recent research shows that a better pedagogical efficiency can be obtained if a new structure for the internal relations inherent in the subject matter is chosen for the electronic version [9]. Remote manipulation of training resources is another paradigm, which can be developed, in the proposed hybrid learning framework. For example, synchronous demonstrations on remote facilities can be performed by the teacher and watched by students. Asynchronous real or virtual experimentation can be conducted using physical hardware at a remote location [1] or interactive simulators.
From a technical point of view, it is anticipated that several new or refined features will be added to the system in the near future. These include a more effective (and possibly voice-based) system for live feedback from remote students. Another feature, which requires some more work is a "web-cam" or an electronic based system for sending at regular intervals high quality images of handwritten documents to remote students via a specific URL. Finally we hope to investigate the use of a speech-to-text system for converting the lecture's speech to a text stream for real-time scrolling and archiving to a text file and/or WWW page.
It is not difficult to imagine that traditional model of education will be affected by the use of asychronous delivery methods. Indeed, one can envision organizations such as software firms who are not conventional providers of educational content, getting into this new business, as educational delivery becomes an attractive and lucrative business proposition. The issue of choosing an educational provider would then become similar to choosing a provider for other online services. Clearly the quality of the program must be measured by some recognized standards of accreditation; the reputation of the institution as well as overall costs will be important considerations. Only time will tell how these developments will fare when compared with more traditional forms of education. It is our view that the two modes of educational delivery will serve complementary purposes, with the traditional student/instructor model being preferred if at all possible, while the asynchronous mode will provide access to students whose circumstances would rule out the preferred mode of learning. Of course, the use of asynchronous online tools such as electronic conferencing (mailing lists, etc) will continue to grow as a complement to the traditional modes of learning.
The University of Florida's approach to technology-enhanced traditional and distance engineering education, represents significant progress in both in the way students learn and in the methods used to deliver courses. The traditional on campus students benefit from having very useful online multimedia supplements and the off-campus students can benefit from the look and feel of a traditional classroom with interactive modes supported by the ALN on which the system is based.
Thus far we have concentrated on the delivery of multimedia and streaming content and this has been relatively successful. As the individual course offering evolve into entire online degree or certificate program, we are now handling the administrative and support structures needed to facilitate these online programs. In addition we are also concentrating more attention to interactive simulations which get students more actively involved in some of the course material.
The MS program which is mainly lecture based (with an optional thesis) has been easier to develop in the lectures on demand mode. We are presently developing the BS program and this is proving to be a much more challenging exercise. To be sure some course which require hands on labs will not be offered entirely online - students will be required to take actual lab courses at convenient locations and these will then be accepted by the University at Florida as satisfying the lab requirement. In addition it is much more important to provide drill and example problems via online simulation or testing for undergraduates and so this is a major focus of the development.
Further information about the the online MS and BS programs at the University of Florida may be found at http://www.list.ufl.edu/online.
The work described in this paper was supported by a grant from the Alfred Sloan Foundation as well and the Southeastern University and College Coalition for Engineering Education (SUCCEED), funded by the National Science Foundation.
[1] LIST, EEL 5718
web page http://csc3.list.ufl.edu:8900/
[2] RealServer
G2 Administration
http://service.real.com/help/library/guides/production/realpgd.htm
[3] Stanford
University http://stanford-online.stanford.edu/demo/index.html
[4] Purdue University
http://ECE.www.ecn.purdue.edu/ECE/
[5] Progressive
Network http://www.real.com
[6] University
of Illinois UC
http://www.online.uillinois.edu/oakley/presentations/CIEC]Links.html
[7] D. Gillet,
G. F. Franklin, R. Longchamp, and D. Bonvin, "Introduction to Automatic
Control
via an Integrated Instruction Approach", The 3rd IFAC Symposium on
Advances
in Control Education, Tokyo, Japan, August 1994.
[8] http://www.wolfram.com/news/mathml/
[9] F. Michau
and D. Munteanu, "Web Server of Pedagogical Documents in Control
Engineering",
7th Annual Conference of the European Association for Education in
Electrical
and Information Engineering (EAEEIE) on Telematics for Future
Education
and Training, Oulu, Finland, 1996.
[10] H. Latchman,
C. Salzmann and D. Gillet, H. Bouzekri "Information Technology
Enhanced
Learning in Distance and Conventional Education", IEEE transaction
on
education
May 1999.
[11] H. Latchman,
C.Salzmann, D. Gillet and J. Kim, "Learning On Demand - A Hybrid
Synchronous
/ Asynchronous Approach", Proceeding of the 1999 IEEE Transactions
on
Education April 1999.
[12] Fawzi
Daoud, "Knowledge Media Learning", proceeding of the 1999 IEEE
transactions
on education April 1999.
[13] RealProducer
G2 Content Creation Guide
http://www.real.com/products/tools/authkit/index.html
[14] WebCT
Company http://www.webct.com