Interactive Technology and Smart Education ppt

Volume 4 Number 4 November 2007SPECIAL ISSUE Papers from the IEEE International Workshop on Multimedia Technologies for E-Learning MTEL Gerald Friedland, Lars Knipping and Nadine Ludwig

Volume 4 Number 4 November 2007

SPECIAL ISSUE

Papers from the IEEE International Workshop on Multimedia

Technologies for E-Learning (MTEL)

Gerald Friedland, Lars Knipping and Nadine Ludwig

Guest editorial

Vector graphics for web lectures: experiences with Adobe

Flash 9 and SVG

Authoring multimedia learning material using open standards and

free software

E-learning activity-based material recommendation system

Educational presentation systems: a workflow-oriented survey and

technical discussion

www.emeraldinsight.com/itse.htm

Interactive Technology

Education PROMOTING INNOVATION AND A HUMAN TOUCH

Interactive Technology

Education

ISSN 1741-5659

Vol 4 No 4 November 2007

Interactive Technology

Education

PROMOTING INNOVATION AND A HUMAN TOUCH

SPECIAL ISSUE

Papers from the IEEE International Workshop on Multimedia

Technologies for E-Learning (MTEL)

Gerald Friedland, Lars Knipping and Nadine Ludwig

Guest editorial

Vector graphics for web lectures: experiences with Adobe Flash 9 and SVG

Authoring multimedia learning material using open standards

and free software

E-learning activity-based material recommendation system

Educational presentation systems: a workflow-oriented survey and

technical discussion

Professor Alistair Sutcliffe, University of Manchester, UK

Anne Adams, UCL Interaction Centre, UK

Petek Askar, Hacettepe University, Turkey

Ray Barker, British Educational Suppliers Association, UK

Maria Bonito, Technical University of Lisbon, Portugal

Marie-Michèle Boulet, Université Laval, Canada

Sandra Cairncross, Napier University, UK

Gayle J Calverley, University of Manchester, UK

John M Carroll, Penn State University, USA

Chaomei Chen, Drexel University, USA

Sara de Freitas, Birkbeck University of London, UK

Alan Dix, Lancaster University, UK

Khalil Drira, LAAS-CNRS, France

Bert Einsiedel, University of Alberta, Canada

Xristine Faulkner, London South Bank University, UK

Terence Fernando, University of Salford, UK

Gerhard Fischer, University of Colorado, CO, USA

Monika Fleischmann, Fraunhofer Institute for Media Communication, Germany

Giancarlo Fortino, University of Calabria, Italy

Gerald Friedland, Freie Universität Berlin, Germany

Bernie Garrett, University of British Columbia, Canada

Lisa Gjedde, Danish University of Education, Denmark

Ugur Halici, Middle East Technical University, Turkey

Lakhmi Jain, University of South Australia, Australia

Joanna Jedrzejowicz, University of Gdansk, Poland

Joaquim A Jorge, Technical University of Lisbon, Portugal

Athanasios Karoulis, Aristotle University of Thessaloniki, Greece

Lars Knipping, Freie Universität Berlin, Germany

John R Lee, University of Edinburgh, UK

Paul Leng, Liverpool University, UK

Anthony Lilley, magiclantern, UK

Zhengjie Liu, Dalian Maritime University, China

Nadia Magnenat-Thalmann, University of Geneva, Switzerland

Terry Mayes, Glasgow Caledonian University, UK

Toshio Okamoto, University of Electro-Communications, Japan

Martin Owen, NESTA Futurelab, UK

Vasile Palade, Oxford University, UK

Roy Rada, University of Maryland, MD, USA

Elaine M Raybourn, Sandia National Laboratories, NM, USA

Rhonda Riachi, Oxford Brookes University (ALT), UK

Kerstin Röse, University of Kaiserslautern, Germany

Joze Rugelj, University of Ljubljana, Slovenia

Eileen Scanlon, Open University, UK

Jane K Seale, University of Southampton, UK

Helen Sharp, Open University, UK

Vivien Sieber, University of Oxford, UK

David Sloan, University of Dundee, UK

Andy Smith, University of Luton, UK

Paul Strickland, Liverpool John Moores University, UK

Josie Taylor, Open University, UK

Malcolm J Taylor, Liverpool University, UK

Thierry Villemur, LAAS-CNRS, France

Weigeng Wang, University of Manchester, UK

Editorial Advisory Board

Honorary Advisory Editor

Dr Claude Ghaoui

School of Computing & Mathematical Sciences, Liverpool John Moores University, Byrom Street,Liverpool L3 3AF, UK Email: c.ghaoui@LJMU.ac.uk

Ever since the advent of automatic computation devices,

efforts have been made to answer the question of how to

properly integrate them and take advantage of their

capa-bilities in education Educational multimedia systems

promise to make learning easier, more convenient, and thus

more effective For example, classroom teaching enriched

by vivid presentations promise to improve the motivation

of the learner Concepts may be given a perceivable

exis-tence in a video show and the observability of important

details can be stressed Video capturing of lectures has

become common practice to produce distance education

content directly from the classroom Simulations allow

stu-dents to explore experiments which would be otherwise

impossible to be conducted physically by students

Today, almost every university claims to have a

strate-gy to utilize the opportunities provided by the Internet

or digital media in order to improve and advance

tradi-tional education However, the question about how

mul-timedia can really make education more exploratory and

enjoyable is as yet not completely answered In fact, we

are just beginning to understand the real contribution of

multimedia to education For example, various web sites

and lecture videos produced as part of the “e-learning

hype” often do not exploit the full potential of multimedia

for teaching For example, how can we support participantinteraction in classrooms and lecture halls better? Whatare the best tools for the development of educational mul-timedia material? How can we make the production ofeducational material easier and existing application morereusable?

In addition, new technologies and trends – such asmobile and semantic computing – open up new andexciting opportunities for teaching with multimedia andthe creation of multimedia learning material How canthese new trends in multimedia research be used toimprove multimedia education or education in general?

In order to find answers to these and many other tions, we organized the second IEEE InternationalWorkshop on Multimedia Technologies for E-Learning(MTEL) in connection with the 9th IEEE InternationalSymposium on Multimedia Based on the success of thefirst MTEL workshop in 2006, our goal was to attractresearchers and educators from the multimedia commu-nity as well as researchers from other fields, such assemantic computing and HCI, who are working onissues that could help improve multimedia education aswell as teaching and learning in general Based on dis-cussion among these experts with different backgrounds,the workshop’s aimed to identify new trends and high-light future directions for multimedia-based teaching

ques-The following special issue of Interactive Technology

Smart Education presents four papers that have been

carefully selected by the program committee for

publica-tion in this journal They have been extended by the

authors according to reviewers suggestions We hope that

these articles are able to inspire even more creativity in

the overlap between human-centered and

technology-centered research The following paragraphs provide a

short overview of the selected articles

Vector Graphics for Web Lectures: Experiences with Adobe

Flash 9 and SVG, presents experiences made during the

development and every day use of two versions of the

lec-ture recording system virtPresenter The first of these

versions is based on SVG while the second one is based on

Adobe Flex2 (Flash 9) technology The authors point out

the advantages vector graphics can bring for web lectures

and briefly present a hypermedia navigation interface for

web lectures that is based on SVG Also, they compare

the formats Flash and SVG and conclude with describing

changes in workflows for administrators and users that

have become possible with Flash

Authoring Multimedia Learning Material using Open

Standards and Free Software deals with avoiding drawbacks

like license cost and software company dependencies at

distributing interactive multimedia learning materials

The authors propose using open data standards and free

software as an alternative without these inconveniences

But available authoring tools are commonly less

produc-tive The proposal is based on SMIL as composition

lan-guage particularly the reuse and customization of SMIL

templates used by INRIA on their technical

presenta-tions The authors also propose a set of free tools to

pro-duce presentation content and design focusing on

RealPlayer as delivery client

E-Learning Activity-based Material Recommendation

System an application to utilize the techniques of LDAP

and JAXB to reduce the load of search engines and the

complexity of content parsing is described Additionally,

through analyzing the logs of learners’ learning

behav-iors, the likely keywords and the association among the

learning course contents will be conducted or figured

out In conclusion, the integration of metadata of the

learning materials in different platforms and

mainte-nance in the LDAP server specified

Finally, Educational Presentation Systems: a

workflow-oriented survey and technical discussion presents an overview of

processes before, during and after an educational

presenta-tion The different processes are presented in form of a

workflow The workflow is also used in order to present,

analyze and discuss different systems including their

individual tools covering the different phases of the

workflow After this overview of systems, the different

approaches are discussed in respect to the workflow This

discussion provides specific technical details and

differ-ences of the focused systems

ACKNOWLEDGEMENTS

The Guest Editors wish to thank Claude Ghaoui, ITSE

Editor-in-Chief, and then dedicated reviewers for theirdetailed and thoughtful work They were:

Abdallah Al-Zoubi, Princess Sumaya University forTechnology, Jordan

Michael E Auer, Carinthia Tech Institute, AustriaHelmar Burkhart, University of Basel, SwitzerlandPaul Dickson, University of Massachusetts, USABerna Erol, Ricoh California Research Center, USARosta Farzan, University of Pittsburgh, USAClaude Ghaoui, Liverpool John Moores University, UKWolfgang Hürst, University of Freiburg, GermanySabina Jeschke, University of Stuttgart, GermanyUlrich Kortenkamp, Paedagogische HochschuleGmuend, Germany

Ying Li, IBM T.J Watson Research Center, USAMarcus Liwicki, University of Bern, SwitzerlandRobert Mertens, University of Osnabrück, GermanyJean-Claude Moissinac, ENST Paris, France

Thomas Richter, University of Stuttgart, GermanyAnna Marina Scapolla, University of Genova, ItalyGeorg Turban, Darmstadt Institute of Technology,Germany

Nick Weaver, ICSI Berkeley, USADebora Weber-Wulff, FHTW Berlin, GermanyMarc Wilke, University of Stuttgart, GermanyPeter Ziewer, Munich Institute of Technology, Germany

We would like to thank all authors for their quickrevision and extension of the articles presented herein.Their commitment made it, again, possible to release thisspecial issue so quickly after the workshop

REFERENCE

Friedland, G., Knipping, L., and Ludwig N (2007), “SecondIEEE International Workshop in Multimedia Technologies

for E-Learning”, Proceedings of the 9th IEEE International

Symposium on Multimedia, IEEE Computer Society, Taichung, Taiwan, pp 343–95.

ABOUT THE GUEST EDITORS

Dr Gerald Friedland is currently a researcher at theInternational Computer Science Institute in Berkeley,California Prior to that, he was a member of the multi-media group of the computer science department of FreieUniversität Berlin His work concentrates on intelligentmultimedia technology with a focus on methods thathelp people to easily create, edit, and navigate content,aiming at creating solutions that “do what the usermeans” He is program co-chair of the 10th IEEE

Symposium on Multimedia and the Second IEEE

International Conference on Semantic Computing In

addition to the second IEEE International Workshop on

Multimedia Technologies for E-learning, he also

co-chaired the first ACM Workshop on Educational

Multimedia and Multimedia Education He has received

several international research and industry awards

Among them is the European Academic Software Award

in 2002, for the creation of the E-Chalk system in

coop-eration with Lars Knipping He is also member of the

editorial advisory board of ITSE

Dr Lars Knipping is a researcher at the mathematics

department at Technische Universität Berlin He belongs

to the board of editors of ITSE and the editorial team of

iJET (International Journal of Emerging Technologies in

Learning) Before joining Technische Universität he

worked as a scientific consultant in a research project for

a state-funded TV broadcaster, the “Sender Freies Berlin”,followed by positions as researcher and instructor at themultimedia group at the computer science department ofFreie Universität Berlin and as lecturer in InternationalMedia and Computing at the FHTW Berlin Dr Knippingreceived his Ph.D degree for his work on the E-Chalksystem and holds M.Sc degrees in both mathematics andcomputer science

Nadine Ludwig graduated from Technische UniversitätIlmenau with a degree in Computer Science in 2005 Inher thesis she described the integration of remote labora-tories in Learning Content Management Systems viaSCORM Since May 2006 Ms Ludwig has been a part ofthe MuLF Center at Technische Universität Berlin as aresearch associate Currently she is working on her PhD-thesis in the field of Semantics and Modularization

of Learning Objects in Cooperative Knowledge Spaces

Vector graphics for web lectures:

experiences with Adobe Flash 9 and SVG

Markus Ketterl

Virtual Teaching Support Center, University of Osnabrück, Osnabrück, Germany

Email: mketterl@uni-osnabrueck.de

Robert Mertens

Fraunhofer IAIS, Schloß Birlinghoven, Sankt Augustin, Germany

Email: robert.mertens@iais.fraunhofer.de and

Oliver Vornberger

Department of Computer Science, University of Osnabrück, Osnabrück, Germany

Email: oliver@uni-osnabrueck.de

Abstract

Purpose – The purpose of this paper is to is to describe vector graphics for web lectures, focusing on the experiences

with Adobe Flash 9 and SVG.

Design/methodology/approach – The paper presents experiences made during the development and everyday use

of two versions of the lecture-recording system virtPresenter The first of these versions is based on SVG, while the second is based on Adobe Flex2 (Flash 9) technology The paper points out the advantages vector graphics can bring for web lectures and briefly presents a hypermedia navigation interface for web lectures that is based on SVG The paper also compares the formats Flash and SVG and concludes with describing changes in workflows for administra- tors and users that have become possible with Flash.

Findings – Vector graphics are an ideal content format for slide-based lecture recordings File sizes can be kept small

and graphics can be displayed in superior quality Information about text and slide objects is stored symbolically, which allows texts to be searched and objects on slides to be used interactively, for example, for navigation purposes The use of vector graphics for web lectures is, however, a trend that has begun only recently A major reason for this

is that multiple media formats have to be combined in order to replay video and slides.

Originality/value – The paper offers in insight into vector graphics as an ideal content format for slide-based lecture

recordings.

Keywords: Lectures, Worldwide web, Graphical user interfaces, Presentation graphics, Multimedia, Teaching aids Paper type: Research paper

Vector based graphics formats offer a number of

possi-bilities for the realization of web lecture interfaces for

slide based talks One major advantage is that they support capturing contents in a symbolic manner which

is a requirement for searching text in a recording (Lauerand Ottmann, 2002) They also offer superior picture

quality Last but not least, vector based graphics formats

enable developers to realize a high degree of interactivity

that can be used for implementing advanced navigation

concepts as described in (Mertens et al 2006d) They also

can be used to tackle a number of layout problems as

fur-ther described in (Mertens et al 2006b).

Vector graphics are, however, not very common in web

lectures This article presents the authors’ experience

with two different vector graphics formats: (scalable

vector graphics SVG ) and Adobe’s new Flex 2 (Flash 9

based) technology for content presentation and control in

the web lecture system virtPresenter

The SVG based version of the lecture recording system

has been used at the University of Osnabrück and at the

University of Applied Sciences Osnabrück since summer

2003 During this time, users with different backgrounds,

knowledge and expectations experienced the system in

every day use The Adobe Flex 2 based counterpart has been

introduced in February 2007 after a seven month

develop-ment and testing period This new version is, apart from

small changes concerning further system requirements and

improvements, in productive use since March 2007

The article is organized as follows: Section 2 points out

the advantages vector graphics can bring for web lectures

and briefly presents a hypermedia navigation interface for

web lectures that is based on SVG Section 3 describes

experiences with this SVG based interface and points out

difficulties that arose during the use of this interface in a

number of university courses Section 4 compares Flash

and SVG with respect to their use in lecture recording

Section 5 introduces the Flash based successor of the

SVG based interface Section 6 describes changes in

work-flows for administrators and users that have become

possi-ble with Flash Section 7 briefly summarizes the work

pre-sented in this article and refers to future projects and ideas

GRAPHICS IN WEB LECTURES

The advantages of using vector graphics for content

rep-resentation in web lectures can be summarized in a couple

of words: vector graphics store content in a symbolic way,

vector graphics can be enlarged without loss of quality

and many vector graphics formats allow for interactive

on-the-fly manipulation of contents The aim of this section

is to show why these properties of vector graphics are

use-ful by showing how each of them improves web lectures

Contents and Interactivity

The original virtPresenter user interface shown in

Figure 1 was developed to implement a hypermedia

navigation concept for lecture recordings (Mertens,2007) Hypermedia navigation consists of the five ele-ments full text search, bookmarks, backtracking, struc-tural elements and footprints (Bieber, 2000)

Full text search is realized by searching the text of theslides in the slide overview Search results are highlight-

ed by an animation that grows and shrinks them edly Both the ability to search in the slides directly and

repeat-to animate search results is based on the properties ofSVG (symbolic representation and manipulation on thefly) Bookmarks are realized as a functionality thatallows for selecting arbitrary passages and storing themfor later viewing or exchanging them with other students Backtracking is implemented by storing theplay position whenever the user navigates to anotherplay position Thus each navigation action can beundone In order to facilitate orientation at the storedplay positions, replay begins at their time index minusthree seconds Structural elements are realized in twoways the simple one of which are next/previous buttonsthat allow navigating to the next or previous slide oranimation step A more sophisticate realization of struc-tural elements is the interactive slide overview imple-

mented in virtPresenter (Mertens et al., 2006c) In the

overview, those parts of a slide that had been animatedduring the original presentation when the lecture wasrecorded can be clicked on with the mouse The record-ing then starts replay at the time index when the respec-tive animation takes place during the lecture To realizethese features, the slide documents are analyzed andscript code containing the respective time indices isadded automatically to the animated elements of a slide

(Mertens et al., 2007) The implementation of this step

was relatively easy due to the symbolic representation ofthe slide elements in SVG Footprints serve the purpose

of showing users which parts of a hyperdocument theyhave already visited In classic hypertext, this is done bycolouring visited and non-visited links differently Since web lectures are time based media, anotherapproach had to be found In virtPresenter, colouredparts of the timeline indicate that the correspondingpassages of the recording have already been watched bythe user Multiple visits are indicated by deeper shad-ings The footprints are stored symbolically as pairs ofstart and end time indices They are drawn on the flywhen a lecture is watched This has been realized by the use of animated SVG rectangles The different colourshadings are created by overlapping semitransparent rectangles

This brief description shows that the properties ofSVG as a vector graphics format have been crucial for therealization of the virtPresenter user interface Especiallythe implementation of footprints, bookmarks and fulltext search has been facilitated immensely by SVG as avector graphics format

2.2 Superior Picture Quality

Good picture quality of lecture slides is important even

for standard usage scenarios (Ziewer and Seidl, 2002)

However, it becomes even more important, when the

lec-ture slides are shown on a large screen as in the scenario

depicted schematically in Figure 2

In this scenario, the lecture is replaced by a

cinema-like session in which the recording of the lecturer and the

slides are presented to the audience on two large screens

This scenario has been carried out successfully at the

University of Osnabrück a number of times (Mertens

et al., 2005) Since the slides are shown on a large screen,

bad picture quality becomes even more obvious than

dur-ing replay on a standard computer display At the

University of Osnabrück, the slides used had been in

SVG and had thus been presented in the same quality as

in the original lecture

The SVG-based version of the viewer interface was first

developed in 2003 and improved in various steps The

main focus of the development was to implement the

hypermedia navigation concept for lecture recordings

described in section 2 and in more detail in (Mertens

et al., 2004).

At the time when development of the SVG based

ver-sion began, SVG seemed to be a promising choice for a

content format to be used in lecture recordings SVG is

an XML based vector graphics format and was expected

to grow in importance We had expected that SVG

ren-derers supporting the required subset of the SVG

stan-dard would soon become available on more platforms

than Windows and that their performance would increase

in order to rival that of Macromedia Flash (now Adobe

Flash) Things have, however, developed in a differentdirection

While all the features described in (Mertens et al.,

2004) could be realized with a combination of JavaScript,SVG and Real Video, the technology used lead to a num-ber of problems in every-day use Loading and renderingspeed has shown to be a major problem when combiningSVG and Real technology Table 1 compares slide load-ing times of the SVG and the Flash based implementa-tion (further described in sections 4 and 5) It also showsloading times for an optimized version of the SVG slides

in which background graphics in the slides (logos) hadbeen deleted to speed up rendering The testing environ-ment was a Windows XP system with an AMD Athlon

64 based processor with 2,01 GHz and 1 GB RAM Thetests were made locally on that system without internetconnection interferences This test indicates the elapsedtime till a slide object is loaded and fully available in themain application

As some interactivity and animation features of SVGthat are only supported in the Adobe SVG Viewer (ASV)had been used in the interface, replay was only possiblewith the ASV for Microsoft’s Internet Explorer (IE) Thisviewer plug-in does, however, exhibit low renderingspeeds and support will be discontinued in January 2008.This fact is especially problematic when many slides have

to be shown at once as it is the case for overviews Alsoswitching from one slide to another happens with anoticeable delay The Real video player buffers data whenusers navigate in the video This buffering also slowsdown the interfaces responding times noticeably.Another problem with SVG was that the plug-inrequired only exists for Microsoft’s Internet Explorer.Even though Adobe had implemented plug-in-versionsfor other browsers, only the one for IE supports the subset

of the SVG specification required for the tion This fact rules out platform independence for the

implementa-Figure 1 VirtPresenter 1.0 user interface

interface Last but not least, the fact that plug-ins are

required for both Real Video and SVG poses an obstacle

for first time users of the interface

The use of the SVG-based interface has been evaluated

in a number of courses In these evaluations, the above

mentioned points have shown to have a considerable

neg-ative impact on user acceptance In 2006, three courses

have been evaluated with a questionnaire developed for

the evaluation of e-Learning at the University of

Osnabrück For abbreviation purposes, these courses are

referred to in the paper as courses A, B and C Table 2

summarizes relevant details on the courses

Figure 3 shows how the students judged download

times of the recordings No actual download was offered

The term “download times” does thus refer to loading

and rendering times of the viewer interface By and large

the numbers in the figure do not seem too critical at first

sight In practice, however, the interface loads

consider-ably longer than other material found on the course web

site Also, the results show that while the loading times

have been acceptable for most students, they have notbeen acceptable for all students

Figure 4 shows how many students reported problemsusing virtPresenter The problem descriptions wereentered as free text answers in the questionnaires Incourse A, no student reported a problem This might bedue to the fact that students were given very detailedinstructions Having a non-technical background, thestudents have very likely followed these instructionsclosely The questionnaires have also shown that all stu-dents in course A used IE In the other courses, the ques-tionnaires have shown that some students did not use IE(even though they had been instructed that using anotherweb browser would cause problems with the interface) Incontrast to course A, course B and C had been attended

by a number of students with technical backgrounds Thequestionnaires lead to the assumptions that some of thesestudents, being used to solve problems by trial and error,have tried to use the interface with other browsers than

IE unregarding the information that it would not work

on these browsers Seemingly unaware of the fact that theinterface was not supported under these settings, the stu-dents reported the system behaviour as faults From oneproblem description it even became clear, that the stu-dent had not installed any SVG viewer

In order to counter the above described effects, a ber of improvements had been devised for the SVG basedversion of the interface For example, a nearly equivalentsolution with QuickTime video instead of Real video thatalso works with SVG for the slide representation and aFlash 6 based thumbnail overview component for fasterslide loading and interface responding This approach ofmixing technology did not solve the problems either Thereason was that the users had to install another plug-in,QuickTime instead of Real as well as the Flash plug-in

num-Table 1 Slide loading with SVG and Flash

Technology SVG SVG optimized Flash

average slide 164* 120* 67

loading time (ms)

430** 243** 81 average slide

loading time

1 video (ms) (Real video) (Real video) (Flash video)

average slide

13 different converted PowerPoint slides System: Windows XP

AMD Athlon 64 Processor;

*outlier here: 520, 635 2,01 GHz, 1GB Ram

**outlier here: 7300, 6349, 2280, 4300

Figure 2 Lecture slides on large screens

Moreover the reaction time of the interface could not be

improved by this approach

As a preliminary workaround, plug-in and browser

checks had been added to the original version These

measures alert users if they try to use the interface with

wrong software settings and thus reduced bug reports

that are due to accessing the interface with wrong

soft-ware setting Also, a number of enhancements had been

added to avoid unnecessary loading of slides when slide

changes happen at a high frequency

These approaches have, however, been limited by the

technology setting in which they had been employed In

order to overcome these problems, we have turned to

Adobe Flex 2 in combination with the Open Source Red5streaming server backend as described in section 5

SVG

In a strict sense, the new interface cannot reach the tion range of the old virtPresenter interface described in

func-(Mertens et al., 2006a; Mertens, 2007) by now.

This is mainly due to the fact that the new version doesnot yet feature an automatically generated thumbnailslide overview which is crucial to a number of function-alities implemented in the SVG based version (Mertens

et al., Mertens et al., 2004, 2006d) The thumbnail

overview is used both to visualize the connection of navigation actions to the structure of a talk (Mertens

et al., 2006d) and to allow structure based navigation on

the level of animations within a slide The latter is ized by clickable slide elements that allow for direct nav-igation to the replay position when the correspondingslide element first appeared on screen during the record-

real-ed lecture (Mertens et al., 2004).

However, the reimplementation was necessary due tofrequent user problems with unsupported computer plat-forms, wrong browsers or browser settings or missingplug-ins The underlying shared infrastructure (Mertens

et al., 2007) was enhanced to export, besides different

podcast formats flash content (Flash video and Flash

slides) (Ketterl et al., 2006b, 2007a) Adobe’s Presenter

(formerly named Breeze) is now also a part of the matic lecture recording production chain This softwarecomponent enables a fast PowerPoint to Flash conversionthat could be fully automated as well This software com-ponent was selected in this new process due to the factthat it is reliable and now even affordable for a smalleruniversity project Today there are some open source orcommercial PowerPoint to Flash export systems besidesthe Adobe product on the market However, AdobePresenter currently seems to be the only system that fits into our automated production chain The other systems could not be integrated in the automatic production chain as they could not be started from other

auto-Figure 4 User problems

Table 2 Course details

Course Full Name of Course Didactic Setting Number of Students

A Fundamentals of Biblical Theology Lecture took place as usual, all students 25

could attend and the recordings had been provided as an add on.

B Internet Technologies Lecture took place at one University and

was transmitted to another one Recordings were provided as an add on A more detailed description of the scenario can be found in (Hoppe et al., 2007) 27

C Managing Innovation and Projects Same as course C 19

Figure 3 Lecture recordings download times

programmes A Problem with Adobe Presenter is

con-stituted by the fact that this component exports only

Flash 6 slides in the current version The communication

between old Flash objects and new Flash 9 objects is not

ideal at the moment Difficult is for example the

han-dling of different old Flash version based slides in a Flash

9 application A prototype version which also features

slide based navigation is depicted in Figure 5 on the left

hand side

Nevertheless, the time for post processing (video and

slide conversion and slide text analysis and building all

the required software files for the interface could be

reduced from previously about three hours down to only

about one hour for a 1.5 hour lecture Of particular

importance is here, that the flash video conversion is

much faster than our previous Real video conversion Our

initial recording format here is still MPEG-2 because of

the fact that this video format is of good quality and can

be converted into many different video/audio formats in

the post processing process

Figure 5 (right) depicts the revised and newly

imple-mented Flex 2 based web interface Besides the objective

of using it on any computer platform without

adjustments, the aim was that people without a technical

background could use the interface as easily as internet

experts On the right hand side of Figure 5 one can find

an area where users can choose from a list of recorded

lec-tures or search text in the recordings Figure 6 shows this

lecture list (section a) and search results (section b) in a

more detailed view The lecture list gets updated over an

RSS notify mechanism Inspirational were our positive

experiences with Apple’s iTunes, their popular Music

Store and the podcast subscriber facility (Ketterl et al.,

2006a, b) The main reason why we do not use Apple’s

iTunes (or other podcatcher software) and the podcast

technology as main distribution facility is, that the

navi-gation possibilities in podcasts are limited compared to

the navigation options in the virtPresenter system

Further inquiries about navigation in lecture podcasts

and how lecture podcasts are being used in contrast to the

normal lecture recordings are ongoing Several

examina-tion results with student users and external users are

described in (Schulze et al., 2007) for virtPresenter and

(Hürst and Welte, 2007a) for a system used at the

University of Freiburg In the revised virtPresenter

sys-tem, users can subscribe to lecture recordings using our

internal university learn management system Stud IP

(www.studip.de) The virtPresenter interface gets updated

and shows the lecture recordings as soon as they are

avail-able Aside from that, external users can subscribe to the

recordings (like subscribing to a normal podcast with a

podcatcher software like Apple’s iTunes) and can view

recordings for example that are open for public viewing

This lecture recording offer is presented over a public

website In short, this means that students as well as

external viewers use the same interface for differentrecordings They do not need to switch between applica-tions and there is no need to follow additional links inother browser windows The interface can also be used if

a link from our lecture website or the LMS points to aspecific lecture or a specific time index in a recording.This is done by interpreting assigned url parameters Thefeature is a further extension of a functionality imple-mented for the SVG based version and described in fur-

ther detail in (Mertens et al., 2005).

Section b in Figure 6 also depicts a possibility to search

in the recordings Users can search not only in one weblecture but in all recordings they have subscribed to Thesearch results are presented in a hierarchical tree overviewsimilar to Adobe’s Acrobat The results can be selectedand are linked directly to the corresponding lecturerecording section

Due to the changeover to Flex 2 technology, users cannavigate fluently in the recordings with a new timescrubber component (see Figure 7) In the SVG basedversion, visible scrolling in the sense of (Hürst andMüller, 1999) was only possible with the slides used inthe recording, in the Flex 2 based version, it is possiblefor both slides and video Presently we highlight slideborders in the timeline and show the lecture slide titledirectly above the respective area of the timeline Colour-coded are the sections which have been viewed already bythe user When a lecturer is using the mouse cursor dur-ing the presentation, this data is also logged with theunderlying recording system and the data can be pre-sented in the user interface as well

The Flex based interface responds considerably quickerthan the old one (see Table 1) Delays resulting from slideloading, jumps to other sections or disturbing videobuffering that we had in the old Real video respectivelySVG based version are not noticeable anymore Even acomplete reload of the system due to a browser refresh isquick The interface was tested on Windows, Linux orMac OS X computer platforms, all with the Flash 9 playerplugin The results described were alike on all platforms

MADE THINGS FAST

For the new implementation of the lecture recordingsystem we used Adobe’s Flex 2 technology (this technol-ogy was introduced in June 06) for the user interface andfor user interaction Flex 2 is based entirely onActionScript 3, which was introduced as a revised andextended programming language as part of Adobe’s newFlash 9 player Flex applications are deployed as com-piled byte code that is executed within the Flash playerruntime The core of Flex is the developer-centric Flex framework, a library of ActionScript 3 objects that

provide a foundation for building rich internet

applica-tions Writing applications with Flex is similar to

devel-oping in NET or Java (Kazoun and Lott, 2007) Also,

Flex provides a wealth of useful components so that

developers do not have to build everything from scratch

Important besides the comfortable developer framework

in our scenario is that neither a special browser version

nor a combination of different plug-ins has to be

installed on the users’ computers (like needed in the

SVG based implementation) The user only needs the

Flash player plug-in for viewing the web lecture

record-ings The current plug-in version is Flash 9, which is

available for browsers on Windows (IE, Firefox and

Opera), Apple (Safari, Firefox) and Linux (Firefox) as

well Normally this plug-in can be installed without

dif-ficulties or special computer knowledge Besides, this

software component is very popular and widespread

nowadays (Téllez, 2007) That means that no special

browser adjustments or compatibility checks are

required The same version will work on different

com-puter platforms as a cross browser solution The plug-in

base for ActionScript 3 is a newly implemented virtual

machine called ActionScript Virtual Machine 2 (AVM2)that converts byte code into native machine code It ismore like a Java Virtual Machine (Java VM) or the NETCommon Language Runtime (CLR) than a browserscript engine The most important advantage is (and this

is a main reason why we are using Flash 9) that the newbrowser environment is faster than previous versions and

it uses much less memory on the computer (Adobe,2007) We could confirm this assertion in our daily workwith the new Flex 2 framework Student users reportthat they like how fast the new interface responds andreacts to user interaction Further user acceptance/prob-lem surveys are planned for February 2008

In order to respond fast, a further component is tant Like mentioned before, a main problem was thevideo buffering of the Real player in the interface A ded-icated and reliable video server is also required Like mostuniversities we have a fairly good server infrastructurebackend Through that we could use Adobe’s recom-mended and expensive Flash Media Server 2 for workingwith recorded lecture videos Instead of this expensivesolution we a have used an open source Flash streamingserver implemented in Java for a couple of months nowwhich is called Red5 (Red5 2007) The adoption was anexperiment, because this open source server deploymentwas not really stress tested, barely documented and onlyavailable in version state 0.6 (currently version 0.6.3 isavailable) The server worked very stable even during thecritical exam time at the end of the term

impor-Our productive streaming system used during thattime was a 2,8 Ghz Intel Dual Core Xenon processorbased Windows XP system with 4 GByte RAM Thisvideo server system is more than adequate with sufficientreserves in case of user request peaks At present there is

no need to use Adobe’s expensive Flash Media Server 2solution in our production environment

Figure 7 Timeline with slide border visualization and

slide title overview

Figure 5 VirtPresenter 2 Flex technology based interface

Figure 6 RSS updated lecture overview

with lecture search

6 BEHIND THE SCENES:

ADMINISTRA-TION AND WORKFLOWS

Lecture recording with virtPresenter makes use of a fully

automated recording and an extended production chain

described in (Ketterl et al., 2007a) While this process is

fully automated, a number of administration tasks still

remained Currently we manage and generate eighteen

web lecture recordings with additional podcasts (Ketterl

et al., 2006a) from different university courses in

differ-ent rooms a week plus some additional recordings for

special occasions like conferences and workshops with

this system This number increases steadily The lecture

recording system is tightly connected to the learn

man-agement system Stud.IP used at the University of

Osnabrück We have also defined more general interfaces

that make metadata like the name of the course, the name

of the lecturer and data for full text search available to

other systems like content portals or search engines

These interfaces also allow for authentication handling

by the other system Thus users do not have to log in

separately in the lecture viewer since they are

authenti-cated externally, e.g by the portal

Normally the recordings are assigned to the web-page

of the course in the university LMS Figure 8 shows what

this integration looks like in our university LMS Stud.IP

The recordings can additionally be tagged with further

meta-data or can be stored in other database systems

wherefrom further platforms can use them as well At

pres-ent we are working on a rights managempres-ent system for the

recordings that will serve the purpose of defining whether

episodes are available for university members, publically

(distribution over Apple’s iTunes music store (Ketterl

et al., 2006a) for example), as part of a course exchange

pro-gramme with other universities or on a pay per view basis

A recurring administration task at the end of a study

term is to bring the web lecture recordings offline on a

computer DVD or a CD for data backup purposes, or for

students respectively lecturers whishing to watch the

lec-ture recordings offline The normal approach in our

pro-duction system was to copy the recorded video, the

lec-ture slides and the complete source code for the web

interface on that offline medium In addition to the fact

that it is not very convenient for users to start the

record-ings by clicking a specific file link in the DVD file

sys-tem we had the drawback that the complete (maybe

copyrighted) material is on that offline medium as well

Over the internet, we had at least user authentication to

protect the content A more attractive and promising way

to reduce administration effort and to keep the content

protected is to use Adobe’s new integrated runtime

envi-ronment called AIR (prior development name Apollo)

AIR stands for Adobe Integrated Runtime

The environment is a new cross-platform desktop

run-time that allows web developers to use web technologies

to build and deploy Rich Internet Applications and web

applications to the desktop (Chambers et al., 2007).

During the last years, there has been an acceleratingtrend of applications moving from the desktop to theweb browser With the maturation of the Flash Playerruntime and Ajax type functionality it became possiblefor developers to offer richer application experienceswithout disturbing page refreshes This means that theFlex implementation of the web lecture system can beinstalled offline on a Windows PC or on a Macintosh sys-tem (a Linux version is promised by Adobe to appear bythe end of 2007) and it will behave like any other appli-cation on the system On Windows, for example, thevirtPresenter web lectures appear now offline in the startmenu and in the windows taskbar As a drawback, usershave to install the AIR runtime on their system

The adoption of this technology in general is still inquestion Why should users prefer a web like application

on their normal desktop computers? Unlike thisapproach there are other projects and ideas that focus

on the web as an operating system (Vahdat et al., 1996)

or new alternative technologies as described in the nextsection

In the literature one can find further examples forusing RIAs on the desktop or ideas for adopting this

technology (Chambers et al., 2007) In our lecture

record-ing production environment, AIR solves some of theoffline related problems We can offer virtPresenterrecorded AIR versions for standard download in case of aRed5 streaming server breakdown Another prospect isthat users do not need to be online while watching thelecture recordings since the AIR application couldinclude all required files The offline application getsupdated through a new interpretation of the associatedRSS files whenever the computer is online and new data(new lecture recordings) can be transferred and updated

in the offline version

For a simple lecture recording data backup mentioned

in the beginning of this chapter, AIR is not an option,due to the fact that the content is encapsulated in theAIR application and it is problematic to disassemble it

RESEARCH

During the last few years, Flash has evolved into an idealcontent format for web lectures Especially the fact thatboth slides and video can be replayed with one singlebrowser plug-in makes web lecture interfaces built uponthis technology easy to use for almost anyone This paperhas demonstrated the feasibility of a Flex 2 based userinterface for web lectures and it has shown that this tech-nology can be used to improve usability and ease theadministrative workload

With AIR it is even possible to protect content in

offline versions of a web lecture Given the fact that AIR

and AIR- or Flash-like approaches (Silverlight (Cohen,

2007), the JavaFX family, or Google Gears) are rumoured

to be supported by a number of mobile devices in the

near future, AIR could also open more perspectives for

interactive presentation of web lectures on mobile

devices If AIR on mobile devices worked just like

con-ventional AIR applications, it would be possible to

pro-duce learning content that can be used for normal

web-sites and for m-learning modules at he same time, that is

without expensive device adjustments Our lecture

pod-casts (audio, video and enhanced podcast versions)

(Ketterl et al., 2006a, b) were a step forward to support

mobile users with fine granulated lecture recordings

In combination with additional mobile self

assess-ments as developed for the system presented here (Ketterl

et al., 2007b) and other systems (Hürst et al 2007a)

learning on the go becomes possible The podcast

tech-nology has a drawback at present for mobile learners

Mobile users cannot give feedback to the lecturer for

example due to technical limitations of devices and of the

podcast technology With full AIR support on mobile

devices, it is likely that these problems could be solved

easily as one AIR application could run on different

plat-forms (mobile, internet and desktop)

Another branch we are pursuing in the Flex based

ver-sion of the interface is implementing social navigation

functionalities that had previously been tested in the

SVG based version of the interface (Mertens et al.,

2006a) Flex 2 does, however, open new perspectives for

social navigation in lecture recordings The reduced

load-ing times allow for editload-ing and rearrangload-ing content on

the client side without having to change its server side

representation It is also easier to embed the player in

other web sites To prove this, some of our lecture

record-ings and the newly implemented Flex 2 based

virtPresenter interface have been integrated as an

application in the social community Facebook An issuethat does still remain to be solved is how navigation can

be facilitated in re-arranged and re-structured content

REFERENCES

Adobe Systems Incorporated (2007), “Flex2 technical overview:technical whitepaper”, available at: www.adobe.com/products/flex/whitepapers/pdfs/flex2wp_technicaloverview.pdf(accessed December 2007)

Bieber, M (2000), in Ralston, A., Reilly, E and Hemmendinger,

D (Eds), Hypertext Encyclopaedia of Computer Science, 4th ed.,

Nature Publishing Group, pp 799-805

Chambers, M., Dixon, R and Swartz, J (2007), Apollo for

Adobe Flex Developers Pocket Guide, Adobe Developer

Library, O’Reilly, Media Inc., Sebastopol, CA

Chambers, M., Dura, D and Hoyt, K (2007), Adobe Integrated

Runtime (AIR) for JavaScript Developers, Adobe Developer

Library, O’Reilly, Media Inc., Sebastopol, CA

Cohen, B (2007), “Silverlight technical articles Silverlightarchitecture overview: technical whitepaper”, MicrosoftCorporation, April, available at: http://msdn2 microsoft.com/en-us/library/bb428859.aspx (accessed December 2007).Hoppe, U., Klostermeier, F., Boll, S., Mertens, R and Kleinefeld,

N (2007), “Wirtschaftlichkeit von Geschäftsmodellen für

universitäre Lehrkooperationen – eine Fallstudie”, Zeitschrift

für E-Learning, Vol 3, No 2, pp 29-40.

Hürst, W and Müller, R (1999), “A synchronization modelfor recorded presentations and its relevance for information

retrieval”, 7th ACM International Conference on Multimedia,

Orlando, Florida, pp 333-42.

Hürst, W and Welte, M (2007), “An evaluation of the mobile

usage of e-lectures podcasts”, Proceedings of the Mobility

Conference on Mobile Technology, Applications and Systems, Singapore, September 2007.

Hürst, W., Jung, S and Welte, M (2007), “Effective

learn-quiz generation for handheld devices”, Proceedings of the 9th

Conference on Human Computer Interaction with Mobile Devices and Services (MobileHCI 2007), Singapore.

Kazoun, C and Lott, J (2007), Programming Flex 2, Adobe

Developer Library, O’Reilly, Media Inc., Sebastopol, CA

Figure 8 Lecture Recordings in the Learn Management System

Ketterl, M., Mertens, R and Morisse, K (2006a), “Alternative

content distribution channels for mobile devices”,

Microlearning International Conference on Micromedia

& eLearning 2.0: Getting the Big Picture, Innsbruck, 8-9 June

2006, pp 119-30.

Ketterl, M., Mertens, R and Vornberger, O (2007a), “Vector

graphics for web lectures: comparing Adobe Flash 9

and SVG”, Workshop on Multimedia Technologies for

E-Learning (MTEL), IEEE International Symposium on

Multimedia 2007, Taichung, Taiwan, 10-12 December 2007,

pp 389-95

Ketterl, M., Heinrich, T., Mertens, R and Morisse, K (2007b),

“Enhanced content utilisation: combined re-use of multi-type

e-learning content on mobile devices”, IEEE Multidisciplinary

Engineering Education Magazine, Vol 2 No 2, pp 61-4.

Ketterl, M., Mertens, R., Morisse, K and Vornberger, O

(2006b), “Studying with mobile devices: workflow and

tools for automatic content distribution”, World Conference on

Educational Multimedia Hypermedia & Telecommunications

EDMedia 2006, Orlando, FL, June 2006, pp 2082-8.

Lauer, T and Ottmann, T (2002), “Means and methods in

auto-matic courseware production: experience and technical

chal-lenges”, Proceedings of the World Conference on E-Learning in Corp.,

Govt., Health & Higher Education, E-Learn 2002, Montreal,

Quebec, Canada, 15-19 October 2002, pp 553-60.

Mertens, R (2007), “Hypermediale Navigation in

Vorlesungsaufzeichnungen: Nutzung und automatische

Produktion hypermedial navigierbarer Aufzeichnungen

von Lehrveranstaltungen”, PhD thesis, Universität

Osnabrück, Osnabrück

Mertens, R., Farzan, R and Brusilovsky, P (2006a), “Social

navigation in web lectures”, ACM Hypertext 2006, Odense,

Denmark, 23-25 August 2006, pp 41-4.

Mertens, R., Friedland, G and Krüger, M (2006b), “To see or

not to see? Layout constraints, the split attention problem

and their implications for the design of web lecture

inter-faces”, World Conference on E-Learning, in Corporate,

Government, Healthcare & Higher Education, E-Learn 2006,

Honolulu, HI, 13-17 October 2006, pp 2937-43.

Mertens, R., Ketterl, M and Vornberger, O (2006c),

“Interactive content overviews for lecture recordings”,

IEEE ISM 2006 Workshop on Multimedia Technologies for

E-Learning (MTEL), San Diego, California, USA, 11-13 December 2006, pp 933-7.

Mertens, R., Ketterl, M and Vornberger, O, (2007),

“The virtPresenter lecture recording system: automatedproduction of web lectures with interactive content

overviews”, International Journal of Interactive Technology and

Smart Education (ITSE), Vol 4 No 1, pp 55-66.

Mertens, R., Brusilovsky, P., Ishchenko, S and Vornberger, O.(2006d), “Time and structure based navigation in web lec-

tures: bridging a dual media gap”, World Conference on

E-Learning, in Corporate, Government, Healthcare & Higher Education, E-Learn 2006, Honolulu, HI, USA, 13-17 October

2006, pp 2929-36

Mertens, R., Ickerott, I., Witte, Th and Vornberger, O.(2005), “Entwicklung einer virtuellen Lernumgebung für

eine Großveranstaltung im Grundstudium”, Proceedings of

the Workshop on elearning 2005, HTWK Leipzig, 11-12 July

2005, pp 197-210.

Mertens, R Schneider, H., Müller, O and Vornberger, O.(2004), “Hypermedia navigation concepts for lecture

recordings”, World Conference on E-Learning in Corporate,

Government, Healthcare & Higher Education, E-Learn 2004, Washington DC, November 2004, pp 2480-7.

Red5 Open Source Streaming Server (2007), available at:http://osflash.org/red5 (accessed December 2007)

Schulze, L., Ketterl, M., Gruber, C and Hamborg, K.C.(2007), “Gibt es mobiles Lernen mit Podcasts? Wie

Vorlesungsaufzeichnungen genutzt warden”, 5.e-Learning

Fachtagung Informatik (DeLFI), Siegen, Germany, September

2007, pp 233-44.

Téllez, A.G (2007), “Authoring multimedia learning material

using open standards and free software”, IEEE International

Symposium on Multimedia 2007 Workshop on Multimedia Technologies for E-Learning (MTEL), Taichung, Taiwan, 10-12 December 2007, pp 383-9.

Vahdat, A., Dahlin, M and Anderson, T (1996), “Turning theweb into a computer”, Technical report, University ofCalifornia, Berkeley, CA

Ziewer, P and Seidl, H (2002), “Transparent teleteaching”,

19th Annual Conference of the Australasian Society for Computers in Learning in Tertiary Education (ASCILITE), Auckland, New Zealand, December 2002, Vol 2, pp.749-58.

Authoring multimedia learning material using open standards and free software

Alberto González Téllez

Departamento de Informática de Sistemas y Computadores, Valencia, Spain

Email: agt@disca.upv.es

Abstract

Purpose – The purpose of this paper is to describe the case of synchronized multimedia presentations.

Design/methodology/approach – The proposal is based on SMIL as composition language Particularly, the paper

reuses and customizes the SMIL template used by INRIA on their technical presentations It also proposes a set of free tools to produce presentation content and design focusing on RealPlayer as delivery client The integration in this e-learning platform of multimedia compositions developed following the proposed technique is also presented.

Findings – Technological support to learning and teaching has become widespread due to computers and internet

ubiquity Particularly e-learning platforms permit the any-time-and-any-place distribution of interactive multimedia learning materials There are commercial tools available to author this kind of content, usually based on proprietary formats This option has some drawbacks like license cost and software company dependency To use open data standards and free software is an alternative without these inconveniences but available authoring tools are common-

ly less productive This shortcoming is certainly important to non-technical authors and it could be solved by open source collaboration.

Originality/value – The paper presents multimedia learning material using open standards and free software.

Keywords: Multimedia, E-learning, Teaching aids, Computer software

Paper type: Research paper

Digital format learning/teaching materials are commonly

used in universities due to classroom computer

avail-ability and to the added capabilities that computer based

delivery offers compared to classic blackboard only

method In fact the classroom has been extended by

ubiquitous e-learning platforms (Sakai, Moodle, WebCT,

etc.) that impose the use of digital format to learning

content Computer authoring tools permit to create

dynamic presentations with animation effects, audio

and video clips to make knowledge transference more

effective For instance the more commonly used tation editor PowerPoint is able to produce narrativepresentations adding a speaker voice track to slides Astep forward is done by tools like eChalk (Jeschke, 2006)that allows the recording of all live activity on a penbased input device or electronic whiteboard, includingthe lecturer voice, and the delivery of recorded content toJava aware web clients

presen-Common presentations authored with office suitestools are intended to be used locally in the computerwere they are stored Web format is supported as anexport option but usually the format obtained is not well

suited for the Internet (i.e slides converted to bitmaps,

lack of streaming support, etc.) HTML extended with

Flash and JavaScript is more suitable for web delivery and

this is the nowadays general choice for web content

authors The main reason is that very good commercial

authoring tools are available and almost all clients and

platforms support this formats In this context Flash is

the part that adds multimedia support and compared to

HTML and JavaScript it is a proprietary format In spite

of being the de facto standard for multimedia on the web

(i.e YouTube is based on Flash) it has the shortcomings

of forcing authors to be linked to Adobe and to its

com-mercial decisions Portability is not a problem because

the Flash plug-in is available on Windows, Linux and

MacOSX

Open format alternatives to Flash are SVG and SMIL,

two XML compliant languages standardized by the

World Wide Web consortium (W3C) XML is a W3C

effort to enforce the definition and implantation of open

and application independent data formats SVG stands

for Scalable Vector Graphics and is defined to design

static and animated vector graphics SMIL (Synchronized

Multimedia Integration Language) (W3C SMIL site)

permits to combine and synchronize several independent

media in a presentation The presence of SVG and SMIL

on the web is nowadays clearly surpassed by Flash but

successful open source initiatives such as Firefox and

Helix can change the scenario in the future

There are two other alternatives to make the web

mul-timedia capable: ActiveX controls and Java applets

ActiveX is a Windows-only technology and it is

success-ful due to the actual domain of Windows clients on the

Internet Java applets are supported in all Java aware

platforms and are very convenient to implement small

web compatible interactive applications We make use of

applets to enrich our learning documents with interactive

simulators (González, 2003) and, as we will see later, to

include multimedia compositions into our e-learning

platform The availability of the Java plug-in for all

com-mon web clients and operating systems makes this

tech-nology a good development platform for e-learning

envi-ronments (Jeschke, 2006)

In this work we propose a technique to develop

multi-media contents for the Internet based on open standards,

particularly SMIL that has been used since many years in

the context of lecture recording (INRIA site, Yang et al.,

2001, Ma et al., 2003, Joukov, 2003, Hunter, 2001) Our

proposal is comparable to the one appearing in (Yang

et al., 2001) but it is simpler and more concrete in the

sense that all the tools and procedures required are

pre-sented and available It includes a set of free and in most

cases open source authoring tools A main goal in the

proposal is multiplatform support (particularly

consider-ing Windows, Linux and MacOSX) on the delivery and

production processes

We have been working in recent years on the utilization

of open and XML compliant formats to produce teachingcontent As a result we have developed an authoring envi-ronment to produce and to manage content based onDocbook (González, 2006, González, 2007) Until nowcontent media was limited to text and static graphicsfocusing on paper format delivery

In the academic year 2006-2007 our university startedPoliformaT an e-learning platform based on Sakai(Mengod, 2006) This has opened some working direc-tions to us One of them is based on the fact that Sakaihas chosen the IMS formats for learning content (IMSsite), particularly IMS content package and IMS QTI lan-guages Our previous decision about XML has been wisebecause Docbook content can be automatically translated

to IMS format by means of XSLT This is quite feasiblebecause Docbook is well structured and format inde-pendent A second working line is based on the possibil-ity of delivering more dynamic content (multimediacompositions) containing animations, video, audio anduser interaction

Our learning content anatomy has text as backbone,

we are classic in this respect Text combined with staticgraphics is delivered in paper format (PDF) and in webformat (HTML) Web format is extended by means ofmultimedia compositions based on SMIL, this extension

is the topic of this work Our multimedia compositionsare classified following the following increasing structur-

al complexity sequence:

1 Static image with voice narration

2 Computer animation or natural video with nized voice narration

synchro-3 Multiple media synchronized with voice narration

or lecturer video

An important decision to make when dealing with timedia on the Internet is to select the target client Webclients only support directly HTML, JavaScript andbitmap graphics (JPEG, PGN and GIF) Other contentlike vector graphics, audio and video require specificplug-ins (Rogge, 2004) and then specific formats Theauthoring of this kind of content is then strongly condi-tioned by the target client Some of the most commonmultimedia clients are:

mul-• Windows Media Player (Microsoft)

• Quick Time (Apple)

• RealPlayer (Realnetworks, Helix)

• Flash (Adobe).

• Mplayer (open source)

• VLC (open source)

Excluding Windows Media Player, only available on

Windows, all players are multiplatform Flash is with no

doubt the one that wins in terms of amount of content

published on the web RealPlayer has been shadowed by

Windows Media Player but it is still alive (release 11.0

has been delivered on November 2007) and it has also the

interesting feature of having a linked open source

initia-tive named Helix (Helix site) Helix was started by

Realnetworks and includes several open source projects

including several players, the Helix server and streaming

formats RealPlayer supports SMIL 2.0, briefly described

on the next section, that allows composition structures

and user interaction capabilities that surpass the ones

offered by proprietary formats (Pihkala, 2006) In

(Bulterman, 2003) SMIL is proposed to encode peer-level

annotations that allow dynamic expansion of multimedia

presentations The counterpart is that SMIL players use

to have limitations on the features supported and even

errors (Eidenberg, 2003) And last but not least there is

no media content standardization for SMIL

After balancing pros and cons we have chosen SMIL as

the language to create our multimedia learning/teaching

material The purpose of SMIL is to define the spatial and

temporal integration of several media in a multimedia

composition and to establish the user interaction with

the composition Previous considerations indicate that it

is advisable to choose a target client among the available

SMIL aware clients This will define precisely the media

formats to use and the SMIL specification portion that is

properly supported and then reliable RealPlayer is our

choice because it is available on Windows, Linux and

MacOSX and it supports an extensive subset of the SMIL

2.0 specification In spite of being a proprietary player it

has the interesting feature mentioned previously of being

related to the open source project Helix RealPlayer

sup-port, among others, the following formats:

• Text: Plain text and Realtext

• Images: JPEG

• Audio: Realaudio

• Animations: Realvideo

• Natural video: Realvideo

Realaudio and realvideo are specially designed for

streaming delivery and are the more convenient audio

and video formats to get good synchronization results in

SMIL constructs played by RealPlayer

SMIL (Synchronized Multimedia Integration Language,

W3C SMIL site) is the XML W3C standard intended to

define the synchronized integration of text, graphics,audio and video in multimedia presentations SMIL per-mits to define the spatial and temporal composition ofseveral media and the interaction between the mediainside the presentation and with the presentation and theuser Because of being XML compliant only a plain texteditor is required to create SMIL documents by hand and

it is also straight forward to generate them automatically.SMIL document structure is similar to HTML; there is

a root element <smil> with two children elements

<head> and <body> The <head> element is the ment header and contains several kinds of metadata ele-ments The most important one is <layout> that definesthe spatial regions on the presentation as shown forinstance in Figure 1 The element defines the features(background color, size, etc.) of the main presentationpanel Element also includes the definition of spatialregions that will contain the presentation media Everyregion is defined by a <region> element that sets its loca-

docu-tion and size, if also assigns a unique identifier (id

attrib-ute) to the region in order to be able to make references

to it from the content part of the document The headersection can also include descriptive metadata in <meta>elements that will permit the document inclusion in aautomatically managed content repository

After the header we have the <body> element thatincludes the references to the media shown in the presenta-tion and their spatial and temporal locations Every media isincluded by means of a media element like: <text>, <img>,

<audio>, and <vidio>, using the attribute src that specifies

the location path of the media file Spatial location is

defined by means of the region attribute that is set to a region identifier defined by the id attribute in <region> elements.

Time behavior is defined by means of a nested sition of <seq> and <par> elements that define sequen-tial and parallel playing, respectively Inside a <seq> or

compo-<par> element we can have <switch> elements intended

to select from a media collection the ones that complywith some conditions (i.e presentation language) Everymedia element has attributes that establish its timing

behavior: begin for the start time, end for the end instant and dur for the media playing duration.

SMIL also has links implemented with <a> elementthat allow user interaction with the presentation Linkscan point to content locations (as in HTML) and to tem-poral locations Temporal links destinations are defined

by means of <area> elements included in temporal mediaelement (i.e locations in a video clip) In Section 4.2 wedescribe how this is performed in our compositions

The French research institution INRIA (InstituteNational de Recherche en Informatique et en

Automatique) has chosen SMIL as the format for their

technical presentations (INRIA site) Our interest in

SMIL began after noticing how well these presentations

are reproduced with RealPlayer in Windows, Linux and

MacOSX It has been possible to analyze these

presenta-tions and to reuse their template because SMIL is an open

format and INRIA has not defined privacy restrictions

INRIA presentations are about one hour long and they

are made using two designs; the first one has a root

presen-tation that links to partial presenpresen-tations that are several

minutes long The second design includes the whole

pres-entation in one SMIL document Our prespres-entations are

con-ceived as small pieces in a text backboned lecture and then

their length will be up to 5 or 10 minutes, then the second

template is more adequate The spatial design of the SMIL

template selected defines the regions shown in Figure 2:

• Title: It includes presentation titles.

• Slides: It shows presentation slides (i.e JPEG

images): It can include sub regions to show different

types of media inside a slide

• Temporal link index: It includes temporal links to

presentation time locations

• Lecturer: It shows a narrative lecturer video.

• Logo: It includes the institution logo.

The presentation timing structure has a root <par>

element that contains the slides sequence, the menu

links and the narrative lecturer video The sequence of

slides is made putting the elements (<img> or <video>)

corresponding to every slide inside a <seq> element If a

slide is made up of different media then it corresponds

to a <par> element that includes all the media The ing control of the slides sequence is implemented by

tim-means of the dur attribute inside every slide element.

The narrative lecturer video covers the whole tion and it is encoded in Realvideo The menu linkspoint to <area> locations in the lecturer video that arealso synchronized with the timing defined in the slidesequence A more detailed analysis can be performed bylooking at the SMIL source of a presentation This can be

presenta-done by clicking at the clip source entry in the floating

menu when a presentation is played with RealPlayer

We have found INRIA technical presentations a goodexample of the SMIL capability to create multimediapresentations Our customization of their design tem-plate, in order to elaborate our multimedia material, isdescribed in Section 4.1

After establishing the technology to use we have to select

a set of good enough authoring tools to produce themedia we are going to include in our multimedia presen-tations We have preference for free, open source andmultiplatform tools In Table 1 we propose a set of freetools available on Windows and Linux Real ProducerBasic is a free product from Realnetworks that permits tocapture and to convert audio and video to Real formats.The converted streams can not be edited inside RealProducer Basic Therefore the media edition, if required,should be performed before conversion

Impress is the OpenOffice presentation editor and wehave found it good enough to produce teaching content.CamStudio and xvidcap are screen video recorders both ofthem open source They allow producing demos or ani-mations by recording screen videos (i.e Impress anima-tions) Finally JEdit is an open source text editor written

in Java with several extensions One that is particularly

Figure 2 INRIA technical presentation

Table 1 Authoring tools Tool type Windows Linux

Audio capture Real Producer Basic Real Producer Basic Video capture Real Producer Basic Real Producer Basic Screen video cap CamStudio xvidcap

Animations Impress Impress SMIL editor JEdit JEdit

Figure 1 Layout section example

relevant here is XML extension that is very adequate to

edit SMIL documents

The multimedia compositions that we are interested

on are mentioned in Section 2 It is very straight forward

to reuse INRIA SMIL template to create these types of

multimedia compositions To produce a type A

composi-tion we only have to make the following changes in the

SMIL template:

• Delete the link menu

• Replace the <video> narration by an <audio>

element

• Delete the <area> elements in the narration element

• Reduce the slide sequence to only one <img>

element

A type A composition is converted into a type B

compo-sition by replacing the <img> element in the slice

sequence by a <video> element To produce a type C

composition we only have to define the slide sequence

and the synchronization between the link menu, the

nar-rative video or audio and the slide sequence A detailed

explanation is given in Section 4.2 An example of a type

C composition is shown in Figure 3 The lecturer video

has been replaced by an audio track and a GIF animation

in order to reduce the amount of storage or network

bandwidth required

The hardware equipment required is very accessible and

it is compounded by a PC, speakers, a microphone and a

digital video camera The production process has two

main steps:

1 Content creation (slides, audio clips, video clips, etc.)

2 Content integration using SMIL

A common content is a slide sequence created by apresentation editor like PowerPoint or Impress.PowerPoint allows exporting a presentation in JPEG orPNG format in such a way that every slide is exported as

a JPEG or PNG file Impress has the HTML exportationoption that also exports every slide as a JPEG file

If a slide includes animation the previous technique isnot adequate An animated slide can be captured usingone of the screen capture utilities proposed in Section 4.The capture process will generate a video clip that will beconverted later into realvideo format in order to get agood result on Realplayer This conversion is performed

by means of Real Producer Basic that supports severalvideo formats as input, like uncompressed AVI and DV.Natural video obtained with a video camera (webcam,camcorder, etc.) can also be included by performing thesame conversion as in screen capture clips We have foundthat a target bandwidth between 256 and 512 kbps forrealvideo gives satisfactory results for both screen recordsand natural video

The presentation narration is produced by recording anindependent audio or video clip for every presentation item.This can be done by means of the capture capability of RealProducer Basic that directly generates realaudio andrealvideo formats The inconvenient is that the capturedclip can not be edited If audio and video edition is neededthen it is required a capture utility (i.e Nero 7) that gener-ates a Real Producer Basic compatible format When all theindividual clips are available in realaudio or realvideo for-mats they are glued into a single narration by means ofrmeditor console utility included in Real Producer Basic.After having obtained all the presentation contentitems and the presentation narration, the next step is tocustomize the SMIL template (i.e using JEdit) Thecustomization process has two dimensions:

1 Spatial Definition of the presentation layout (slide

region, link region, title region, etc.)

2 Temporal Definition of temporal behavior and

syn-chronization (slide durations and time link locations).Temporal design is the most complex and it is per-formed in three steps:

1 Get the duration of every individual narration clip(tslide,dur) This is indicated by Realplayer when play-ing the clip

2 Obtain the sequence of slide starting times (tslide,start).This can be computed from tslide,dur values using aspreadsheet

3 Design the temporal link index by grouping slidesand getting the location of every anchor in the pres-entation timeline from tlink,startvalues

Figure 3 Customization example