Alberto González Téllez
Dept. of Computer Engineering, Polytechnic University of Valencia, Spain
Keywords: Content reuse, Slides, Docbook, SMIL, XML, XSLT.
Abstract: Slide presentations have become a common method used by lecturers to prepare and deliver teaching
content. The amount of effort invested in authoring this material is usually very high and then it is desirable
to be able to reuse it. Unfortunately commonly used slide editors offer very little reuse capabilities. In this
paper we present a method of authoring teaching slide presentations that has the advantage of being able to
automatically port slide contents into other delivering formats (textbooks, synchronized multimedia
presentations, e-learning platforms, etc). The technique is based on using an XML compliant language with
available tools that are user friendly and free.
Slide presentations have become a common method
to make lectures more effective and efficient (Alley,
2005). To produce good slide presentations requires
carefull design and working time (Alley et al.,
2005), in order to avoid countereffects like those
exposed in (Parker, 2001). The amount of effort
invested by lecturers producing their presentations
is, after some years, quite big. But as a result they
find themselves with hundreds of slides packed in
proprietary format. In order to take the maximum
profit from this investment it would be desirable to
be able to automatically reuse this material as much
as posible (several final formats, topic units, exercise
collections, multimedia presentations, etc).
A way to separate content from format is to use
XML compliant markup languages (W3C rec 04,
2004). This need appeared for instance on the web
and this was the reason why the W3C developed
XML specification. A widespread XML language
oriented to book production is Docbook (Docbook,
2006, Walsh, 2003, Stayton, 2005). It was born in
the seventies under the SGML domain with the aim
of being used to document computer products,
particularly the UNIX operating system. Nowadays
Docbook is available in both SGML and XML, and
it is being used as a general documentation tool. A
dialect from Docbook is Slides that is oriented to
slide presentations authoring. Slides markup reuses
some of Docbook content elements and it basically
changes the hierarchical structure. There are also
specific XSLT style sheets (Docbook, 2006) to
generate HTML and printing format from Slides
Besides of being technically feasible to
automatically reuse content it is important to have
productive and easy to use authoring tools. XML
editors tend to be much less friendly than slide
presentation editors because they usually show the
text markup instead of a WYSIWYG preview.
XMLMind XML Editor or XXE (XXE site, 2006)
does not follow this line. This editor is intended to
be used by non technical authors that are interested
in XML capabilities that are not available in
proprietary formats.
The content of the paper is organized as follows:
first we describe briefly Slides markup and style
sheets. Then we propose some authoring and format
generation tools we have found very effective.
Finally we describe two slides reuse applications:
textbooks and multimedia presentations.
As stated previously our proposal to author reusable
slides, is based on XML, particularly on the
Docbook dialect Slides. Norman Walsh defined this
language to be used for authoring slide presentation
in a similar style as other proprietary tools.
Slides markup language is specified by a public
DTD and it includes a subset of Docbook content
elements. The hierarchical elements differ from
Docbook to make them suitable for presentations.
González Téllez A. (2007).
In Proceedings of the Third International Conference on Web Information Systems and Technologies - Society, e-Business and e-Government /
e-Learning, pages 366-371
DOI: 10.5220/0001263703660371
The root element is slides and it has as children:
slidesinfo and a sequence of one or more foilgroup.
The slidesinfo element includes the next element
Title and subtitle.
Author information.
Meta information (release, revision, etc).
Some of slidesinfo elements are generated by
default by the style sheets (i.e. title and author). To
show other elements style sheets customizations is
required [6].
A foilgroup element contains a sequence of one
or more foil elements. A foil element corresponds to
a slide and then can contain a title and a sequence of
one or more content elements.
The set of content elements available inside a foil
is a small subset of content elements defined inside a
section element in Docbook. These content elements
can be classified as:
Text block and text inline.
Lists of items (ordered, unordered and
Figures and media objects.
Links (local, external and web).
A speakernotes element is also available to
include lecturer guide notes that are not intended to
appear in the presentation.
When a Slides document is completed it has to be
transformed into presentation format. To accomplish
this there are available XSLT style sheets (Docbook,
2006) to generate chunked HTML and XSL-FO.
This later format can be transformed into WordML,
OpenDocument, PDF or RTF in order to get a final
printed format. Slides style sheets can also be
customized to get a satisfactory presentation format
in combination with CSS.
To apply the style sheets to the slides document
we need an XSLT processor. There are many
publicly available in both native code like libXML
(Libxlst site, 2006) and Java like Saxon (Saxon site,
2.1 Theory and Exercises
Contents in slides can be classified as theory and
exercises. In order to be able to clearly distinguish
and manage both types of elements it is convenient
to be able to identify them. This can be
accomplished with the next slides markup
o The role attribute of foil elements is restricted
to the values: theory, exercise.
o Inside a foil, with its role attribute set to
“exercise”, enunciate elements are marked
setting its role attribute to “enunciate”. This
has to be done only to direct foil descendants.
With this simple customization we can easily
separate theory from exercises. Applications of this
capability are for example: publishing an exercise
book, create an exercise repository, etc. In order to
make exercises more effective solutions can be
filtered when the presentation format is generated.
The slides DTD and XSLT style sheets are the
technical basement of the proposal. In order to make
it attractive to final users it is required the
availability of user friendly tools that hide as much
as possible XML related details.
The tools required are an editor and a format
generator. After looking for sometime among the
many available free and commercial products we
have chosen XXE editor (XXE site, 2006) and XFC
XSL utility (XFC site, 2006). They are commercial
products from XMLMind but there is a free license
for both tools that offers enough functionality to
fulfill our requirements.
3.1 Proposed Editor
XXE stands for XMLMind XML Editor, it is Java
written general purpose and friendly XML editor.
XXE is specially customized to Docbook, Slides and
XHTML. There are two views of document content,
one shows the XML tree structure and the other is a
pseudo WISYWIG view, based on CSS, that is
proven to be very comfortable and productive. This
friendlier preview is shown in figure 1.
The application has the common main menu and
two toolbars. The one below the main menu gives
quick access to the most common operations. Below
the quick access toolbar there is the navigating
toolbar that permits quick location of:
XIncluded elements.
Elements in the document element subtree
of the active element.
Sibling elements and ancestor/descendant
The main panel is divided into the editor panel
and the tools panel. This later panel permits:
Attribute setting.
Spell checking.
Searching and replacing.
Selecting especial characters.
Locating validity errors.
Figure 1: WISYWIG like preview in XXE.
There are two license modalities for XXE:
standard and professional. The standard edition is
free and it has been artificially limited in
functionality. The most relevant features not
available are:
Schema based validation.
Documents can not be saved to FTP and
WebDAV servers.
It could not be deployed using Java Web
XSL-FO plug-ins do not work inside XXE.
None of these missing features is relevant for the
proposed application because we use DTD
validation, our content files are stored in the local
file system and we generate format from another
application. Then standard edition is the sensible
3.2 Proposed Format Generator
The tool we propose to generate presentation format
is XFC (XMLMind Format Converter) XSL utility.
It is also written in Java and it has a free personal
XFC is in fact an XSL-FO processor that
generates WordML, OpenDocument and RTF. The
XSL utility also allows generating other formats like
HTML and PDF. It includes also Saxon XSLT
processor and FOP converter from XSL-FO to PDF.
XFC XSL utility is customized to generate
HTML, chunked HTML and printing formats from
Docbook and Slides documents. The corresponding
XSLT style sheets are included in the distribution.
The transformation process is described in figure 2.
The tool is completely customizable the only
restriction is that Saxon can not be replaced. New
transformations can be defined by setting the style
sheet, defining its parameter values, passing the
transformation output to a external program (i.e. an
XSL-FO processor or a script) and setting an output
format viewer (i.e. web client, PDF viewer, etc).
Figure 2: Format generation with XFC XSL utility.
3.3 Collaborative Authoring
Docbook and Slides documents are plaintext, this
allows to use concurrent versioning systems like
CVS to collaborative authoring teaching material.
We use a CVS server on Linux and the java client
SmartCVS which is very user friendly and free for
non commercial use.
Next we are going to describe two interesting
applications of the proposal capability of flexible
reuse. First we will port slides content into textbook
by means of transforming slide presentations into
Docbook. The second application consists of
including slides content into synchronized
multimedia presentations based on SMIL (W3C rec
13, 2005).
WEBIST 2007 - International Conference on Web Information Systems and Technologies
4.1 Reusing Slides Into Textbook
When a lecturer begins with a subject he or she
needs to write a sketch of content in slide format
normally based on preexisting information sources.
In this first approach we propose to use Slides
language as the source.
From the Slides source a chunked HTML format
can be generated directly just applying Slides HTML
style sheets. A printed format, more convenient for
students can be obtained translating Slides to
Docbook and then applying Docbook XSL-FO
styles sheets and an XSL-FO processor to get final
printable format. The XSLT style sheet required to
convert Slides to Docbook is very simple because it
has to change only the document structure. We can
also take advantage of the role attribute that is
always present in Slides and Docbook elements and
does not have predefined values.
Figure 3: Slides to textbook conversion.
A special attention is paid to exercises. We
implement exercises in Docbook by means of
qandaentry elements. Originally this element was
defined to write FAQ collections but it is very well
suited to exercises also. Qandaentry elements are
wrapped in qandaset elements. On the other side
qandaentry elements have as children one question
element and one or several answer elements.
Then the conversion from Slides to Docbook is
performed in the next way:
o Slides root slides is converted to Docbook
root chapter.
o Slides foilgroup elements are converted to
Docbook section elements.
o Slides foil elements with its role attribute set
to “theory” are converted to Docbook
simplesect elements.
o Slides foil elements with its role attribute set
to “exercise” are folded in a qandaset element
and converted to Docbook qandaentry
Foil children with attribute role set to
“enunciate” are folded inside a question
Other foil children are folded inside an
answer element.
The conversion process is shown graphically in
figure 3.
When the authored material has reached a critical
mass then the source can be changed to Docbook
Focus now will not be on the presentation summary
but on producing a complete textbook. From that
moment the Slides document will be generated from
the Docbook document by means of custom XSLT
style sheets. The Slides document content will be a
subset of Docbook content and the lecturer has to be
able to define this subset. Content selection will
follow the next criteria:
Text selected from:
o Inline marked text (i.e. emphasis).
o <literallayout> elements
o <programlisting> elements.
Media as block and inline elements.
Common block text elements (i.e. paragraphs) are
filtered in such a way that only inline marked text is
selected. The most common inline text element is
<emphasis>. Also block elements inside a literal
layout or program listing elements are included in
the Slides document.
Tables content and media elements are not
filtered as they are supposed to be relevant in the
presentation format.
As before exercises are specially considered. The
textbook can be used as exercises repository then a
mechanism is required to select what exercises will
be included in the presentation format (i.e. as
examples), this can be accomplished by using the
role attribute of qandaentry elements. Selected
exercises will generate a foil element according to
the previously described design.
XLST style sheets required to perform Slides to
Docbook and Docbook to Slides conversions are
available at (Slides2Docbook site, 2006) and
Docbook2Slides site, 2006), respectively.
4.2 Porting Slides to Multimedia
In order to keep the student attention alive it is
convenient to make class presentations as much
dynamic as possible. A way to achieve this is to
combine several media elements, not only text and
graphics, into a timeline. An open XML standard
designed specifically with this goal in mind is SMIL
that stands for Synchronized Multimedia Integration
There are several SMIL clients (RealPlayer,
Ambulant, QuickTime, etc) and authoring tools
(LimSee2, GoLive, GiNS, etc). The fact that SMIL
is a compliant XML language allows other
alternatives to create SMIL presentations: a plain
text editor or automatic generation through XSLT.
The later alternative is particularly attractive in our
case because we can use as input to the XSLT
transformation our slides documents.
Slides language is capable of including media
objects. An example is to add to every foil element
an audio stream file that will expand the slide
content. Other audio and video object can be added
in order to make the presentation more attractive.
All the previous objects have to be located in the
screen and scheduled in time. SMIL allows to define
a layout on the screen, decomposing it into regions,
and to schedule media objects in time in sequence
and/or in parallel.
In spite of SMIL being a standard, there are many
details that have to be considered to produce SMIL
presentations: the way text is included and
formatted, the media formats supported, the types of
linking mechanisms, etc. This means that a
particular client has to be selected in order to
guaranty that everything will work. We choose
RealPlayer because it is a widely spread client and it
has a considerable SMIL 2.0 support (Real
Networks, 2004). Other interesting features of this
client are: it is free, it is multiplatform and it has
complementary authoring tools also free and
multiplatform like RealProducer Basic.
An institution that has choosen SMIL as their
presentation tool is INRIA (Institut National de
Recherche en Informatique et en Automatique) in
Grenoble, France. Their interest in SMIL is proven
by the fact that they have developed an open source
SMIL authoring tool called Limsee. A good example
of the SMIL capabilities can be seen at the INRIA
presentations site (INRIA, 2006)
We have proposed a method to author teaching slide
presentations based on XML, particularly on the
Slides language. Our experience using this method
has shown that losing some user friendly features,
present in more conventional slide editors, is paid
off by the ability of reusing content. In the context of
teaching topics at graduate level and above, this
benefit is particularly relevant.
An important aspect of the proposed method is
the authoring tools, particularly the XXE editor that
has proven to be very productive to non XML
As present and future work we are developping
Docbook and Slides customizations, at markup and
style sheets levels (González, 2006), in order to
better fulfill teaching material authoring
requirements. We also are working on porting Slides
document to SMIL as much automatically as
possible using RealPlayer as the target client.
Other interesting areas for future development
will try to take advantage of other XML related
capabilities like content repository management
(Dhraief, 2001) and to port content to e-learning
platforms (González, 2005, Mengod, 2006).
Alley, M., 2005. The Craft of Scientific Presentations.
Critical Steps to Succeed and Critical Errors to Avoid
4th edition, Springer.
Alley, M., et al. 2005. Pilot Testing of a New Design for
resentation Slides to Teach Science and Engineering.
35th ASEE/IEEE Frontiers in Education Conference.
Indianapolis, IN, 2005
Parker, I., 2001, Absolute PowerPoint, The New Yorker,
28 May.
W3C rec 04, 2004. XML specs.
Docbook, 2006. Docbook main sites.
Walsh, N., 2003. DocBook: The Definitive Guide.
Stayton, B., 2005. DocBook XSL: The Complete Guide.
Sagehill Enterprises.
XXE site, 2006.
XFC site. 2006,
Libxslt site. 2006,
Saxon site. 2006,
González, A, 2005. Entorno web para la generación y
corrección automatizada de exámenes basado en XML
WEBIST 2007 - International Conference on Web Information Systems and Technologies
y Java, I Congreso Español de Informática, Granada,
González, A, 2006. Teaching Document production and
management with docbook. WEBIST 2006. Setúbal,
INRIA, 2006. Multimedia presentations site
W3C rec 13, 2005. SMIL 2.1 specs
Rutledge, L., et al., 1999. Anticipating SMIL 2.0: The
Developing Cooperative Infrastructure for Multimedia
on the Web, in Proc. The Eighth International World
WideWeb Conference (WWW8), Toronto, Canada.
Real Networks, 2004, Real Networks Production guide,
Dhraief, H., 2001. Open Learning Repositories and
Metadata Modelling, in Proc. of Int. Semantic Web
Working Symposium, Stanford, USA
Mengod, R. PoliformaT, the Sakai-based on-line campus
for UPV - history of a success. 5th Sakai Conference,
Vancouver, BC, Canada 30 May - 2 June 2006
Slides2Docbook site, 2006,
Docbook2Slides site, 2006,