Long but good source for handheld devices.

Helsinki University of Technology 12th May 2005

Telecommunications Software and Multimedia Laboratory

T-111.590 Research Seminar on Digital Media

Digital TV Services in Handheld Devices

A Compound Document Format for Handheld

Devices

Mikko Pohja

44305U

A Compound Document Format for Handheld

Devices

Mikko Pohja

HUT, Telecommunications Software and Multimedia Laboratory

mikko.pohja@hut.fi

Abstract

Handheld devices have recently gained access to the Internet. The devices differ

a lot from traditional desktop computers what comes to, for instance, browsing

the Web. Several parties have defined specifications, which facilitates creating

Web documents for handheld devices. In this paper, we have defined document

profile, which combines many document formats targeted for handheld devices.

The result is compound document profile, through which can be realized multimedia

presentations in handhelds. The profile has two configurations, from which

one can select depending on target device's resources.

1 INTRODUCTION

Nowadays, handheld devices have access to the Internet. However, there are some

limitations in web browsing with restricted devices. For instance, screen size and processing

power are notably smaller than in desktop computers. To browse current web

pages with handheld devices can be difficult or even impossible. Several consortiums

have addressed this problem by defining markup language specifications for handheld

devices.

World Wide Web Consortium (W3C) has published XHTML Basic and CSS Mobile

Profile for handheld devices. W3C has also defined Scalable Vector Graphics

(SVG) recommendation to present vector graphics in XML format. Through SVG the

graphics can be presented properly even though screen sizes varies a lot between different

devices. Recently, W3C started a Compound document formats working group1.

Its aim is to specify the behavior of some format combinations, namely XHTML, SVG,

Synchronized Multimedia Integration Language (SMIL), and XForms combinations.

Open Mobile Alliance (OMA) has introducedWireless Application Protocol (WAP)

2.0 as a declarative environment for handheld devices. XHTML Mobile Profile (XHTMLMP)

is part of Wireless Markup Language (WML) 2.0, which is the markup language of

1http://www.w3.org/2004/CDF/

1

WAP 2.0. Vodafone has made a proposal how to integrate XHTMLMP and SVG

Tiny2.

In this paper we review what kind of requirements handheld devices set to a compound

document format and what kind of format is suitable for handheld devices.

Aspects, which were taken account when defining requirements, were representation

of text, images, and 2D graphics, temporal layout of a document, and user interaction.

We defined language profile, which consists of existing and proposed formats of

markup languages. In addition, the profile is divided to two configurations depending

on device's resources. That is, most restricted devices are not able to handle all the

graphical and temporal features of the defined profile.

The rest of the paper is organized as follows. In next Section is discussed the

markup languages used in this paper. Section 3 defines the requirements of the language

profile and Section 4 introduces the implementation of the profile. Future work

is defined in Section 5. Finally, Section 6 gives the conclusions of the work.

2 BACKGROUND

Hypertext Markup Language (HTML) is used to create structured hypertext documents

on the World Wide Web (WWW). HTML is based on Standard Generalized Markup

Language (SGML). The Extensible HyperText Markup Language (XHTML) is a family

of document types and modules that reproduce HTML, reformulated in eXtensible

Markup Language (XML). XHTML is the successor of HTML. Several versions of

XHTML have been developed.

For handheld devices, there are XHTML Basic and XHTMLMP profiles. XHTML

Basic includes the minimal set of XHTML modules defined in Modularization of

XHTML (Altheim, 2001). XHTMLMP is a superset of XHTML Basic. In addition

to XHTML Basic elements, XHTMLMP contains elements relating to forms, presentation

and style element and attribute. Also, XHTML 2.0 (Axelsson, 2004), which is

under development at W3C, is suitable for handheld devices. XHTML 2.0 is successor

of the HTML languages, but it is not backward compatible with them. The origins and

relationships between markup languages are depicted in Figure 1.

Cascading Style Sheets (CSS) is a method for adding style to Web documents.

CSS has also various levels and profiles. CSS Mobile Profile is for handheld devices.

OMA has defined a WAP CSS profile, which is superset of CSS Mobile Profile. WAP

CSS has the same properties as W3C's CSS Mobile Profile and in addition few WAP

specific properties (WAP Forum, 2002).

SVG suits well to present 2D graphics on the WWW. The advantage of vector

graphics is that it does not lose its accuracy when zooming. That is, graphics can be

adjusted to all screen sizes. Integration of SVG to XHTML is vital to enable content

usage in all devices. Currently, there are several SVG specifications defined by W3C.

For mobile devices, there are two profiles, namely SVG Tiny and SVG Basic. Tiny is

2XHTMLMP+SVGT, http://lab.vodafone.com/public/XHTMLMP-SVGT-Recommendations.html

2

(Forum Nokia, 2003)

Figure 1: Relationship between markup languages.

for restricted mobile devices whereas Basic is targeted for higher level mobile devices.

(Capin, 2003)

3 REQUIREMENTS

The definitions of the requirements for the language profile are based on the ones used

in Honkala et al. , 2004. The paper defines language profile for multimedia XML

language. The profile should cover following aspects: spatial and temporal layout,

support for media objects, and support for user interaction (i.e., links, user input, and

submission). The requirements are described in Table 1.

Defining spatial layout is vital in Web documents. There are two options to do it.

One is XHTML and other is SMIL. SMIL provides only absolute position for laying

out, whereas XHTML provides also flow layout. For hanheld devices, there are several

XHTML profile options to select (e.g., XHTMLMP, XHTML Basic, or XHTML 2.0).

In addition to layout, also text and images can be handled by both SMIL and XHTML.

2D graphics can be represented as any image in a SMIL or XHTML document. In

addition, it can be represented through SVG.

Temporal layout can be defined either by SMIL (Ayars, 2001) or Timesheets (ten

Kate et al. , 2000, Jalava et al. , 2005).

W3C has recently published a working draft of use cases and requirements for

compound documents by reference (Appelquist et al. , 2005). The document defines

requirements for a profile, which comprises XHTML, CSS and SVG. The profile sup-

3

Table 1: General requirements of the compound document format.

Requirement Description

Spatial layout Defines positions of media elements

in a presentation.

Text Handling of a text flow in a presentation.

Images Handling of images in a presentation.

2D graphics Method to handle and display 2D

graphics in a presentation.

Temporal layout Synchronization of media elements of

the presentation.

User interaction

Links Hyperlinks of a presentation.

User input Method to retrieve user input.

Submission User input must be submitted to

server.

ports presentation of rich multimedia content. Some of the requirements are also used

in this paper for requirements of compound documents in handheld devices. The requirements

are discussed in next Section.

4 IMPLEMENTATION

In this section, we discuss the proposed language profile. First, we define with what

technologies the requirements set above can be fulfilled. Second, we create two con-

figurations from the profile to adjust it for different handheld devices.

4.1 Technologies

In the proposed format, we chose XHTML and CSS to define the spatial layout instead

of SMIL. XHTML provides more options to define the layout. In XHTML, elements

can be laid out absolutely, relatively, or with the flow. Also, XHTML is particularly

meant for textual layout, whereas SMIL is rather for positioning and synchronizing

media objects. Laying out the images can be realized easily with both technologies.

>From several XHTML versions, we chose XHTML 2.0. XHTML 2.0 is a novel

solution, which provides generally useful set of elements, which can also be handled

by handheld device. XHTML 2.0 corrects errors and deficiencies identified in earlier

versions of the HTML. For instance, in XHTML 2.0, the structure of document

can be defined more strictly. A document can be divided to sections and subsections.

4

Table 2: Technology resolutions for the compound document format.

Requirement Description

Spatial layout XHTML 2.0 + CSS

Text XHTML 2.0

Images XHTML 2.0

2D graphics SVG

Temporal layout Timesheets

User interaction

Links XHTML 2.0

User input XForms

Submission XForms

The headers are decided automatically according to embedding levels. Also, many

elements and attributes have more general usage, which makes the language more versatile.

Finally, it uses XForms as a form technology, which offers a lot improvements

compared to current HTML forms.

Temporal layout can be realized either by SMIL (Ayars, 2001) or by Timesheets

(ten Kate et al. , 2000, Jalava et al. , 2005). SMIL has some limitations (ten Kate et al. ,

2000, van Ossenbruggen et al. , 2003) to be used as a common multimedia declaration

language for all XML documents. Especially, when a document consists of several

XML languages. In addition, when using SMIL the layout is defined by SMIL, too.

Timesheets works the other way around. It is integrated into a host language's layout

system, which makes it possible to control single elements inside a document. In short,

Timesheets assigns temporal relations between elements and sets styles for elements

in temporal manner. Timesheets was selected to describe the temporal layout since it

integrates better with compound documents.

SVG was chosen to represent 2D graphics. SVG's advantage is that it supports

portability. Single SVG graphic can be used in devices with very different characteristics

(e.g., screen size). There are two SVG version options for handheld devices. The

selection between SVGT and SVG Basic depends on configuration.

The user interaction is realized by XHTML 2.0. Any XHTML 2.0 element can

be a hyperlink through href attribute. XHTML 2.0 contains XForms module, through

which validated entries and server submission can be realized.

The technology resolutions for the requirements are summarized in Table 2.

To combine all the technologies together, we have defined some technology requirements

for the format. W3C's working draft of use cases and requirements for

compound documents by reference (Appelquist et al. , 2005) was instrumental in defining

the requirements. The requirements are discussed in Table 3. XHTML 2.0 is used

as host language and it embeds both SVG and XForms elements. SVG is compounded

by reference, but XForms by inclusion as usual in XHTML 2.0. User must be able to

5

Table 3: Technology requirements of the compound document format.

Requirement Description

Compound SVGT by reference. Use XHTML object element to embed

SVGT documents.

Compound XForms by inclusion. XForms elements are mixed with XHTML

at element level as defined in XHTML 2.0

specification.

XHTML 2.0 is used as a host language.

SVGT and XForms are combined with

XHTML 2.0.

Support for user interaction. Event and focus management between components

should be supported.

Must support grid, flow, overlapping

layouts.

Elements may be positioned absolutely or

with the flow. In addition, they may overlap

each other.

Must define transparency support

for SVG backgrounds

Support for static SVG background images.

interact with all the components of a document. That is, there has to be way to move

focus between components, as well as events must flow from a component to another.

Specific definitions of event and focus management are left as future work.

Laying out is critical part of combining languages. In the proposed format, the

elements can positioned either absolutely or with the flow of elements. In addition,

they may overlap each other. It must be possible to layout and render elements from

different languages on top of each other. Finally, SVGT graphics must be able to use

as background images in a presentation.

4.2 Configurations

Handheld devices have different characteristics. In this paper, we defined two configurations

of the language for different devices. One is for mobile phones with restricted

resources and other is for more powerful devices like PDAs. In compact configuration

we have to go short of certain features to make it possible to use the profile in restricted

devices. The comparison of selected technologies is discussed in Table 4.

The compromises were done in the areas of 2D graphics, temporal layout, and form

technology. For restricted configuration we chose SVG Tiny profile, while full configuration

includes SVG Basic profile. Both profiles are meant for handheld devices as

discussed in Section 2. Possibility to define temporal layout was left out in restricted

configuration. Finally, restricted configuration uses XForms Basic (Dubinko & Raman,

2003) as form technology, whereas full configuration uses XForms Full. The

differences of the profiles are the following features. XForms Basic Profile processors

6

Table 4: Comparison of the two configurations.

Requirement Restricted Full

Spatial layout XHTML 2.0 + CSS XHTML 2.0 + CSS

Text XHTML 2.0 XHTML 2.0

Images XHTML 2.0 XHTML 2.0

2D graphics SVG Tiny SVG Basic

Temporal layout None Timesheets

User interaction

Links XHTML 2.0 XHTML 2.0

User input XForms Basic XForms

Submission XForms Basic XForms

may support only XML Events Basic, XForms Basic Profile processors may implement

a subset of an XML Schema processor, and they may treat some XML Schema

datatypes as string.

5 FUTURE WORK

A major detail, which we left as future work, is focus management. It is common

problem for all compound document formats. The problem arises if there is not a

pointing device in use. User must use, for instance, arrow keys instead. The navigating

order of the focus points of a document, the rules how to jump from part of a document

to another, and handling of dynamic cases have to be defined.

Another future work is a reference implementation. By implementating an user

agent, which supports both of the configurations defined in this paper, we could verify

our design of compound document format for handhelds.

6 CONCLUSIONS

Handheld devices can browse the Web nowadays. However, the Web documents are

not often suitable for handhelds. We have defined general requirements for a handheld

device language formats. According to the requirements, we defined a compound

document format, which utilize several existing and proposed markup languages, to

be to used as a markup for restricted devices. Since handheld devices have various

resources, we defined two configurations, which differs by their graphical and multimedia

features. The limited configuration is meant for small mobile phones, whereas

full configuration is meant for PDAs and equivalent devices.

Focus management should be still defined for the format. This is a common problem

for all compound document formats and it should be solved generally. Also, we

7

would need an user agent, which supports the proposed format. With the user agent,

we could verify our results.

REFERENCES

Altheim Murray. 2001 (April). Modularization of XHTML. W3C Recommendation.

W3C. Available at http://www.w3.org/TR/xhtml-modularization/.

Appelquist Daniel, Mehrvarz Timur & Quint Antoine. 2005 (April). Compound Document

by Reference Use Cases and Requirements Version 1.0. Working Draft.

W3C. Available at http://www.w3.org/TR/2005/WD-CDRReqs-20050404/.

Axelsson Jonny. 2004 (July). XHTML 2.0. Working Draft. W3C. Available at

http://www.w3.org/TR/xhtml2/.

Ayars Jeff. 2001 (August). Synchronized Multimedia Integration Language (SMIL

2.0). W3C Recommendation. W3C. Available at http://www.w3.org/TR/smil20/.

Capin Tolga. 2003 (January). Mobile SVG Profiles: SVG Tiny and SVG Basic. W3C

Recommendation. W3C. Available at http://www.w3.org/TR/SVGMobile/.

Dubinko Micah & Raman T.V. 2003 (October). XForms 1.0 Basic Profile. Candidate

Recommendation. W3C.

Forum Nokia. 2003 (April). WML to XHMTL Migration. Tech. rept. 2.1. Nokia.

Honkala Mikko, Cesar Pablo & Vuorimaa Petri. 2004 (December). A Device Independent

XML User Agent for Multimedia Terminals. Pages 116-123 of: IEEE Sixth

International Symposium on Multimedia Software Engineering.

Jalava Teppo, Honkala Mikko, Pohja Mikko & Vuorimaa Petri. 2005 (April).

Timesheets: XML Timing Language. Member Submission. W3C. Available at

http://www.w3.org/Submission/xml-timing/.

ten Kate Warner, Deunhover Patrick & Clout Ramon. 2000 (May). Timesheets - Integrating

Timing in XML. In: WWW9 Workshop: Multimedia on Web.

van Ossenbruggen Jacco, Hardman Lynda, Geurts Joost & Rutledge Lloyd. 2003

(May). Towards a Multimedia Formatting Vocabulary. Pages 384-393 of: Proceedings

of the twelfth international conference on World Wide Web.

WAP Forum. 2002 (January). Wireless Application Protocol WAP 2.0 Technical White

Paper. White Paper. WAP Forum.

8



Serkant KARACA
Senior Software Engineer
Ankara/Turkey
http://kurtlar-vadisi.blogspot.com
http://www.internetteyim.net