Website Design

For Information Technology Professionals

Tom Worthington FACS

Visiting Fellow, Department of Computer Science, the Australian National University, and Director, Tomw Communications Pty Ltd, Canberra

For Internet, Intranet, and Document Systems (COMP3400/COMP6340), May 2002

Note: 2005 Version now available


  1. Introduction
  2. WebML: modeling language for designing Web sites
  3. Guidelines for Site Design
  4. SOCOG Olympic Web Site and the Disabled
  5. The Web on Small Screen: I-mode phones and Interactive TV


The World Wide Web (Web) is a computer application which has become very popular. It provides a network of information resources, a uniform naming scheme for locating the resources, protocols for accessing them and hypertext navigation among the resources. Information technology professionals have decades of experience with computers and telecommunications. As web sites become larger and more complex IT professionals can contribute their experience into how to build such systems. IT professionals also need to learn about how graphic designers, librarians, journalists and marketers build web sites, to be able to design software for them.

The Discipline of Web Site Design

The information architect of a large, complex web site should be two things: someone who can think as an outsider and be sensitive to the needs of the site's users, and at the same time is enough of an insider to understand the site's sponsoring organization, its mission, goals, content, audiences, and inner workings. In terms of disciplinary background, the information architect should combine the generalist's ability to understand the perspectives of other disciplines with specialized skills in visualizing, organizing, and labeling information.Rosenfeld & Morville (1998)

Rosenfeld & Morville (1998) argue that information architecture is a new field, not clearly part of an existing discipline. However, people have been designing information systems for electronic digital computers for about 50 years. The Australian Computer Society defines Information Technology as:

"the development and application of computers and communications-based technologies for processing, presenting and managing data and information" (ACS 1997).

After being involved in pioneering many early web sites, IT professionals have tended to be relegated to a technical support role in web site development. This can cause problems as the scale of web sites and their increasing functionality exceed the design skills of graphic designers, librarians, journalists and marketers. To be able to design web software and assist in the design of large scale web projects, IT professionals need to adapt their skills and learn some new web design skills.

What is the World Wide Web?

There is no formal definition of what the World Wide Web (Web) is. However, the World Wide Web Consortium (W3C), which develops web specifications, describes it as:

... a network of information resources. The Web relies on three mechanisms to make these resources readily available to the widest possible audience:

  1. A uniform naming scheme for locating resources on the Web (e.g., URIs).
  2. Protocols, for access to named resources over the Web (e.g., HTTP).
  3. Hypertext, for easy navigation among resources (e.g., HTML).

W3C (1999)

The web can be summarized as:

Web Mechanisms
Locating uniform naming scheme URI
Accessing Protocols HTTP
Navigating Hypertext HTML

The web was built on the foundation of the pre-existing Internet, which was built on pre-existing information technologies. Many of the challenges which look new to web site developers were faced by IT developers decades before the Internet or the web existed. IT professionals can use the experience learn from the past to add new features, such as support for mobile devices and interactive television (iTV), to the web.

The ACS Core Body of Knowledge

The Australian Computer Society divides the aspects of IT into:

Areas of Knowledge

  1. Computer Organisation and Architecture
  2. Conceptual Modelling
  3. Database Management
  4. Data Communications and Networks
  5. Data Structures and Algorithms
  6. Discrete Mathematics
  7. Ethics/Social Implications/Professional Practice
  8. Interpersonal Communications
  9. Program Design and Implementation
  10. Project Management and Quality assurance
  11. Information Security
  12. Software Engineering and Methodologies
  13. Systems Analysis and Design
  14. Systems Software

From Core Body of Knowledge for Information Technology Professionals (ACS 1997)

The major challenge in IT is to discover what the client wants, what might achieve that need, to reliably build a system to meet the need and then ensure the system operates into the future. As web sites become larger and more complex the Body of Knowledge built up for IT can be applied to the web. Some examples:

Web Design

Designing Navigation Systems: Examples

Browser Navigation Features

Commonly used conventions for web navigation have been established by popular browser programs. Some applications, such as micro-browsers embedded in appliances, may justify not using these navigation conventions. Also it is possible to switch off or override some browser controls for some applications, such as public web kiosks. However, for most web site design it should be assumed that the conventions will be worked with.This also assists where additional hardware or software is used to provide access for the disabled.

Building Context

Readers can get lost as to which web site they are in and where they are. The navigation system can provide clues as to where the user is and was. Examples are a navigation bar across all web pages or a fixed menu frame in a multi-frame site. Examples are the fixed top and bottom menus on the the Commonwealth Government Entry Point. and the fixed left frame on the Department of Defence home page.

Remote Navigation Elements

With most web sites based on a paper book model, it is natural to include a title page (or "splash page"), table of contents, a site map (detailed table of contents) and an index. In some designs the site map doubles as a more accessible table of contents. Few sites have an index, but instead have a search facility. A search facility provides less context for the reader than an index, as the search finds what is there, not what else may be near it.

EdNA (The Education Network Australia) has a "splash page" which appears for a few seconds before the contents page loads. The site map uses a simple design with headings from the contents page and then one level of details. A confusingly powerful search facility is provided. An index called a "browse map" is provided.

Conceptual Design

High-Level Architecture Blueprints: No widely accepted modelling or diagramming techniques exist for web site design. IT professionals may be able to adapt Conceptual Modelling methods, such as those used for relational data and Software Engineering Methodologies used for requirements analysis functional and technical specifications process, data and object orientation models.

An example of an attempt to develop web site modelling is WebML (Ceri et al 2000), of which the entity-relationship modeling section will be discussed in detail later.

Structure Charts and Site Maps

Lengel (2002) describes the use of Structure Flow Charts in chapter one as an aid to designing a web site. In chapter two they describe the use of site maps to describe the web site for the user and comment "the flow chart you developed for your site in chapter one is a site map". In practice site maps rarely use diagramming and usually use indented text. Also the use of the term "flow chart" does not match that of computer science, where a flow chart show the flow of control in a computer program, not the structure of the data or of the software modules.

Site Map as a Diagram:

Site map for Australian Heritage Commission web site

From: Site map for Australian Heritage Commission web site, Australian Heritage Commission, 1999

Site map as text:

From: EdNA Online - Sitemap, Education Network, 2001

Worthington Methodology for Quick Web Site Design

The design is based of a paper book provides a familiar hierarchical structure for a web site:
Book Web Site
Cover Splash Page
Front Matter, Foreword, Preface Home Page, About Page, FAQs, Contacts
Table of Contents Main Menu
Chapters Pages
Back Matter, Bibliography, Index Other Sites, Index Search

To quickly turn a conventional book or report into a web site:

  1. Obtain an electronic copy of the approved, typeset document.
  2. Save the document in web format.
  3. Replace presentational markup with structural markup and style properties. For example replace explicit font sizes and colors with standard heading level codes (such as <H3>) and styles. The Tidy program assists with this.
  4. Resize or delete excessively large graphics.
  5. Separate content into web pages: cover page to "splash page", table of contents to home page, chapters to web pages.
  6. Add extra navigation, such as a replacement for the paper index, some hypertext and sub-menus.
  7. Coalesce some pages and edit some content. For example condense part of the introduction onto the home page.
  8. Check the document and after approval publish it online.

This process can reduce the production of a web site from months or weeks to a few hours. If there is time later, graphics and additional styling can be added to enhance the look of the site.

Examples of quick sites:

Public web sites need publicity to attract readers. The content of the site can be summarised and converted to announcements and presentations.

WebML: modeling language for designing Web sites

Introduction to WebML

Web Modeling Language (WebML) is a notation for specifying complex Web sites at the conceptual level. WebML enables the high-level description of a Web site along multiple dimensions:

From: "Modeling the CDNOW Web site" by Aldo Bongio ( ZIPped PPS, 323 Kb; PowerPoint Viewer required) (Ceri et al 2000).

The remainder of this section will discuss the data modelling aspects of WebML using excerpts from a the pre-publication draft of the book "Designing Data-Intensive Web Applications" by Ceri, Fraternali, Bongio, Brambilla, Comai, Matera, February, 08 2002 (used by permission of the authors).

Data Content

Focusing on the first dimension (data content):

Data modeling is one of the most traditional and consolidated disciplines of the Information Technology, one for which well established modeling languages and guidelines exist. For this reason, WebML does not propose yet another data modeling language, but exploits the most successful and popular notation, namely the Entity-Relationship (E-R) model.

The essential elements of the Entity Relationship model are entities, defined as containers of data, and relationships, representing semantic associations between entities. Entities are described by means of attributes and can be organized in generalization hierarchies. The WebML data model slightly simplifies the classical Entity Relationship data model, by leaving out a few primitives, like generalized N-ary relationships, which are relationships involving more than two entities, and relationships with attributes. These omissions do not diminish the expressive power of the WebML data modeling language, because the missing primitives can be simulated by the use of entities and binary relationship, which are relationships involving only two entities. A positive consequence of this simplification is that the WebML data model is simpler to learn and more compatible with the UML notation for class diagrams, which allows only binary relationships.


... An entity represents a set of objects of the real world, characterized by common properties and having an autonomous existence. Examples of entities are person, car, artist, and album. An entity is a statically defined concept, a descriptor of the set of common properties. Associated to an entity there are the actual objects, which exhibit the properties specified by the entity. These objects constitute the population of the entity, and they are the real content of the application at a certain point in time: a certain set of persons, of cars, and so on. The objects of an entity all share the common properties (also called attributes) dictated by the entity they belong to: persons and artists have a name, cars have a model and a make, albums have a title, a year of publication, and so on. The notion of entity corresponds to the concept of object class of UML, but for the fact that UML classes permit the designer to specify not only the attributes of objects, but also the functions applicable to them, in the spirit of object-oriented modeling.


Two Entities


Attributes are used in the WebML model to represent the properties of real world objects. Examples of attributes are the name , address, photo, and the personal home page of a person. The WebML graphical representation of attributes is depicted ... attributes are listed inside the entity box, below the entity name. In the example, attributes Title, Year and Cover are defined for entity Album, and attributes FirstName, LastName, Biography and Photo characterize entity Artist. WebML attributes are restricted to take at most one value, and it possible that an object has no value for some of its attributes. As the notation ... suggests, WebML attributes correspond to class attributes in UML.


Graphic notation for entities and attributes


... one or more attributes of each entity may be marked as a primary key, to express that they uniquely identify the entity objects. Key attributes must satisfy a few restrictions, not requested for regular attributes. Their value must be defined for every object of the entity, and there should not exist two objects with the same value of the key attributes. The specification of primary keys is optional, but their specification is recommended because they are useful also for the hypertext model, where key attributes can be used to represent the entity instances in certain navigational structures, like indexes, offered to users to let them select objects out of a collection. ... Attribute Title is chosen as the primary key of Album , while the pair of attributes <FirstName, LastName> is the primary key of entity Artist. Graphically, key attributes are distinguished by a small key con placed at the right of the attribute’s name.

primary keys

Graphic notation for primary keys

Data Types

WebML attributes are typed, which means that they assume values from well-defined domains. WebML supports standard built-in data types, as well as user-defined data types. The former are the primitive types typically provided by the most common programming languages, the latter are defined by the user. Built-in data types include...

Data type Brief description
String A “short“ sequences of characters
Text A “long“ sequences of characters
Integer An integer numerical type
Float A floating point numerical type
Boolean A true or false value
BLOB A Binary Long OBject, for example an image or a video, which must be handled in a special way because of its size. Blob types can be further refined by expressing their MIME type (for example image/gif)
URL A Uniform Resource Locator of a Web page

User-defined data-types are specified as enumerated domains, which are finite, ordered sets of values. For example, the user may specify a new type named SupportType with the three different values, e.g., “CD”, “Tape” or “Vinyl”...

IS-A Hierarchies

The WebML data model permits the designer to organize entities into a hierarchy, where they share some common features. This IS-A hierarchy (also called generalization hierarchy or specialization hierarchy) has one super-entity and one or more sub-entities. The sub-entity inherits all attributes defined in the super-entity and may add locally-defined attributes and relationships. For example JazzArtist and PopArtist are sub-entities of Artist, and JazzArtist has an extra attribute Instrument, specifying the instrument played by a jazz artist.

primary keys

Graphic notation for IS-A hierarchies


Relationships represent semantic connections between entities, like the association between an artist and his/her album... The WebML notation supports binary relationships without attributes, which are relationships connecting two entities. In this respect, WebML relationships are simpler than in the Entity Relationship model, where relationships may connect more than two entities and have attributes, and also simpler than class associations of UML, which are binary but may be extended with attributes.

Relationship roles can be annotated with minimal and maximal cardinality constraints, denoting respectively the minimum and maximum number of objects of the destination entity to which any object of the source entity can be related.

Based on their maximum cardinality constraints, relationships are called as “one-to-one”, if both relationships roles have maximum cardinality 1, one-to-many”, if one relationship role has maximum cardinality 1 and the other role has maximum cardinality N, or “many-to- many”, if both relationships roles have maximum cardinality N. ... WebML graphic notation for a relationship called Publication, defined between entity Album and entity Artist. An album is associated to exactly one artist (cardinality 1:1), and each artist may be associated to several albums (cardinality 0:N); thus, the role from album to artist is mandatory, while the role from artist to album is optional. The relationship is “one-to-many”, from one artist to his multiple albums.

primary keys

Graphic notation for relationships

Complex Example

... a simple data schema describing albums and artists information. Artists publish albums composed of tracks, and have reviews of their work. Artists have a name, a biography, a picture, date and place of birth, and, possibly, date and place of death. Artists are partitioned into jazz artists and pop artists, the former being characterized by an instrument. Albums have a title, year of publication and a cover picture, and each track of the album is characterized by a number, a title and its duration. Each album is available on different supports, like CD, tape and so on, and each support has a list price, a discount and an actual price. For each review associated to an artist, the author of the review is known and the text containing the review is available. Cardinalities between entities are also defined: a review is associated to a single artist, an artist may have several reviews and may have composed several albums; an album is of a unique artist, may have different supports, and contains many tracks; finally, a track belongs only to an album.

The Album, Artist and Review entities include the definition of primary keys, either consisting of a single attribute or of multiple attributes. Conversely, Support and Track are examples of entities for which no meaningful primary keys can be defined, because for both entities no “reasonable” combination of attributes is sufficient for identifying uniquely the entity’s objects. For example, attribute Title may appear a good candidate key for entity Track, but is discarded because there may be two tracks in different albums with the same title.

Data schema

Data schema in WebML for the Cdnow Web Site

This technique might be applied to analysis of a complex site, such as

Guidelines for Site Design

Web site designers must balance the needs of the intended users of the web site, the demands of the client and the capabilities of the network and computer equipment. Unlike computer applications for use in an organisation, the web designer has little control over the computer network and equipment used to display their work. Ideally the web site should be first designed to function on a small screen with one column of monospaced, monochrome font text. The design can then be enhanced with fonts, color, images and multiple columns, if required. The overall structure of the site should be designed using a site map after use of E-R modelling or other analysis techniques. The ANU's Timetable system could benefit from having its structure made viable to the user.

Aspect Ratio and Display Size

The web designer should try to cater for the greatest range of screen sizes and shapes possible (techniques for this are discussed below). Lengel (2002) in chapter three, suggest that most modern monitors have a display size of 800 pixels wide by 600 pixels high. However, many older computers and web appliances have a screen 640 x 480 pixels. Hand held devices and internet TVs have smaller screen(discussed in detail below). Most screen have an aspect ratio of 4:3, but wide screen TVs are 16:9 and some PDAs have square 1:1 screens. An application designed specifically for a 800 x 600 screen may not be usable on a 320 x 240 internet TV. A particular problem is a fixed layout which does not allow the text to wrap to fit on a narrower screen, as some devices do not have horizontal scrolling to allow all text to be seen.


Web pages are traditionally layed out in a similar way to printed text, in a regular pattern, with more general information at the top left (for English publications). In general web pages should be laid out following this pattern, but having too rigid a design can result in a dull design and the reader being unable to distinguish one page from another in the web site. The layout should, as far as is possible, allow for screen of different resolutions. Cascading Style Sheets (CSS) include commands for positioning elements on a web page. However, these parts of CSS are not well supported by some browsers and it may be necessary to use <TABLE> commands to position text.


Text should be sized relative to the default size set by the browser and styled using Cascading Style Sheets (CSS) based on the structural elements of the document. The ANU has a standard style sheet "":

H1, H2 { font-family: Arial, Helvetica, sans-serif; font-style: normal; font-size: medium; color: #000000 }

H3 { font-family: Arial, Helvetica, sans-serif; font-style: normal; font-size: medium; color: #000000 }

H4 { font-family: Arial, Helvetica, sans-serif; font-style: normal; color: #000000 }

H5 { font-family: Arial, Helvetica, sans-serif; font-style: normal; color: #000000 }

H6 { font-family: Arial, Helvetica, sans-serif; font-style: normal; color: #000000 }

TBODY, BODY, TABLE, TD, BLOCKQUOTE, FORM, LI, OL, UL, DD { font-family: Arial, Helvetica, sans-serif; font-style: normal; font-size: small; color: #000000 }

P, p { font-family: Arial, Helvetica, sans-serif; font-style: normal; font-size: small; color: #000000 }

A { font-family: Arial, Helvetica, sans-serif; font-style: normal; font-size: small; }

A:hover { color:red }

This is referenced in the Discover The Australian National page HEAD:





<LINK href="/templates/anustyle.css" rel="stylesheet" type="text/css">

Later in the HEAD an additional style of tinylink is defined for use anywhere on the web page:

<STYLE> .tinylink { font-family: Arial, Helvetica, sans-serif; font-size: 8pt} </STYLE>


In the body of the document some font commands are used:

<BODY ...

<FONT SIZE="-2" face="arial, verdana, helvetica, sans-serif"> <center> Please direct ...

As well as the additional style:

<A HREF="/comments.php" class="tinylink">...

The effect of the styles cascade, that is the local style overrides the definition in the header and that overrides the external definition. In this case the font size will be changed, while the font color will be unchanged.

Fixing text at, 12 points, for example as suggested by Lengel, will cause difficulties if the user's screen is significantly smaller or larger than that the page was designed on. Web browsers have a default font size which should have been set to be readable on the particular screen. Most text on the web page should be set at the default size and other text, such as headings, should be set relative to default size. Serif fonts (such as Times Roman) are generally considered easier to read for main text, but sans serif (such as Helvetica) are better for low resolution screens and headings. As with text size specific fonts should be avoided and style commands used to suggest fonts for specific elements of the text. As an example a heading should be marked as a heading <h1> and then styles added for text size and font.

In this way if the browser does not have the specific font (or the user can't see it), another can be substituted in a systematic way. Utiltiy programs, such as Tidy, can be used to extract collect text defnitions from a web page into a CSS style sheet. An example of a web page using individual defninitions is Links on the World Wide Web (Spool 1999). An example of a web page using an internal CSS style sheet is The Federal Government 30% Rebate (HIC 2001). An example of a complex style sheet is that for the W3C's documents: This includes comments indicating some style commands cause difficulties with some web browsers.


GIF, JPEG and PNG provide lossless formats for perfectly rendering images, such as logos and diagrams. JPEG provides a lossy format, which will render a not completely accurate version of the image, which is useful for photographs. Images should be sufficiently large to display on the largest anticipated screen, but not so large that they waste bandwidth and storage. Each image which conveys information should be accompanied by a text caption which describes the image for readers who can't see the images. Decorative images do not require captions (and should have a null ALT tag). Including the size of the image in the IMG statement will aid the web browser to lay out a web page quickly. Similar considerations apply to animations, sound and video.


The web site should have a consistent color scheme, using a small number of colors. This can be best achieved by using styles to apply colors to all elements. Color should not be relied on to convey information as some people can't see some colors, or may have a monochrome screen.

Spool (1999) suggests that longer text links work better than cryptic ones: "When people are searching for information, they pick links based on their expectation of the page the link will take them to.". Image links when combined with text may be better:

Users sometimes made their decision and clicked on the text (which wasn’t a link) before the GIF links even appeared, so clearly the additional information helped them decide...

A weak positive correlation exists between image links (such as graphics that look like buttons) and user success. However, we saw one minor problem with image links: because they don’t change color after the user clicks them, there is no visual cue that they have already been traversed. Users did use visual cues to initially identify the image links by moving the mouse across the page and watching for the cursor to change to Netscape Navigator’s pointing finger.

Web Site and the Disabled: The SOCOG Olympic Web Case

In August 2000 the Sydney Organizing Committee for the Olympic Games was found to have engaged in unlawful conduct by providing a web site which was to a significant extent inaccessible to the blind. How was an analysis of the content of such a large web site carried out? What features were looked for? How was the impact of changes to the site estimated?

Excerpts from: Olympic Failure: A Case for Making the Web Accessible

Web accessibility guidelines have been developed to assist designers to make web sites which are available to the greatest range of users of the Internet. The most respected guidelines are those from the World Wide Web Consortium (W3C 1999):

For those unfamiliar with accessibility issues pertaining to Web page design, consider that many users may be operating in contexts very different from your own:

Content developers must consider these different situations during page design. While there are several situations to consider, each accessible design choice generally benefits several disability groups at once and the Web community as a whole. For example, by using style sheets to control font styles and eliminating the FONT element, HTML authors will have more control over their pages, make those pages more accessible to people with low vision, and by sharing the style sheets, will often shorten page download times for all users. (W3C 1999)

On 7 June 1999 a complaint was made to HREOC, by a blind person that he was unlawfully discriminated against by SOCOG in three respects: the failure to provide braille copies of the information required to place orders for Olympic Games tickets; the failure to provide braille copies of the Olympic Games souvenir programme; and the failure to provide a web site which was accessible to the complainant. The Ticket Book component of the complaint was dealt with on 30 September 1999 and Souvenir programme, 27 March 2000.

The complainant requested that the HREOC make the following orders or declarations:

  1. That SOCOG include ALT text on all images and image map links on the website;

  2. That SOCOG ensure access from the Schedule page to the Index of Sports; and

  3. That SOCOG ensure access to the Results Tables on the web site during the Olympic Games. (HREOC 2000)

A decision on the web accessibility was delivered 24 August 2000 (HREOC 2000), with SOCOG found to have engaged in unlawful conduct by providing a web site which was to a significant extent inaccessible to the blind. The web site ordered to be made accessible by the start of the Sydney Olympics. On 6 November 2000 (after the Sydney Olympics) the web site was found to only be partly complaint and $20,000 damages were awarded (HREOC 2000c).

On 27 April 2000 the following pages were inspected (these pages may have changed since this inspection was carried out):

The web pages were displayed using Internet Explorer (Version 5.0 for Windows 95), with image display switched off (to simulate use by a blind person). The display of the web pages was inspected for usability and then an examination of the HTML source code made. It should be noted that a more detailed analysis would be required to more formally rate the site, but this was not required, nor was it feasible.

Problems found and compliance with Web Accessibility Guidelines were summarised as:

i. W3C Recommendation Web Content Accessibility Guidelines 1.0, 5-May-1999 at ( ).

From: Guideline 1. Provide equivalent alternatives to auditory and visual content. Checkpoints ( ):

“1.1 Provide a text equivalent for every non-text element (e.g., via “alt”, “longdesc”, or in element content). This includes: images, graphical representations of text (including symbols), image map regions, animations (e.g., animated GIFs), applets and programmatic objects, ascii art, frames, scripts, images used as list bullets, spacers, graphical buttons, sounds (played with or without user interaction), stand-alone audio files, audio tracks of video, and video. [Priority 1]”

Guideline 5. Create tables that transform gracefully. ( )

“5.1 For data tables, identify row and column headers. [Priority 1] For example, in HTML, use TD to identify data cells and TH to identify headers.

5.3 Do not use tables for layout unless the table makes sense when linearized. Otherwise, if the table does not make sense, provide an alternative equivalent (which may be a ). [Priority 2]“

From: Guideline 1. Provide equivalent alternatives to auditory and visual content. Checkpoints ( ):

Checkpoints:1.1 Provide a text equivalent for every non-text element ...

For image maps, either use the “alt” attribute with AREA, or use the MAP element with A elements (and other text) as content. “

ii. AusInfo Guidelines at Equal access and the web: some issues ( )

“Availability of information and services in electronic form via the web has the potential to provide equal access for people with a disability; and to provide access more broadly, more cheaply and more quickly than is otherwise possible using other formats...

Current text readers and Braille output devices are not able to deal with information or links presented only in graphics or photographic format. “

iii. IBM’s Web Accessibility Guidelines and Checklist at

Version 2.1 January 11, 2000 ( )


1: Images and animations. Use the alt=”text” attribute to provide text equivalents for visuals. Use alt=”” for visuals that do not convey important information or convey redundant information.

2: Image Maps. Use client-side image maps to provide accessible text for image map hot spots. If you use server-side maps, provide equivalent text links.


9. Tables. For tabular data, use the CAPTION element and/or the summary attribute. For tables with complex row and column headers, use the headers attribute on cells.”

A demonstration showing two options for implementing accessibility features for the menu was prepared later (Worthington 2000b):

Option 1: ALT Text on the existing image map

In this implementation, a ALT tag with the name of the sport is added to each rectangular region of the existing image map. As an example, alt="Archery" is added to the first area:

< AREA shape="rect" coords="1,1,80,12" href=
sports/AR/home.html" alt="Archery">

There is no change in the appearance of the menu for most users. However, the effect of including the ALT tags is that a browser can display the text labels in place of the graphics, if required.

Option 2: Simple text menu supplementing or replacing the existing image map

Simple hypertext links are used for the name of each sport:

programs/index.html?/eng/sports/AR/home.html"> Archery </A>

These instructions could be used in a simple list of sports, or in a table as show below. The list has been changed to run alphabetically across the columns, rather than down, to aid braille readers (which read one line of horizontal text at a time). The colour of the table background and text can be changed to match that of the current graphic. However, those colours may be difficult for some people to read:







Canoe/Kayak Slalom

Canoe/Kayak Sprint

Cycling: Mountain Bike

Cycling: Road

Cycling: Track





Gymnastics: Artistic

Gymnastics: Rhythmic

Gymnastics: Trampoline




Modern Pentathlon






Synchronized Swimming

Table Tennis




Beach Volleyball

Volleyball Indoor

Water Polo



However, some disabilities, such as Aphasia involve reading problems and graphics can improve access (CDARU 2001). A simple way to achieve this for the Olympic Games would be to use the standard sports pictograms, along with text in the sports menu. The pictograms were already used on the sports web pages linked from the menu. Thus:

Archery Archery

Accessibility of web sites for the disabled is a subset of the more general issue of usability of IT applications. Applying the discipline of disabled access can provide insights into better web and general computer application interfaces for non-disabled users. As an example accessibility features can be used to make hand held wireless web devices more practical, by improving usability of a small screen and lowering bandwidth requirements.

The Web on Small Screen: I-mode phones and Interactive TV

A well designed web site can be adapted to devices very different from a conventional personal computer. Two examples are web telephones and interactive television set-top-boxes


I-mode provides a web service on a mobile telephone screen, using a proprietary mark-up language based on HTML:

First introduced in Japan in February 1999 by NTT DoCoMo, i-mode is one of the world's most successful services offering wireless web browsing and e-mail from mobile phones. Whereas until recently, mobile phones were used mostly for making and receiving voice calls, i-mode phones allow users also to use their handsets to access various information services and communicate via email.

In Japan, i-mode is most popular among young users, 24 to 35 years of age. The heaviest users of i-mode are women in their late 20s. As of November 2000, i-mode had an estimated 14.9 million users.

When using i-mode services, you do not pay for the time you are connected to a website or service, but are charged only according to the volume of data transmitted. That means that you can stay connected to a single website for hours without paying anything, as long as no data is transmitted.

From I-mode FAQ, WestCyber Corporation, 2000 (no longer on-line)

I-mode's data format evolved from Compact HTML (cHTML). It is intended to be used on a portable device with small memory, low power CPU and small monochrome screen with one font. Unlike the Wireless Application Protocol (WAP) which uses its own Wireless Markup Language (WML), I-mode's data format is based on a subset of HTML, so that I-mode pages should be upward compatible with most web browsers:

3.2 Features of Compact HTML

The Compact HTML is a subset of HTML 2.0, HTML 3.2 and HTML 4.0. We describe the major features which are excluded from Compact HTML, as follows. We define that Compact HTML includes GIF image support...

I-mode web pages look like early, simple web page and like those designed for disabled access. It is not clear if I-mode web pages can be created which will be acceptable for use on normal web browsers, of if their simple design will be unacceptably dull.

An example of an I-mode web site is the Japan Times - Mobile Edition which assumes a small screen size:

gets breathing

chain operator
Co., which is
currently pro-
tected from
creditors un-
der the corpo-
rate rehabili-
tation law ...


The Japan Times
(c) All rights

The People's Daily - I-mode displays better on a conventional web browser by allowing text to wrap:

China Urges Advanced Countries to Help Developing

A Chinese central bank official on Sunday urged advanced countries to coordinate their macroeconomic policies and keep the major currencies' exchange rates stable to create an environment conductive to stability and growth for developing countries.
(04.30 14:0)
Top of Group Top Page

Interactive TV

Neon Technology NTV - 500 entry level set-top Advanced set-top-boxes (STBs) for digital television provide web-like features for the viewer. These allow the viewer to look at the electronic program guide and other information. While previously based on proprietary formats, more recent boxes use some form of HTML. An example of a proposed standard is Multimedia Home Platform (MHP), which attempts to adapt existing Internet and web standards for to digital Television (DTV). MHP uses DVB-HTML, which is an adaption of XHTML. However, MHP has not been widely deployed. Some Internet TV boxes, such as the Neon Technology NTV - 500, use the TV just as a display device and use a conventional dial-up telephone line for Internet access. Canberra's Transact broadband system currently uses Motorola Streammaster 5000 STBs. See: Web Page Display on Transact .


ATVEF is similar in concept to MHP, but less ambitious in scope and addresses US TV requirements more than European. The ATVEF manual is far more readable than MHP:

1. How to place TV in a web page (using <OBJECT> and <IMG> tags)

The OBJECT and IMG tags are used to place the TV picture in a web page, for example:

<object data="tv:" width="60%" height="60%">

<img src="tv:" width=320 height=240>

From: Enhanced Content Specification, ATVEF, 1998, Appendix B: Content Format Notes

Web-t and iTV characteristics

Web telephones and iTV share common characteristics:

Web content can be generated to provide adaption to specific devices. However, the target device may not know, or may change, or resources may not available to closely tailor for a large range of device. W3C's accessibility guidelines were specifically developed with the needs of small devices in mind, as well as disabled users. It is therefore possible to use the guidelines to design content to suit a generic hand held web device, iTV and larger screens.

Typical Small Device

While it is possible to create web sites for specific telephone and TV web devices, a reasonable target to aim for is a device with a one Quarter VGA (QVGA) screen (320 x 240). Web pages designed for this size can display on a conventional larger PC screen, as well as a PDA or a TV. There is no exact size in pixels for a TV set, as TV standards differ across the world and the edge of the screen may be masked, so not all the image is displayed. Small devices may have limited or no Java, they are more likely to have ECMAScript. While CSS may be supported, it may not produce a readable display on a small screen.

Diagram of Screen sizes

A QGVA screen can be simulated by folding a sheet of A4 paper in half four times, then unfolding it twice and cutting out the middle quarter. When placed over a typical PC SVGA screen the hole in the paper is about the size of a QVGA screen:

Diagram of Screen sizes

WebTV STB Emulator

An early internet TV device sold in the USA was WebTV. This was purchased by Microsoft and renamed Microsoft TV. While this service is not marketed in Australia there is a useful emulator "The WebTV Viewer" to simulate the TV browser on a personal computer:

It can tell you if your Web content is appropriate for the receiver. It has tools to show you how content is altered to look the best on a television screen. There is a special window that shows how tables and images are scaled... The Viewer does not have a TV tuner, so all of the broadcast pictures are simulated.

From: The WebTV Viewer, Microsoft Corporation, 2001

See: wTV: Adapting Weather Pages for Internet TV.


Further Information

A presentation version of this document is available: For an example of the multi-format technique proposed in this document, set your web browser to use the accompanying style sheet . This will omit sections of the document marked with the class definition "optional" and leave a large margin before titles marked "newslide". Set the browser is set to use a large font size and select the frames version of the document, for a slide-show type of presentation.

See also:

This document is Version 1.0 - 7 May 2002:

Modified from the 2001 notes.

Comments and corrections to:

Copyright © Tom Worthington 2002.