AGOCG logo
Graphics Multimedia VR Visualization Contents
Training Reports Workshops Briefings Index
This report is also available as an Acrobat file.
Back Contents
Authoring and Design for the WWW



When information is sent from one computer to another, the maximum speed at which it can travel is dictated by the capacity of the link as it were the diameter of the pipeline. Bandwidth is therefore measured as an amount of data per second. For a given route from one computer to another, the maximum speed of data transfer is dictated by the point of worst bandwidth in the chain. Transfer rates are seriously affected by network traffic, so vary depending on whether use is at peak times or quiet times.


The software which enables the user to find Web documents and use them is a Browser. Browsers are available for different kinds of computer. The simplest will work on a small inexpensive computer, though it must of course be connected to a network. As more complex multimedia documents are published on the Web, larger, faster computers are required to support the appropriate browsers.

Clickable maps

A picture can be a Web hypertext anchor as easily as a word. However, HTML does not inherently support pictures which trigger different links depending on where they are clicked. These are clickable maps, and currently require three components on the server: the image, a map file specifying which regions do what, and a CGI to handle the process.

Colour table

Anything displayed on a computer which supports only 256 colours (the most common figure) will have its colours assigned to the particular 256 colours available at the time, dictated either by the machine's operating system or the current application (such as a browser). This is the colour table, also known as the palette or CLUT (colour lookup table)

The colours in a graphic which was prepared using one colour table will look wrong (often completely wrong) if represented using another lookup table. This can happen if the colours in the originator's palette are not successfully recreated for the user by the browser package. It is vital to use a suitable palette when preparing images for the Web, and to test images on a variety of machines. See the Web pages listed under Technical Details in Web Resources, p110.


Compression is an essential tool in overcoming the problems of bandwidth. Files on the server are compressed, transmitted in compressed form, and decompressed at the destination. Time is used up in the decompression process, but this compares favourably with the delays caused by slow downloading. Different type of data are compressed in different ways. See also Graphics.


A picture file of 100K uses as much data as 17,000 words. While pictures may have advantages both in term of conveying information and enhancing the look of a Web page, they should be kept to file-sizes (after compression) below 30K.

For a picture which occupies all of a standard display, 640 pixels wide by 480 pixels deep, the number of bytes required before compression is 300K. This assumes that the picture uses one byte (8 bits) for each pixel. Such a bit-depth' will give an adequate representation of most images, and is standard for GIF files, commonly used on the Web. Using more bits per pixel would mean a greater amount of data to be transferred: for 16-bit, twice as much as for 8-bit; for 24-bit, three times as much. A disadvantage of using 8-bit colour is that it involves the use of colour tables.

For two useful sites on graphics compatibility and JPEG and GIF file-formats for the Web, see the Technical Detail section of the Web Resources.

Graphics paint and draw

There is an important distinction between graphical information stored as pixels, units of fixed size representing colour and tone, here called paint graphics, and that stored as the geometry and other attributes of drawn objects.

Graphics on the Web have been dominated by paint format images stored as GIF or JPEG files. However, as publishers and users become frustrated by their limitations, draw formats will become more common.

The distinction between paint and draw graphics has implications for the way in which graphics are originated, edited, stored and displayed. Draw files consume file-space on an object-by-object basis, so simple images are economical to store, complex ones more extravagant. They are often more economical than equivalent paint images. For the Web, the single most important disadvantage of paint images is that their resolution is set when they are created. Even if a browser or a plug-in provides a facility to zoom in on an image for a closer view (most do not) eventually the user will only see bigger and bigger pixels. For a draw image however, the geometry can be scaled in a much more useful way, for example allowing a user to study detail in a large diagram at one moment and get an overview of it the next. When a draw object is enlarged, it will be redrawn on the screen with acceptable resolution

Many kinds of images, particularly photographic, cannot only be represented in paint format, but for any sort of diagrammatic information draw formats are preferable. File-formats which can represent drawn imagery include Acrobat and Shockwave for Freehand.


Helper applications were the forerunners of plug-ins. While plug-ins deliver foreign' file types within the main browser, helper applications open a separate window for their own display. Helper applications can normally also be used independently of the browser, unlike plug-ins.


Text for the Web is marked up' using HTML, the Hypertext Markup Language, which is an informally agreed (but developing) standard. Originally authors and publishers were obliged to get to grips with the details of HTML, but increasingly they can rely on software tools which protect them from this level of detail.


Hypertext is text (and these days other media too) delivered in an interactive electronic environment. Users on viewing one part of a hypertext can interact with parts of it (for example by clicking the pointer over words, headings or pictures) to view another related document.

HTML was designed with hypertext in mind, and text is till the skeleton on which everything else hangs. Hypertext linking in HTML is more sophisticated for text than for other media types: using text, it is easy to make structures where individual parts of texts (words, phrases) lead to specified parts of other texts. However, when it some to other media, this is not so easy. It would be very difficult within the Web to make a structure for filmic sequences, where a particular frame in one movie linked to a frame in another. Similarly, while clickable image-maps allow one picture to have various discrete active areas, it would be difficult to jump to a particular part of another picture (or even of the same picture).


Outliner packages allow users to switch between top-level views of texts and more detailed views. Unlike some proprietary hypertext systems, HTML does not easily support this idea of folding and unfolding levels of detail.


A page' on the Web is one chunk of information having its own address' or URL (Uniform Resource Locator) so that the user's computer can find it. A page will be a single file on the server, but may have embedded in it other files such as graphics, sounds and animations.

There is no limit to the length of a page: the term does not denote a unit of fixed size as in the pages of a book.


Many software packages of all kinds have become modular to the extent that they can be enhanced with additional functions by mini-programs which are connected to them. A Web browsers which support plug-ins does not itself need to be able to handle all conceivable types of data, since a special requirement is catered for by the appropriate plug-in. Example uses include sound, increased interactivity and the display of video. The plug-in must be installed by the end-user. Some data types also require special software on the server where they are sited.

Plug-ins are not normally usable independently of the browser.


For documents to be available on the Web, they must be stored on a suitable Server. This is a network-connected computer of moderate to high specification, and runs a Server Application a software package which is capable of responding to requests for files which it then transmits to users.


A site on the Web uses a Server to offer a collection of related Web pages. Not all the pages which notionally comprise a single site need actually be on a single server, or even in one physical location.


All computer data, whether graphics, text, sound, video, is stored in similar digital files which are measured in bytes and their multiples. The size in bytes affects: Of these three, the most important for the Web by far is the transfer rate, because the speed with which data can be moved from the server to the user dramatically affects the usability. If users have to wait too long for information, or for a response to an interaction, they will usually give up. Significant improvements in transfer rate can be achieved using a variety of compression methods.


Tables are grid-like structures containing other elements. The grid itself need not be visible. Tables can be used to position elements more precisely on the Web page than by other means, but at the cost of decreasing the flexibility with which the information can fit into different displays.


Testing of a Web site includes both technical testing, to ensure that it works in all respects, and user-testing, to ensure that it achieves its intended purpose. See The Design Process (p72)


Text is very economically stored on computers, especially in unformatted form (with no information about position, style, typeface etc). The average length word (five letters plus its following space) occupies 6 bytes. So ten thousand words occupies 58K, and two PhD theses each 80,000 words long would easily fit on a single floppy disc (capacity 1.4Mb).

Text in computer graphics represents a special case of the difference between Paint and Draw (see also Graphics). Most importantly, draw text is live' text, which can be searched, sorted and generally manipulated as text. Paint text, on the other hand, while it is originated using text-editing tools, is not stored in the computer as text but as a pixel-based graphic: it is just another picture.

Live text can be

Since live text is rendered onto the screen each time the text is viewed, its appearance will suffer greatly if the font information (dictating the correct shape and spacing of the letter forms) is not available every time it appears. The commercial nature of fonts is one of the principal difficulties with which any system for the transfer of electronic texts must deal.

The advantages and disadvantages of paint text are essentially the opposite of those for live text.


Most operating systems for desktop computers now have optional extensions which convert text to the spoken word, with greater or lesser realism and sophistication. Text which is really graphics cannot be spoken in this way.


Uniform Resource Locators are internet addresses comprising three elements 1 the protocol, which indicates what kinds of messages can be exchanged 2 the domain name, which identifies the server using a unique address 3 the pathname or location on the server where the relevant file or other resource is located.

WWW, the Web

The Web is one application of the Internet which, by linking up networks, provides connection of computer to computer around the globe. Originally mainly textual, the Web is now a multimedia publishing mechanism which allows anyone with a Server to make documents available to other computer users. The kinds of publishing' of which the Web is capable are more varied and flexible than the term implies; also, there are many users of the Web which need not be World Wide'.
Back Contents

Graphics     Multimedia      Virtual Environments      Visualisation      Contents