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



How HTML works in the browser

Using HTML, the appearance of the document on the end-user's machine is dictated by a combination of the functional description decided by the document author, incorporated in the file (left in the diagram) and the formatting instructions for HTML incorporated in the browser software, perhaps modified by the preferences of the user (right). By contrast, other computer-generated documents such as those made with a word-processor are fixed with regard to their appearance when they are made.

The final formatted appearance of the document is a combination of the author's functional description of its elements together with the interpretation made by the browser package (as modified by the user's preferences)

In HTML the author specifies that an element is one of a set of pre-defined types, such as Sub-Head No3 or List Item, and a definition stored in the browser software specifies the visual interpretation of this type. It is thus a system for specifying the semantic structure of the document, albeit crudely, rather than its mere appearance. The author specifies the function of an element and the browser software applies an appropriate format.

The core of an HTML document is text, of two kinds: text which is to be shown to the end-user and that which instructs the browser software how to behave. In appearance and function it therefore resembles old- fashioned word-processing and print-specification files in which textual codes tags' enclosed in angle brackets are visibly embedded, and which are capable of turning on and off the various styles. The stream of text in the HTML file is split by the browser package into its two components, resulting in a display where only the content-text is visible, but its appearance and behaviour have been dictated by the now invisible tags.

Word-processing packages use codes (usually concealed from both author and reader) to give visual styles to elements of the document. For example a subhead might be in 14pt Baskerville Italic. HTML avoids this specific visual formatting.

When a browser package opens an HTML file (whether locally from a disc or over a network), it reads the content-text and tag-text as a continuous stream. The tags are indicated to the browser by angle-brackets (< and >).

The tag-markers < and > ensure that the tags are split off from the content text by the Browser. The tags are interpreted as instructions for formatting, interaction etc, while the content text is displayed to the user

For each element of the content, such as a headline, there will normally be a start-tag and an end-tag (eg. <H1>My headline</H1>) so the parsing process is quite simple. On finding a start tag, the browser sets the type size, weight, indentation etc. to that specified for this type of element. On finding an end-tag, it turns it off again (usually returning to the default specification for body-text, though it is unwise to assume this).

If the element is a link to another document, then the browser software can interpret mouse-clicks on that word or phrase as an instruction to display the specified document, to jump' to that document. If an element calls for the display of another media type, this is then handled in an appropriate way by the browser.


  1. A paragraph of plain text containing one hypertext link to another document is opened by the browser.
  2. The plain text (bracketed by <p> and < /p> tags) is formatted according to the defaults of the browser package:
    black type
    12 point
  3. The embedded hypertext link is displayed as follows:
    (but otherwise the same as the text around it)
  4. The user alters the preferences for the Browser display:
    the type is now, say, all 10 point Helvetica
    the hypertext link is shown in green

A significant outcome of this empowerment of the user in relation to the document's appearance is that for those users who are visually impaired or are struggling against adverse viewing conditions, it is a simple matter to set the type styles used by the browser to more acceptable sizes.

Automating design

HTML takes further the approach adopted by advanced word-processors and page-layout packages in allowing marked-up text to be automatically laid out. Once the semantic structure of the document has been decided and encapsulated in the necessary codes, no further intervention is needed (or indeed possible). A text document is marked-up, or the text is inserted into a prepared template and the design is taken care of by the standards of HTML, modified by the end-user's preferences.

This automation deprives graphic designers of many opportunities to exercise the subtleties to which they are accustomed, but it makes for an efficient production-line approach to the laying out of large quantities of text.

This approach has a lot to offer Education, where many documents share a common structure, for example course module descriptions. Fully marked-up text can be automatically generated from a database which stores every module description. For such documents which share a structure and to a large extent a style across a whole institution, there is very little to be said for everyone creating their own formatted text. In addition, where cross- reference to a common text is required, for example to the attendance regulations, the hypertext link which invokes this single document can be inserted automatically in every module document. If it is altered, it appears in its new form in every single location where it is cited.

Back Next Contents

Graphics     Multimedia      Virtual Environments      Visualisation      Contents