The World-Wide Web (also known as WWW or just the Web) is the provision of independent distributed servers which can work together and link together is a seamless way. The protocol each of these servers use is defined as the HyperText Transfer Protocol (HTTP) to transmit documents in the HyperText Mark-up Language (HTML). HTML allows documents to contain formatting information and links to other media such as graphics, audio and video. The free distribution of descriptions ,  of this mark-up language and information about how to construct servers and clients to display the information has led to a very large number of sites implementing and using this approach.
Public domain implementations of the HTTP servers are available from several sites and for several platforms so that servers can be run from UNIX systems, PCs and Apple Macintoshes. The client software to allow access to servers over the network and the display of the retrieved HTML documents is also available for these platforms and others. In addition for each of the platforms there is a growing choice of client program. This is encouraging for the long term future of the Web approach to providing information, as it seems likely that free clients and servers will continue to be available even as commercial systems start to appear.
In addition to supporting documents written in HTML for the WWW clients, the Web offers access to other information services. Allowing access by clients to other servers means that information already being made available on existing gopher and ftp sites can be immediately integrated. A further feature of the Web is that access is possible to any program on the server. This allows easy extension of the WWW system to specialised areas such as database access, on line information gathering and security checking.
3.1. Alternatives to the World Wide Web
The World-Wide Web has now established itself as one of the fastest growing technologies in the computer world. The Web and its protocols and display clients offer advantages over earlier systems such as gopher in the way in which they support structured documents and linking across different servers. However the existing system does have some weaknesses. In particular communication between the server and the client is not built into the design. Each connection with the server is a new connection and can not be simply matched up to an earlier connection. Software authored for local delivery can also take advantage of additional facilities and produce systems that are easier to use then those currently possible using the web. This may change with later versions of servers, clients and protocols. In addition some alternatives are becoming available, for example the Hyper-G system  which offers similar facilities but has additional features to allow greater integration between clients and servers and identification of users.
3.2. Using World-Wide Web for authoring or delivery
The facilities for authoring in the World-Wide Web have not been very sophisticated. HTML documents are basically plain text with mark-up indication (for example <h1>A Heading</h1> is an HTML sequence to define that text as being a level 1 heading). This means that any editor can be used to create the HTML documents. Tools are now appearing to help with this task, either sets of word processing macros (e.g. CU-HTML) or custom programs (e.g. HoTMetaL). There are now a variety of these to help in the authoring task, particularly in the creation of valid HTML documents.
The Web approach to authoring documents is not as sophisticated as that used in typical authoring tools (such as those discussed in ). This is partly because of the need for further development of the tools to assist the creation of the documents and also the need to introduce further features into the servers and client, for example tables, pop-up windows and in-line video. Some of these features are planned for future releases such as HTML2 and HTML3 . There is though a fundamental difference between the client-server approach of the Web and most other authoring approaches. On the Web it is under the control of the client how things are expected to appear on the screen. For example one client may choose to render a level 1 header in black bold Helvetica font at size 32point, while another may render it as blue italic Times-Roman size 12point. These local decisions mean that it is difficult to ensure that a given document appears as intended by the designer, the designer also can not predetermine the delays that will be imposed by using a remote server.
This variation in appearance and interaction response times is not a problem for many uses of the Web where the underlying information is what is important, but where the design has been made to ensure usability these changes can hinder the use of learning material. In practice many Web pages have been designed on assumptions based on the use of the most common client with its default settings. Some control can be regained by using active graphic maps  in place of text but the current implementation of these does not allow adequate feedback and response times become even more dependent on the network performance.
The majority of courseware has been produced for delivery to local machines using authoring systems targeted at those machines and for the reasons outlined above this is likely to continue to be the case. The role of a courseware server then can be to deliver this material in a simple and flexible manner. As the separate strand of courseware developed for the Web (and similar technologies) develops and becomes established this will be able to be served alongside the more conventional packages.