This report is also available as an Acrobat file.

Contents
Authoring and Design for the WWW

Appendix Two: Some basic facts

The Web is electronic, computer based

Computers are used by the publisher to

store documents
these documents are disk-files containing text, still images, animations, video, sound and interactive elements
each file normally contains information of a single type, rather than mixed media, but the browser software is capable of integrating these into the appearance of a single document
'serve' the documents to the users over networks
server software on the publisher's computer responds to requests sent by the user's computer, sending the files in return
run computer programs that assist the user
ancillary computer programs may be run on the server, for example to get information from a database on the user's behalf

Computers are used by the user to

request documents from servers
whenever the user operates a control in the browser software, or triggers an active element in the current document, this is transmitted back to the server as a command or request, leading to a response by the server
present documents to the user
documents are displayed on the user's computer, in a way appropriate to their file type, so text files appear in a scrolling window, sound files are passed to the sound-output hardware, etc.
separate components (text and graphics files retrieved from the server for example) can be composited together by the browser software into the appearance of a single document. The files which are brought together in this way need not be on the same server.

The Web is based on networks

The World Wide Web depends on networks connected to each other, using agreed forms of addressing and protocols for transferring files. It uses a variety of different forms of physical connection, which can in principle be ignored. When using the phone, we do not normally worry about how the physical connections are made, and the same ought to be true of the Web. In practice, publishers and users may need to deal with such issues as the bandwidth of the immediate network, and of the gateways which connect it to the rest of the world, and this may require them to influence the network and computing policies of their HEI.

The Web costs the user nothing

The Web is free, to the extent that there are currently no charges based on the duration or distance of Web transactions. Use of the Web costs nothing if the HEI provides

a computer
suitable software
connection to the internet
enough time to make use of it

The HEI pays fees to be connected, but does not (yet) charge the user personally. The academic connection carries with it certain restrictions, such as prohibiting commercial use (but this does not mean that academic Web publishers cannot inform their users about courses and other services that they offer).

An increasing number of sites on the Web are charging a fee for access to their information. This trend is likely to continue and poses serious problems for future Education. For the entrepreneurial individual or institution however, it represents an opportunity to publish in a new form which may eventually yield a new income.

Costs of the Web to the publisher

Publishers of Web documents may find that their HEI provides all the facilities described above plus the use of a Web Server on which to mount their files. As the Web becomes more a part of everyday life, losing its novelty value, the willingness to provide material support and services for free is likely to decline. It will become increasingly important to budget at the outset for server space and for the development and maintenance of pages (see Creating a maintainable site).

Requirements for users

The minimum requirements for the Web user are a computer connected to a network, a web browser package such as Netscape, and an IP address, normally provided by the HEI's computing services. The user may also have ancillary software in the form of plug-ins and helper applications for reading data other than text and simple graphics. The user's computer can be a PC, a Macintosh, a Unix machine, anything for which network connection is possible and browser software is available.

Requirements for creating documents

The content creator needs a computer on which to originate and edit pages, and the individual components of those pages. Basic textual pages can be made with the simplest text editor. More specialised tools allow the HTML code tags to be visually differentiated from the text content, to make editing easier. Greater sophistication is provided by special-purpose tools or by page-layout tools which have adaptations for making Web material.

An increasing number of packages whose main purpose has been the production of paper documents are now claiming to provide facilities for the production of HTML. Examples include Microsoft Word, ClarisWorks, Quark Xpress and PageMaker (the Adobe packages Illustrator and PageMaker also offer direct output to PDF files). Claims of HTML-compatible output are easy to make and need to be treated with a certain amount of suspicion. While it is easy to see how an ordinary word-processor or page-layout package can be adapted to output the relevant codes to make a headline bold (for example), it is much less easy to imagine that they provide access to more advanced (and useful) facilities of HTML such as identifying various kinds of lists. Also unclear is how the production of hypertext links might be handled, or of clickable maps. In some cases, the aspects of HTML authoring which are difficult and tedious to do by hand' such as tables are the very aspects which are not automated by these extended packages.

An advantage of extensions to existing software packages is that the learning process will be a short one if the author/designer is already familiar with it in the context of paper publishing. Also, where both paper and Web output are required, there may be economies (but with the usual caveat that writing and designing for paper and screen are not the same). An alternative approach is taken by packages like Adobe PageMill, specifically for the graphical layout of Web documents. It is likely that this sort of package will grow in importance and became more capable than in its early versions.

For additional media, additional tools are required. These include creation and editing tools for images, sounds, movies, 3D models and fully interactive components.

A study of authoring packages has been undertaken for the AGOCG SIMA initiative, see Web Resources.

Requirements for serving documents

The computer which acts as a server must be fast and reliable, and capable of efficient networking. The computing power required is (crudely) the number of simultaneous users multiplied by the bandwidth required for an adequate service for each of them, so a very popular site will make greater demands on the server than one which supports only a few users at a time. It is necessary to budget for the worst case of multiple simultaneous connections, since if users find a site inaccessible at peak times, then almost by definition, a maximum number of users have been disappointed at the time when they most wanted access.

A web server software package alone will be sufficient for basic Web publishing. Greater sophistication is provided by CGIs which offer additional functions for authors and users, for example clickable maps pictures on which the user can click at a particular location to access various sources. CGIs can be found on a variety of shareware sites. To enable communication between a Web page and a database package, a CGI plus a suitable database package will be required. Staff able and willing to create or modify CGIs may be needed.

Additional data provision such as streamed audio sound which can be delivered continuously over the Web rather than needing to be first downloaded to the user's computer may require additional server software. It is important to estimate the extra pressures on the network and the server associated which would arise from providing a stream of audio data to several simultaneous users, compared with simpler pages where for much of the time users are viewing the material passively at their machine and little network traffic is involved.

A full checklist of requirements for Web publishing is provided under Policy.

Getting support

Any computer which can write text can be used to create simple Web pages, and this can be done independently by an individual, but in order to publish those pages the individual needs goodwill from the HEI's computing services who will provide both time and commitment.

Connection on-campus

The commonest form of connection is one provided within the user's institution, typically by the computer centre. It is one which the academic need know little about, since all the technical set-up will probably be done by computing staff. They should be happy to provide the relevant browser software (which is usually free to HEIs) and the physical access to the network. However, staff on campuses which are not networked may need to lobby for their needs to be recognised. There are different policies in institutions about who should pay for the various components of connection to the Web. It is important to talk to the computer centre about what is needed and why. It may be necessary to prepare a good case beforehand, using information from a colleague or someone in another HEI who already has access.

Connection from home

Users can connect to the Internet from home via their HEI, rather than paying a commercial internet service provider which would typically mean paying both a subscription and a connect charge. While some might see the idea of being connected to work when at home as threatening, this low-cost Internet access can also be seen as a perk of the job.

Connection to the HEI is achieved using two modems, one the user's responsibility and connected to their phone line at home, the other in the HEI. Setting up a modem should be easy, but often advice is needed to explain unexpected problems. Once set up, stored configuration settings enable the modem to know how to connect in future. It is then just a matter of running an appropriate program to remake the connection whenever it is needed. The HEI's modem will not be dedicated to a single user but will take calls from a number of off-site users. Normally the user will need passwords, imposed in order to prevent unauthorised access. It is effectively impossible (for all but the dedicated hacker) to get connected without the cooperation of the HEI's computing services. Some HEIs are far ahead in providing this kind of access for both staff and students, while others have hardly begun to recognise the possibilities.

The dial-up user is connected via public telephone lines to the HEI's network. Speeds will be lower than when connected directly to the network internally.

All the time the user is connected to the HEI, even if not actually doing anything, charges are adding to the user's personal telephone bill. However, the phone charges are for connection to the HEI perhaps only a local call away and are unaffected by the distance away of the Web sites which the user visits'. It is as cheap to look at a document in India as one in the UK. Home dial-up users cannot make phone calls at the same time as using their modem (unless they lease a second line) nor can they be phoned by others while connected.

The faster the connection to the HEI the user wants, the more expensive the user's modem. However, there is no point in exceeding the speed of the HEI's modem at the other end. Some HEIs provide modems for staff, but most do not. It is worth checking with computing services staff to find out if they recommend particular modems before committing to a personal outlay.

How users see hypertext in the Web

In a Web browser such as Netscape, the first thing the user sees on opening a file (probably one on a distant server) is a scrolling window of text and perhaps graphics. Web users can view documents by any of the following means:

open a document explicitly, if they know its 'address'
search a variety of indexes for it, by title, keyword and so forth
open it from the document they are currently viewing by interacting with a hypertext anchor' an item in the current document such as a word or picture. The link must have been created by the author of the document.

A Web document may (unusually) contain no items which link to other documents, in which case users, if they wish to open another document, must open it explicitly. However, the likelihood is that the document contains links to other document parts. These parts may be elsewhere in the same document, or they may be in other files, perhaps local, perhaps on completely separate servers at huge geographical distances.

If the current file does contain links to other document parts, then this is indicated using colour, underlining, or other style attributes, which are often configurable by the user (and are stored in a Preferences file). In Netscape, as in most systems, the activation of a link is achieved simply by clicking on the hot text'. If, having traversed a link, the user wishes to return to the original document, the standard facilities of the browser provide a go back' button for this purpose.

In addition to the formatting which indicates links to other documents, the HTML coding of documents used on the Web also provides facilities for differentiating between body-text and various levels of headings, and for formatting structured elements such as a variety of lists.

How the Web finds resources

Browser software finds material using URLs Uniform Resource Locators. In order for a document stored somewhere on the Web to appear on the user's screen, the browser must have access to this address information.

A URL contains more than just the address of a distant server. The first part of the URL tells the browser software how to exchange information with it. In the case of straightforward links to HTML documents, it will be http (for HyperText Transfer Protocol). So common is this that many Web browsers allow it to be omitted when typing in a URL manually.

Connecting

When the Browser program is run, it will initially attempt to open whatever URL is set up in its configuration or preferences file. This can be altered for future occasions by the user. A dial-up user using a modem may need to run a configuration program first to actually open the connection to the HEI's modem. From then on, opening further URLs is done manually or by interacting with the document on display. While at one time only items of text or whole pictures could act as triggers to open other URLs in this way, an increasing range of interactions can now produce this result, including clicking on parts of a picture or interacting with a fully multimedia component.
Open Location dialogue box

The user enters a URL by hand, in this case in a dialog box provided by the Open Location option in Netscape

The user moves the pointer over a hypertext anchor which, when clicked, will open the corresponding URL. In this example from Netscape, the URL associated with the hypertext link is displayed in the status bar at the bottom of the window

In addition, browsers may provide:

Back reload the last URL visited
History List reload a URL visited in the current session
Bookmarks reload favourite URLs, whose addresses are stored with the user preferences
Reload reload the current URL, for example in case of problems with its current display

Contents

Graphics Multimedia Virtual Environments Visualisation Contents