[FIGURE -- what a geek]
HTML is the publishing language for the World Wide Web. Behind the scenes, every document you see on the Web - whether covering the national news, giving information about university courses, selling books or describing a local cat show - is written using HTML. But why is HTML necessary? Why is a browser needed to display HTML documents, and what is the role of the server? These questions, as well as simple descriptions of Java, scripting, multimedia and other Web-associated technologies, form the subject of this chapter.
From your home in the United Kingdom, in the small English town of Chipping Sodbury, you decide to find out what entertainment there is in Bath, the famous city only 15 miles away. A company publishes What's On in Bath online, and you have read that this is available on http://www.bath.info.uk/ as a series of Web `pages' on the Internet. You decide to take a look. You switch on your computer and start it up in the usual way. Just as you might click on an icon to start up your desktop publishing package, so you click on the icon to load your World Wide Web browser. The browser is a piece of software that allows you to display certain kinds of information from the Internet.
Your modem, which connects your computer to the telephone line in the street via an ordinary household telephone socket, makes a series of electronic beeps as it dials up the nearest Internet point of presence and puts you onto the 'Net. Via the interface provided by your Web browser, you can now enter the Internet address, otherwise known as the URL, or Uniform Resource Locator, of the file containing the front page of What's On in Bath. As with all other information on the Web, this would start with `http://', for example:
HTTP (HyperText Transfer Protocol) is the name of the Web's own transmission protocol. Web pages are sent over the Internet to your computer courtesy of HTTP.
The difference between the Web and the Internet
Do not make the rather common mistake of confusing the Web with the Internet itself: the Internet simply provides the medium for the Web to run on, just as a telephone line provides the medium for telephone conversations. What the Web does is provide the technology for publishing, sending and obtaining information over the expanse of the Internet. How the Internet actually works may be a matter of interest, but the Web user does not need to know about it in any detail.
ProtocolsA protocol consists of salutations exchanged between computers: `good morning', `be with you in a tick', `file coming down the line now' and so on. Each service on the Internet has its own protocol: its own personal way of sending files around the system. The protocol for the Web is HTTP. File Transfer Protocol (FTP) is another common protocol of which you may have heard.
This fictitious home page would have perhaps a photo showing the famous Roman Baths on the first page, together with a short paragraph introducing the town, and a menu of icons that call up specific information on Bath's museums, parks, bus-tours and so on. These icons provide hypertext links for you to click on-screen `buttons', enabling you to home-in on the information you want. Hypertext links are in many ways the most important feature of the Web. In the case of our imaginary Web pages, clicking on hypertext buttons mostly displays data that is held on computers somewhere in Bath itself. Sometimes, however, a hypertext link may fetch a file from somewhere quite different on the Internet. If you click on the title of a play taking place in Bath, information about the performing company may be fetched across the Internet from a computer thousands of miles away in the United States. From that point on, you might be able to call up pages that tell you about other plays by the same company. With the Web, you can depart on your own private `tour' of information on a chosen subject whenever the inclination takes you. This is by virtue of the hypertext links that span the globe.
You can also pay for goods and services over the Web. You could, for example, decide to order some tickets for a play by filling in a form displayed on a `page' of an online magazine. You might buy a $20.00 seat for `As You Like It' at the Bath Playhouse and then pay by credit card. The example below is taken from a different application on the Web. You can see now what a Web page looks like.
You can see on-screen buttons to click on: these are the hypertext links. There are two types: underlined words and pictures. These buttons serve to fetch up new information across the Internet. Also, you can see the title of the page and a URL at the bottom of the screen telling you from where the information originated.
[FIGURE movienet screen]
Looking at this screen and other pages from the Web, you can see that the Web provides the following:
A means for publishing online information. The Web enables you to lay out text and graphics of the `pages' on the screen, and to insert titles, photos, captions and so on.
A means for retrieving online information via hypertext links, at the click of a button. You can also use Web search programs to find the information you want.
An interface for looking at information retrieved. This is done by virtue of a Web browser.
A Web page may also contain forms for conducting commercial transactions across the Internet and include other applications; for example, spreadsheets, video clips, sound clips and so on.
The Web is, therefore, simultaneously a means of online publishing, a way of accessing, storing and retrieving information, as well as a means of sending, acquiring and querying data across the Internet. Most importantly, the Web allows the use of hypertext links that can take you to any computer on the Internet. All these functions work by using the hardware: the wires, the cables, the computers, the satellite links that are used to send information from computer to computer using Internet protocols.
Before the Web, the Internet was largely the domain of computer nerds and others who delighted in the abstract and concise. Interfaces to early Internet applications required almost mathematical precision to operate, and programs such as FTP and Telnet were purely command-driven. Later, when file-retrieval services such as WAIS and Gopher became popular on the Internet, these had the advantage of being menu-driven, but still were rather cumbersome.
Although the Internet still was using interfaces suitable for the more technically inclined, software for home and office markets long ago departed from command-driven applications. The idea of a graphical interface using Windows, which originated from work at Xerox PARC (Palo Alto Research Center) in the 1970s, was popularized in the early 1980s, primarily through Apple Computer Inc.'s Macintosh computer. IBM Corp.'s Personal Computer and other manufacturers' PC-compatible machines followed suit with the introduction of Microsoft Corp.'s MS-Windows operating system a few years later. At the same time, windowing systems were introduced to the UNIX workstation market, which is now dominated by the X-11 interface. It has, however, taken until the middle 1990s to introduce a simple and reliable user interface for accessing information over the Internet. The Web's popularity has doubtless been in part because it offers a simple point-and-click interface, which immediately makes it more accessible to a much wider range of users. Another important factor must have been email, which, as it has gained popularity, has made the idea of sending information across networked computers that much more acceptable.
What is a URL?
This is a much-simplified explanation for the novice.
Given the fact that there are vast numbers of Web servers, the question remains of how can HTTP possibly locate just the file you want from somewhere on the Internet. The answer is through the use of the URL, the Uniform Resource Locator. This is rather like the telephone number of a computer on the Internet, together with information appended to specify the exact file to be sent to your machine. Taking the URL for our fictitious Web pages in Bath, we can see the general pattern that URLs adopt. Look at
`http' is the name of the Web protocol used to access the data across the Internet.
`www.bath.avon.uk' is the Internet name or the domain name of the computer on which the information is stored. The `www' indicates that this is a World Wide Web server. `uk' is the country code for the United Kingdom.
To be more specific about the file you want from the server in question, you add to the URL a path and a file name. For example:
which would point to a file called `buses.html' in a directory called `time_tables'. We can imagine that `buses.html' contains bus timetables for the city of Bath, which can be conveniently called up on your screen.
URLs can specify files to be accessed using protocols other than `http'. A URL beginning with `ftp://' points to a file to be fetched using FTP - the File Transfer Protocol. Meanwhile, a file beginning with `mailto://' links to an application which allows you to send an email message to a pre-defined address, and `news://' points to a USENET newsgroup and uses the Network Transfer Protocol to transfer data.
There are various conventions when it comes to URLs. Take the letters `.com' for instance. These mean `company' as in: `http://www. microsoft.com'.
Non-profit organizations may use `.org' and educational establishments use `.edu'. Similarly, there are codes for countries. A URL may end in `.us' for the United States, `.fr' for France, `.au' for Australia and so on. Some of the conventions for URLs are listed in Appendices E and F.
How are email addresses different from URLs? Email addresses follow a different format: firstname.lastname@example.org. For example: tiptoes@hawks. uni.edu.
The string of letters following the @ sign identifies the machine to which the mail will be sent, whereas the name of the person who will receive the mail is given directly before the @ sign. There may be more than one person logging on to the machine to read mail. That is why the name of the person becomes important in an email address.
The Internet is a vast network of interconnected computers. Just as AT&T, France Telecom, British Telecom, and other countrywide or regional telephone networks now are joined together to form a global telephone system, so it is with the Internet. The many thousands of computer networks that make up the Internet are joined together on a global scale, so that any Internet computer can communicate with any other.
How can your computer know how to find someone else's computer on the other side of the world? Just as each telephone in the world has its own unique telephone number, so each computer connected to the Internet has its own computer number. This is known as its IP, or Internet Protocol, address. However, because IP addresses consist of long series of numbers that are cumbersome to remember and type, you rarely come across them in everyday Internet use; most people prefer to use the parallel system of naming computers. This is the system of Internet host names, sometimes called Internet addresses or even domain names. Whereas a computer on the Internet may have an IP address such as 220.127.116.11, it may have a more manageable Internet host or domain name such as www.drizzle.org. Servers consult a globally distributed directory to map each host name onto the corresponding IP address.
Once you are connected to the Internet, you can theoretically `dial' anywhere you want. This gives you tremendous freedom. The cost of Internet access remains the same regardless of whether or not you talk to a computer in Australia, England, the United States, or in the next town. This is because you pay only for the time spent on the line to your nearest Internet point of presence. And, just as when you make a phone call to another country, you are hardly aware that your voice is traveling along foreign telephone lines, so it is, theoretically, with the Internet. Although data sent from a computer linked to a network in an office in Paris may travel across constituent networks en route to Los Angeles, as a user, you do not have to worry about the path it takes. The only evidence you have that the requested information comes from far away is when a part of the Internet system is in heavy use and you have to wait.
The French system Minitel is used in ways similar to the Web
It is interesting to note that the French have been using a public system of information access for a number of years, which, in many ways, has occupied the same niche as the Web, although not nearly so sophisticated nor as versatile. Called Minitel, the system consists of hardware in the form of a computer, keyboard and modem originally given away free to subscribers of France Telecom, the national French telephone company. Minitel is very easy to use and is extremely popular - most homes have one. Minitel has existed as a household `pet' for more than 10 years. Using your Minitel, you dial up on the Minitel online service, and then proceed through a number of menus to book yourself a seat on a train, read the latest recipe for la mousse au chocolat &agrav; l'anglaise, inspect your bank balance, book hotel rooms and so on.
The basic components of the Web are shown in the following illustration. They are:
Web servers, which are computers that hold information for distribution over the Internet. In the example application in the diagram, one Web server might hold the text and graphics of the online magazine What's On in Bath, and another server might hold information on which seats are available for a particular concert. The magazine would be formatted using the Web's own publishing language, HTML (HyperText Mark-up Language). The data on available seats and their price would be held in a database with links to specific forms that are published using HTML.
Servers, which can be PCs, Macintosh systems or UNIX workstations: it is the server software that makes them special, rather than the computer itself. That said, servers need to be fairly up-market machines. Servers do need to be left on all the time, so that people can access the information on them whenever they want. Another important point about servers: they are relatively difficult to set up. If you are a non-technical person who wants to publish on the Web, the best thing to do is to rent some space on someone else's server.
Web clients, which can be PCs, Macintoshes and other computers that are connected to the Internet and which can retrieve information from Web servers. A Web client is the computer on your desk. PCs, Macintoshes, UNIX workstations and even simple terminals can run client software. Different client software is marketed (or is given away free) for different platforms. Thus, Mosaic has both a Macintosh and a PC implementation.
HTTP protocol, which is used to transmit files between servers and clients. When you click on a hypertext link or fill out a form in a Web document, the results need to be sent across the Internet as quickly as possible, and then to be understood by a server at the other end. Instructions such as `send me this file' or `get me that image' are carried by the Web communications protocol, HTTP. This protocol is the `messenger' that fetches files to and from servers, and then delivers results to your computer every time you click with a request. HTTP has its counterparts in other Internet services: FTP, file transfer protocol, and Gopher are protocols that obtain different sorts of information from across the Internet.
Browser software, which is needed by a Web client for displaying text, images, video clips and so on. This is supplied under the umbrella name `browser', of which Mosaic, Microsoft Corp.'s Internet Explorer and Netscape Communications Corp.'s Navigator and Communicator browsers are probably the best-known examples. Browser software gives you the ability to scan information retrieved from Web servers, as you would browse through a book. It also gives you facilities for saving and printing information obtained on the Web.
[FIGURE -- globe p.8]
A key point that helped make the Web successful is that it is multi-platform. What this means is that it does not matter what kind of computer you are using; you can still view information published on other, usually incompatible, machines. Thus, a PC user can access information published on a Macintosh; a Macintosh can access information published by a PC; UNIX users will find that they are compatible with everyone. The trick is that each computer in its own way can assimilate HTML. What happens when an HTML document gets to its destination is up to the computer on the receiving end. It can display paragraphs in Helvetica 20-point type if it pleases, and headings in Times 14-point bold type, if that is the font available. Whereas the document is transmitted computer-to-computer in a standard format, individual browsers may display it quite differently, depending on the capabilities of the hardware and software in the computer on the receiving end.
To publish information for global distribution on the World Wide Web, you need a universally understood publishing language, a kind of mother tongue, which all computers on the Web can potentially understand. You also need a commonly understood communications protocol for sending published information `down the wire' from computer to computer. This should enable users to download information to their machine at the click of a button, and also to send back information (your address, a credit card number, a query to a database and so on) with little effort.
The publishing language used by the Web is called HTML (HyperText Mark-up Language). Using HTML, you can specify which parts of your text are to be headings, paragraphs, bulleted lists, and which parts are to be rendered in bold-face type, in italicized type and so on. You can use HTML to insert tables into documents, to write equations, to import images and to format fill-out forms for querying databases at a distance. (Some of these features are specific to HTML 3 and are not supported by earlier versions of HTML.) The HTML language itself is very flexible and not difficult to use, although, as with all tasks associated with computing, patience is a necessary virtue for authoring hypertext. Part of the Web's appeal is that almost anyone with a reasonable PC, Macintosh or UNIX computer can publish information without being unduly technical. Judging by the variety of publishers on the Web today, HTML is within the grasp of many. A simple example of HTML can be seen at the beginning of Chapter 3.
[FIGURE in margin -- long distance jump p.9]
The initial protocol for the Web was very, very simple. The client sent a request: `GET this filename' and the other end sent back the file and closed the connection. And that was it. There was no content type to tell you what kind of file was being sent. No status code. Just the file. The client therefore had to guess what it had been given and this developed into a fine art. First, the browser would look at the file extension to see if there were any clues, such as ˙GIF or ˙HTML, and then it would look at the beginning of the file in case the first few bytes gave the game away - all rather precarious.
Then along came MIME, a kind of multimedia extension to email. This was soon adapted to HTTP so that now, when you receive a file, you actually have a status code and a content type. This content type tells you whether or not the file is text/HTML, video/MPEG, image/GIF and so on, which gives the browser a chance to call up the correct viewers to display the file. The burden of finding out what kind of file it was has now moved to the server. On UNIX and DOS, servers still play this game of `guess the file format'; whereas on the Macintosh, this is stored as part of the file itself.
The HTTP we use today is the product of collaboration between CERN and a group at the National Center for Supercomputer Applications (NCSA) at the University of Illinois at Champaign-Urbana. Innovations since that time included security features such as Netscape's secure socket layer (SSL), and more recently, the ability to keep open the connection to the server so that the server can make multiple requests. The World Wide Web Consortium is looking into ways of improving this protocol, and into methods of text compression that enable information on the Web to arrive at its destination much quicker.
The Web is expanding not only in terms of how much information it holds, but also in terms of the variety of information it holds. The figure below illustrates this general trend. On the left-hand side, you can see the Web as it started, predominantly as a medium for publishing information in textual form, and then it progressed to include photos, diagrams and so on.[FIGURE -- more variety on the web p.10]
Toward the right-hand side of our diagram, we show the latest features at the time of writing: the introduction of plug-in modules, the Java programming language that enables all manner of small applications to be sent down the wire and used within Web applications, and so on.
Many of the new features of HTML are associated with this departure away from the simple document and toward a Web which combines text, graphics, video, audio, applications such as spreadsheets, front ends to databases, virtual reality applications and all sorts of other `objects'.
[FIGURE -- you called sir p.11]
At the time of writing, Java is a very popular buzzword. Java is a programming language that enables programmers to write small applications that are sent over the Web and then executed on the client. It is an object-oriented language related to the C language. What is so special about Java is that programs can be sent to any machine with the right browser to understand them; furthermore, the code will run safely without any risk of adversely affecting the client.
From the user's point of view, Web documents suddenly become much cleverer. It is as though your browser automatically creates features, such as the ability to run a spreadsheet or to play a piece of animation, right in front of your eyes, without you having to load any extra software. The code required to do such tricks is compiled into a special binary format and executed by a Java interpreter in your machine. Java applications are called `applets' (which means `small applications' - it would be another thing entirely to send full-blown application software over the Web) and are small pieces of code that rely on libraries of Java `classes' in the browser. These libraries indicate that the browser has a certain amount of processing knowledge resident in the browser itself: the browser knows how to create a window, respond to events, paint things within a window, draw text within a window and so on. The applets arriving across the Web capitalize on this knowledge and use the library routines to do something useful, such as displaying a simple spreadsheet or a piece of animated graphics.
But, what happens if the applet calls upon a class that is not available at the client end? In such cases, the class has to be fetched over the Internet, a process that indeed may be rather slow. For this reason, it is best to limit the number of these `additional' classes.
Scripts are small programs, which are transparent to the user and which go on `behind the scenes' fine-tuning Web pages in one way or another. The author can write a script in one of several available scripting languages. Scripts themselves may have one of several functions. The classic function is rendering a form to seem `smart' so that it interacts with you as you fill it in. Thus, you might arrange for a script to check each form field in turn to make sure it is properly filled in, and then to advise you if you have mistakenly left out some critical piece of information.
In the future, scripts may be used to create various special effects, such as animation. The later chapter in this book on Scripting examines some of the current and upcoming uses for scripting. Scripts can also be used to tailor components of a particular program, such as a spreadsheet program. In this instance, application software developers will supply off-the-shelf components for a program. Let us suppose that the components are for a spreadsheet program; with a script, you can easily integrate off-the-shelf components into the application, a feat that would have previously required many programmer hours to compile. But with the hypothetical script you have written, this almost seamless integration happens in much the same way as it does with Visual Basic; here, the components are written in a kind of systems language and then, the script enables off-the-shelf components to be integrated into the application itself. If scripts are half as good as they promise, they can save many a programmer's headache.
If HTML can have all sorts of items embedded in it - pieces of animation to be played by your browser, small programs to run on your computer, and even music - then surely the browser must have some way of knowing what kind of beast is coming down the line. A browser can sense what type a file is and therefore it can know what to do with it by using the Internet media type. The Internet media type is a code sent with the file. The browser uses this code to work out what software is needed to interpret and display the data. Some examples include the following: the Internet media type for an HTML document is text/HTML: for a JPEG image, the Internet media type is image/JPEG.
Once the media type has been recognized and the correct software called up to deal with the file, the information in the file can be displayed embedded in the document; that is, it is displayed in an area of the document as though it were simply part of it. Another way of looking at this is that information can be `plugged in' to a document. Indeed, the phrase `plug-in' is commonly used to describe downloading embedded files that are not in HTML, and which require software at the browser end to come to the rescue and display them.
A browser may allow, for example, a PDF (Portable Document Format) file to appear down the wire. The browser sees that this has an Internet media type Application/PDF. It then calls up the Adobe Acrobat reader to display the PDF file. Once the media type is known, the browser loads the correct application software and interprets the file.
Early work on representing documents focused on rendering instructions needed to print the documents. Work by IBM on GML (Generalized Mark-up Language) focused on an alternative approach, whereby standard document structures such as headers, paragraphs, lists and so on were marked up by tags inserted into document text. The emphasis on document structure rather than on rendering instructions, made it dramatically easier to move documents from one system to another whether for display on simple terminals, line printers or sophisticated typesetting machinery.
This work led to the Standard Generalized Mark-up Language, which is an international standard ISO 8879:1986. SGML enables you to define a grammar for marked-up documents that defines the ways in which tags can be inserted into documents. For instance, list items only make sense in the context of a list, and table cells only make sense in the context of a table. SGML's formal way of describing the grammar is called the Document Type Definition.
Global hypertext makes worse the problem in moving documents from one system to another; for example, we have Macintosh systems, PCs, a variety of UNIX boxes, simple terminals and even speech I/O devices for the visually impaired. SGML proved ideally suited for this application. Tim Berners-Lee chose SGML to define the HTML document format for the World Wide Web. HTML is formally an application of SGML. The HTML Document Type Definition (DTD) formally defines the set of HTML tags and the ways that they can be inserted into documents.
To the uninitiated, DTDs may seem rather intimidating. This book tries to act as a guiding hand to explain the HTML mark-up language and how to apply it so that it creates documents for Web publishing. HTML is not a static document format, but is evolving rapidly from its simple beginnings as conceived by Tim Berners-Lee.
One of the areas in which the World Wide Web Consortium (W3C) has shown great interest is how to make the Web accessible to the blind. In the case of HTML, this has involved two things in particular, discussed below.
First of all, the IMG tag for inserting images (as explained in our chapter Graphics on the Web) is to be superseded by an entirely different tag (OBJECT), which enables the browser to display textual mark-up as an alternative when a visual image appears on the screen.
Suppose you have a photograph as part of your Web pages. Browsers will usually display the photograph as the author intended. A browser used by a blind or visually impaired person will, however, be set up to display a paragraph or two of descriptive text instead of the photograph. This text must be included by the author for the benefit of text-only browsers and may contain hypertext links, itemized lists and all kinds of other mark-up. Seeing such text, the browser will read aloud the words to the user and even use appropriate intonation for bold, emphasized text and so on.
By including textual mark-up as an alternative to images, the blind or visually impaired Web user is in a far better position than previously. The still-popular IMG tag allows only very limited alternative text in lieu of images and thus puts it out of favor with those who rely on speech generation.
Second of all, the W3 Consortium is encouraging authors to separate the structural aspects of their documents from those aspects of the documents that are merely to do with layout. Once this has been done, the software that enables the text to be synthesized and read aloud to the blind or visually impaired user has got an easier job to do. Think about it: if a page is marked up solely in terms of headings, paragraphs, lists, and other structural items, then it becomes much simpler to understand the HTML and to render the content into speech. But the moment that extraneous information about font size, alignment of text, margin width, color and so on is mixed in with the general mark-up, then translating that mark-up into a spoken equivalent becomes much harder.
An ardent enthusiast when it comes to making the Web available to the blind is Dr T.V. Raman. A clear and original thinker, Raman - himself blind - has contributed widely to discussions and negotiations with browser vendors on the subject of Web access for people with disabilities, and also is an active participant at the Math Working Group. He has also written the Emacspeak Speech Interface, which is a full-fledged, speech-output system that enables the blind and visually impaired to access the Web with a line-mode browser within a UNIX environment.
A new plug-in for math should be ready by August 1997. This is a simple-to-use piece of software which allows easy publishing of mathematical and scientific expressions on the Web. It is written by Dave Raggett under the auspices of the World Wide Web Consortium.
This book you may notice, does not cover how to mark up math on the Web. In this book's previous edition, math was inserted as a chapter in its own right with the happy anticipation that math soon would become part of the HTML standard. Unfortunately, that never happened, and, although certain browsers have implemented the math spec as it stood, for the main, math on the Web is still in the wizards' pot; to be sure, there is no end of disagreement on the recipe.
Just as HTML has its own Working Group, so does Math. The HTML Math group includes representatives from companies who produce special software for typing and editing mathematical formulae on computers; examples of such software are Waterloo Maple and Wolfram. There are also representatives from scientific publishers such as Reed-Elsevier Ltd, as well as the American Mathematical Society. Each has its own view of the correct way to do things, and Dave Raggett of the World Wide Web Consortium, who did the original HTML Math proposal, has his own ideas, too.
Given the widely disparate views of group members, evolving a standard for marking up math on the Web is a challenge, indeed. Nonetheless, the Math Working Group eventually hopes to come up with a proposal for concrete notation of HTML math, with the initial deployment via plug-ins, which is add-on software that enables browsers to cope with math notation. Meanwhile, mathematicians, physicists, chemists and other scientists are finding it very difficult to use ordinary HTML to express mathematical formulae and scientific expressions.
It is strange to think that the Web originally evolved for the benefit of physicists to facilitate communication of ideas and papers.
Both Netscape and Microsoft are at last implementing style sheets for Web pages.
Chapter 9 in this book explains how to use style sheets, and throughout this book you will see references as to how to use the style sheet language, CSS (Cascading Style Sheets) to obtain the layout effects you want. Cascading Style Sheets potentially gives the user nearly as much control over the look and feel of material as would a conventional desktop publishing package, and so, it relieves HTML of the burden of non-standard extensions, rendering HTML available for its proper role of structuring information. World Wide Web Consortium members Håkon Lie and Bert Bos, with contributions from Chris Lilley, Dave Raggett and others developed CSS. Microsoft is implementing CSS on its Internet Explorer browser.