|Chapter 4. The World-Wide Web
4.1 The Internet and the World Wide Web
4.2 The Origin of the World-Wide Web
4.5 Client-Server Paradigm
4.7 Uniform Resource Locators (URLs)
4.9 File Transfer Protocol (FTP)
4.10 Websites, Homepages, and Web Rings
4.11 The Last Word on URLs
4.1 The Internet and the World Wide Web
In spite of all the interest in the Internet and electronic mail, most of the latest excitement about the Internet has been because of the newest development -- the World-Wide Web. First, let us be clear about the difference between the Internet and the World-Wide Web. The "Internet" refers to the physical side of this global computer network plus the TCP/IP software and associated protocols for sending things around on the Internet. The "World-Wide Web" refers to a body of information -- an abstract space of knowledge available on the Internet. The World-Wide Web is a collection of information available from literally thousands of places in the world ... all accessible as if it were coming from some common location or set of locations.
It has been said that if the Internet had not existed, it would have had to be invented to accommodate the World-Wide Web. There had to be some way to transmit the enormous amount of information on the Web from one place to another quickly and easily. The Internet had already accomplished getting the infrastructure into place for that purpose.
4.2 The Origin of the World-Wide Web
Unlike most of the major developments in the history of the Internet, the World-Wide Web did NOT come from the United States. In fact, the origin of the Web is with a collection of European high-energy physics researchers located at CERN (Conseil European pour la Recherche Nucleaire). CERN's members are located in a number of different countries. Although not computer scientists, they had become adept at using electronic mail and FTP (see section 4.9 below) for transmitting information among all the CERN participants. But, this was time-consuming and not particularly easy to do. They thought that there must be a better way to share scientific information (including charts, figures, and illustrations) among the scientists. Furthermore, they really wanted a way -- when looking at the information from one experiment -- to be able quickly to access information about any of a number of related experiments.
In 1989 Tim Berners-Lee came from England to join CERN. He considered the CERN problem of information sharing. He proposed a hypertext system that became the forerunner of today's World-Wide Web. Tim created the first Web software in October 1990. A system similar to what we now know as the World-Wide Web was first made available outside CERN on the Internet in the summer of 1991. Berners-Lee's concept was that of a means of sharing research information and ideas effectively throughout an organization. (In 1994 Tim joined the Laboratory for Computer Science at MIT as Director of the World-Wide Web Consortium which coordinates Web development worldwide.)
The underlying aspect of the Web that is most important is its use of "hypertext". Hypertext has actually existed as a concept a long time ... even before the Internet. In fact, the notion of what we now call "hypertext" was first discussed by Dr. Vannevar Bush, Director of the United States Office of Scientific Research and Development, in an article entitled "As We May Think" that first appeared in the July 1945 issue of The Atlantic Monthly. Ironically, this article appeared at about the time that early research was being conducted that would lead to today's modern electronic digital computers. Bush did not refer to his idea as "hypertext". Instead, he talked about a device and coined the term "memex" for it ... that would allow an enormous quantity of information to be stored and accessed in an associative fashion in nearly exactly the way hypertext works on the World-Wide Web. In one of those delicious ironies of science, it took over 45 more years for the concept of computer and hypertext to get together in the World-Wide Web.
A good way to envision hypertext is to imagine reading the encyclopedia. You stumble upon an article in the "W" volume about World War II. While reading that, you see a reference to US Army General George S. Patton. This momentarily catches your attention. You put down the "W" volume in favor of the "P" volume and begin reading about General Patton. While reading that article, you read about Patton's work as a tank commander and become interested in learning more about tanks. You put down the "P" volume momentarily and begin reading an article in volume "T" about tanks. What you have been doing is quite illustrative of hypertext. A reference in one article (World War II) has diverted your attention to another article (Patton) in which another reference (tanks) has sent you off to read yet another article. The terms "Patton" and "tank" are functioning exactly like hypertext works -- allowing you to leave one body of information in favor of something temporarily more interesting. Of course, you can finish the tank article, return to read the remainder of the article on General Patton, and finally return to complete your reading of the World War II article. This is the way that hypertext is often used.
The Web relies on hypertext as its means of interacting with users. Hypertext is basically the same as regular text. It can be stored, read, searched, or edited. It does have an important exception -- hypertext contains connections within the text to other documents.
Hypertext links can be followed with ease. After reading all or part of the information on any page, the Web user has the option to follow another hypertext link or return to the page from which the current one was accessed.
Hypertext links are called "hyperlinks". The major organization of the Web is determined by the hyperlinks on each Web page. Obviously, documents can themselves have links and connections to other documents. Continually selecting links takes you on a free-associative tour of information. Hyperlinks can create a complex virtual web of connections. In fact, if you are accessing Web pages from several different geographic areas and drawing lines from one Web page to another, the diagram you are drawing after a while can begin to look like a series of lines (not unlike the strands of a spider web) connecting the various sites where Web pages come from. This image of a set of connections looking like a spider web and the fact that Web pages often have all sorts of different other Web pages from which they can be reached, creates a situation not unlike that of a spider web. Thus, the term "World-Wide Web" seems quite appropriate.
4.5 Client-Server Paradigm
How does all this "clicking" and "appearing" work? What is it about the Internet and Web that allows a click on a hyperlink to cause a Web page to appear? The principal characteristic of the Internet that allows this is the so-called "client-server paradigm". The client-server paradigm has been a stalwart of the Internet from its very inception. It pre-dates the Web. The basic idea is that the Internet consists of millions of interconnected computers. Some have information that others want.
Let us think of a cafeteria line as an example. A person moving through the line is a "client". She desires certain types of food to be placed on her tray on request. For example, she asks for a plate of roast beef and creamed potatoes. She waits for the person behind the counter, the "server", to retrieve the roast beef and creamed potatoes and to place it onto her tray.
This is exactly the way the Internet works. There are appropriate servers and clients for virtually every type of information transfer on the Internet -- including electronic mail. But, let us investigate specifically how this process works for the World-Wide Web.
A "file" in computer terminology is a collection of related information. For example, suppose that a file is information about travel destinations in Spain. Each line of the file contains information about one location. All of the lines together constitute the whole file. Let us now see how this relates to sending information from server to client on the World-Wide Web.
Any computer on the Internet with one or more Web pages (files) must have some software running referred to as "Web server" software. Because of the demands for Web pages from all over the world at any time of day, Web server software really must run 24 hours per day and 7 days per week. Otherwise, when a Web page is requested, the requester will get no response and is unlikely to ask again for a Web page from that location ... and is likely to say bad things about the location to his friends.
Thus, Web server software runs continually on an Internet computer ... with no purpose "in life" except to wait for a request for one of its Web pages. Then, it springs into action -- retrieving the Web page and then sending it over the Internet to the computer that requested it.
There is sometimes some confusion about whether the term "server" refers to software or a computer. Actually, the answer is both! The Web server software at the MPSNet Cancun website may certainly be referred to as MPSNet Cancun's Web server. But, this software is running on the Internet node computer www.cancun.mpsnet.com.mx. (Thus, their website is accessible via http://www.cancun.mpsnet.com.mx). So, in a sense www.cancun.mpsnet.com.mx is the server computer. So, it is perfectly acceptable to refer to the Web server software being run on the Web server computer ... thus, referring to both software and computer as the "server". Context should make it clear which is intended in any instance.
Client software is any program (such as a Web Browser which will be discussed in Chapter 4) that requests information from a server. Thus, the typical Web browser (such as Netscape Navigator, Netscape Communicator, Microsoft Internet Explorer, etc.) is a client program. In response to a mouse click by the person using the browser, this client program sends a message (just like a small electronic mail message) to the appropriate server. It requests that a Web page be sent and then waits for the arrival of the page.
Upon the arrival of the requested Web page, the client software displays the page for the perusal of the person using the browser. Clients and servers communicate via TCP/IP over the Internet. Thus, the World-Wide Web is intimately tied to the Internet. It is the communication path that clients use to request Web pages from servers, and that servers use to send those Web pages to clients.
As discussed above, there is sometimes some confusion about whether the term "client" refers to software or a computer. The same situation applies as for the term "server". The Web browser running on your computer may certainly be referred to as your client software. But, this software is running on your Internet node computer which is acting as a client computer. So, it is perfectly acceptable to refer to the Web client software being run on the Web client computer ... thus, referring to both software and computer as the "client". Again, context should make it clear which is intended in any instance.
Another useful feature of client-server interaction is what we call a "persistent connection". Have you ever noticed that it may take a long time for the first Web page to arrive from a server, but subsequent Web pages from that same server seem to arrive somewhat faster? One reason for this is that clients and servers may establish a persistent connection so that the communication path between them is left open after the first page is sent. It should not be surprising that after looking at a Web page from a server, there is a very high probability that the next Web page requested might be from that same server.
The World-Wide Web has come along at a very fortuitous time. There had been other attempts to collect a large quantity of information and make it available on the Internet, but these really had not caught the attention of most people outside the scientific community. One of the major problems was that textual information (that is, letters and numbers) is a limited medium through which to communicate. Unfortunately, previous "large quantity of information" attempts had been mostly limited to textual information. Typical electronic mail is good example of this limitation. Until recently, most of us could only communicate words via electronic mail. It is only recently that the capability of electronic mail has evolved to allow us to communicate pictures and even sounds.
At the time that the Web was becoming popular, another activity was happening simultaneously. Research into digitizing information was experiencing tremendous progress. Virtually all forms of media are now yielding to digitization efforts. For example, any of us who have listened to a Compact Disk (CD) version of a song were in fact listening to the result of translating the original music into a series of binary digits (bits) -- ones and zeros. This digitization can be quite precise and can reproduce the sound later with almost breathtaking sameness. Digitized sound is not subject to "wearing out". Even more important from a computing standpoint, digitized sound can be stored in a computer file of ones and zeros and transmitted from one computer to another.
Simultaneously, research was proceeding into digitizing images. Suppose that an image appears to be a beautiful deserted island basking under the gentle glow of the setting sun. But, in reality the file that produces this image is nothing more than a table of row, column, and color numbers. Any image (photograph or drawing) can be placed in a scanner that moves across its surface from left to right and top to bottom breaking the image into a table of dot-like "picture elements" (called "pixels"). The scanner decides the best color to use for each pixel and builds a table of row numbers, column numbers, and pixel color numbers. For example, row 1 column 1 might be a dark brown pixel. Row 3 column 12 might be a light yellow pixel. Row 25 column 254 might be a pale green pixel, etc. How many colors may be used? That depends on how many binary digits (bits) can be used for each pixel color. Using only one bit will work fine if the colors are all either black (0 bit) or white (1 bit) and will lead to a very small file. But, most images employ more colors than just black or white.
Three bits will allow eight colors, sort of like this:
So, for example, if the pixel in row 25 column 254 is green, the digitized file would contain 010 for that pixel. Obviously, most pictures (particularly photographs) contain more than two, or eight, or 256, or 1024 colors. They contain millions of colors. By expanding the number of bits, it is possible to digitize an image so precisely that its rendering later is nearly indistinguishable from the original. The only price paid by using a lot of bits to represent colors is that the file (even for a small picture) can be quite large. Why do we care about the size of a file that represents a digitized image? Because on the Web that file must travel from the computer where it is now (the server) to the computer where someone wants to see it (the client). With a slow telephone connection, a large image file might take several seconds to transmit from one place to another. And even though several seconds seems very quick, Web pages often have five or six images (or more) -- meaning it might take more than a minute for all the image files to arrive and to be displayed.
Thus, images on the Web represent a continual push-pull between using large images with millions of colors versus using smaller images with just a few colors. The former leads to a very satisfying look, but can be frustrating in respect to the amount of time for all the images to "arrive". The latter can generally be viewed very quickly, but may not have the professional look we have come to appreciate from magazines and television images.
Since it is now possible to digitize sounds and images, the next step is obvious. Combine the two to create digitized video. Obviously, if the size of single image can often be a performance issue, a video will contain thousands (or more) of such images along with sound. So, again the speed consideration becomes important.
There are basically two ways to handle video on the Web. The first (and historically original) way is to send an entire video file from one computer (server) on the Internet to another. The receiving computer (client) waits for the entire video file to arrive before attempting to "play" the video. This has been the tradition with most media on the Internet ... wait to get the file before doing anything with it. This still works fine for small video files, but presents a problem for lengthy (or live!) video. A lengthy video file can be enormously large -- in fact, far too large to fit on the disk of most computers. The waiting time can also be quite annoying. The newer, more modern solution is what is called "streaming video". In this situation, the video file is not delivered as one monolithic entity. Instead it is broken into smaller packets. Each packet is delivered from the server computer to the client computer. When enough packets have arrived at the client computer, it can begin playing the video. The beauty of this arrangement is that, as each packet is used, it can be discarded. There is no need to keep the entire file on a disk. Obviously, this is the only sensible solution (and much like television works) for live video. By definition one cannot wait for ALL of live video to arrive to begin viewing it. This same technique can be used for "streaming audio" as well.
4.7 Uniform Resource Locators (URLs)
The World-Wide Web consists of a tremendous amount of hyperlinked multimedia information located all over the world. The obvious question is how does one computer tell another that it wants a certain Web page. This is accomplished by using something called a Uniform Resource Locator (URL). A URL is exactly like an address. It uniquely identifies some Web page, resource, or service.
For example, the URL for the MPSNet Cancun website is
Let us examine this URL. First, the term "http" stands for HyperText Transfer Protocol -- yet another of the many protocols that make the Internet work. Specifically, the HyperText Transfer Protocol (HTTP) is used to send World-Wide Web information from one computer to another. The protocol is usually followed by a colon and two slashes. Next comes the name of the Internet node on which the Web server is running. Thus, there is a computer (most likely located in Cancun, Mexico), named www.cancun.mpsnet.com.mx. Recall from the client-server section above that both this computer and the Web software on it may be referred to as MPSNet Cancun's "web server".
The relationship between HTTP and TCP/IP is a situation similar to sending priority mail. Generally the item is first put into a plain envelope and then that envelope is put into another envelope as required by the priority mailing service (like Federal Express in the United States). When the outside envelope arrives at its destination, someone opens it and removes the inside envelope. It is generally marked in some way so that the recipient knows what to do with it. That is exactly what happens with the Internet with cooperating transmission protocols.
The Web page is first put into the HTTP envelope. Then the HTTP envelope is put into the TCP/IP envelope. The TCP/IP envelope functions exactly like the priority mail envelope. It contains all the information and functionality to allow the file to travel from one computer to another on the Internet. It really does not matter what is inside the TCP/IP envelope. It could be a Web page, or email message, or data file being sent from one branch of a business to another. Being inside the TCP/IP envelope allows it to travel in the appropriate manner along the Internet. Now, the TCP/IP envelope arrives at the computer which requested the Web page. The computer opens the TCP/IP envelope and removes the inner envelope. It is an HTTP envelope. The receiving computer knows exactly what this is and hands the contents of that envelope over to the appropriate software (a Web browser, see the next chapter) so that the Web page can be viewed.
So, that explains the "http" at the beginning of most URLs. But, some URLs do not begin with "http". Some start with "gopher" or "ftp" or something else. The next two sections discuss Gopher and FTP.
Gopher is an Internet-based information system that pre-dates the World-Wide Web. There is still a large amount of gopher information available on the Internet, but production of this material has almost come to a halt because of the popularity of the World-Wide Web. The gopher system features menus of items that can be selected by typing the appropriate letter or number or by clicking on the appropriate line with a mouse. But, gopher was predominantly a text-based system. Nearly all gopher information placed on the Internet consisted of just letters, numbers, and punctuation -- creating hardly the excitement engendered by today's visually-appealing (and even occasionally auditorily-appealing) Websites. Gopher was a useful system, but just did not generate much enthusiasm in the Internet community and certainly did not excite the non-Internet public.
But, all the existing Gopher information is still available using the Web. Any URL such as gopher://soochak.ncst.ernet.in requests information from the Gopher server (rather than the Web server) at soochak.ncst.ernet.in (a gopher site of travel information in India). Since gopher material is textual, this can easily be displayed by a Web browser ... even if it is usually not as visually-exciting as Web material.
4.9 File Transfer Protocol (FTP)
FTP (which stands for File Transfer Protocol) is also a system which predates the Web. Suppose that you have written a report on travel in New Guinea and want to make it available to your friends. Before the emergence of the Web, there were two main ways to make such information available. First, of course, was the possibility of sending this report to friends via electronic mail. The obvious question is which friends would be interested and do you really want to send out lots of (perhaps unwanted) email messages containing a long report. The other way was to use the File Transfer Protocol (FTP).
With FTP, you would copy the report to a special directory (or folder) on your computer. In fact, this would have to be a directory (or folder) accessible to an FTP server. Then, you would make your friends aware of the report via electronic mail or some other means. Then anyone interested in the report could use her FTP software (her FTP client) to connect to your computer (and FTP server) in order to obtain her own copy of the report. FTP is a very simple-minded system. In its typical use a file is simply copied from the remote computer to the local computer. That is, if one of your friends wants a copy of the New Guinea travel report, she uses her FTP client on her local computer to request that file from your FTP server on your remote computer.
All the FTP process does is to copy the file to the local computer. It is up to the person at the local computer to decide where to store that file and what to do with it upon receipt. For example, if your report is a Microsoft Word document, then the recipient might direct the guinea.doc file to be placed in the directory (folder) where she usually keeps her Microsoft Word files. Then, she could start Microsoft Word and import the file to read it.
There are two basic ways in which the FTP process works -- (1) public FTP and (2) private FTP. (1) In public FTP, any file that is available for FTPing on a computer may be accessed by anyone who simply knows the address of the file. No account number or password is required of the client to copy the file. You might want to do this with your New Guinea travel file if you really do not care who makes a copy of it. (2) In private FTP, the client must provide an account and password in addition to the address of the file. This allows file access only to those individuals to whom you have given the appropriate account and password ... protecting the file from general access.
The most popular software for FTPing files these days is visual software that shows the files of the local computer on the left and the files of the remote computer on the right. File transfer can be done by clicking the mouse on the desired file and then either dragging it over to the local files or by clicking on a "transfer" button.
By the way, FTP provides a lot more capabilities than what has been discussed here. In some instances files can be transferred from the local computer to the remote computer (although this is rarely allowed in a public FTP situation). Also, it is possible to transfer groups of files by clicking on several of them at once and then dragging the whole group or clicking on the "transfer" button.
A URL such as ftp://wuarchive.wustl.edu/multimedia/images/jpeg/m/marvin.jpg will work with a Web browser. It simply requests the file from a public FTP site and brings it to the client computer. Some files can be viewed within the Web browser (such as pictures, sounds, and video). Others are handed over to "plug-ins" -- software associated with the Web browser to handle files that are not generally part of Web pages. Finally, if all else fails, the Web browser will simply ask where on the local computer a file should be stored. Then, the Web browser can be exited and the file can be "viewed" with some other appropriate software.
4.10 Websites, Homepages, and Web Rings
Generally, when creating information on the World-Wide Web, this will consist of several Web pages with links among them. This is called a "Website". The Website beginning at http://www.travelsites.com consists of several Web pages concerning such things as "Where to Go", "How to Get There", "Where to Stay", "I Need a Vacation", "Tickets Now", and "Travelsites Membership". The entire collection of inter-connected Web pages may be thought of as one considers a travel brochure or a magazine -- lots of related pages in one location or Website.
In most cases when a URL appears in a book, magazine, newspaper, or even on another Web page (such as http://www.travelsites.com) this URL addresses the Web page that the creators of the Website want you to visit first. This is very similar to the cover of a brochure or magazine -- the page that typically has a link to each of the others and is the place to which you return to begin exploring other aspects of the Website. This cover page is referred to as the Website's "Homepage". Thus, a homepage is typically the Web page by which you should enter a Website.
Sometimes, a group of related Websites (for example, 15-20 travel sites in Spain) will contain links to each other. This creates what is called a "Web Ring" -- a collection of related Websites. Participating sites on a Web Ring are in effect jointly publicizing each other. Someone who visits one is just a click away from any other. This cooperation can be useful to avoid redundancy -- so that a site on a Web Ring need not repeat information that is available at another. It also usually leads to more visits for each of the "ringed" sites than if they were not part of the Web Ring.
4.11 The Last Word on URLs
The URL discussed above (http://www.travelsites.com) is not complete. In fact, few of the URLs used above are really complete URLs. That is because a URL should tell not only the protocol (http) and the location of the Website (www.travelsites.com), but also the name of the file being requested.
A Web page is represented by a file of text that tells what words are to appear on the page, what images are to appear on the page, what hyperlinks are to appear on the page, and how all these items are to be displayed. In the vast majority of the cases these files are created using something called the HyperText Markup Language (HTML). HTML is a "markup language" in the sense that a file created using HTML contains a number of tags such as TITLE -- that tells the title to be used for the page and BGCOLOR -- that tells the background color to use for the page.
Most files on the Web that tell how a Web page is supposed to look are HTML files and end with the file extension .html. So, a really complete URL would look something like
That is, at the Travelsites website this requests the file wherestay.html -- which is a file created using the HyperText Markup Language and is presumably about where to stay when traveling. (At least that's what "wherestay.html" would suggest to most people!) Thus, http://www.travelsites.com/wherestay.html is a complete URL.
You will usually see complete URLs that end in the file extension .html, but occasionally may see some that end in .htm. This is also an acceptable file extension for an HTML file. This is allowed because there are still some computers with files available to Web servers that use an old file naming system known as the "8.3" system. In this system each file name could consist of a first word of no more than 8 characters and an extension of no more than 3 characters. If the Travelsites website were on a machine that used the 8.3 system, then the URL http://www.travelsites.com/wherestay.html would have to be something like http://www.travelsites.com/wheresta.htm.
Why does the URL http://www.travelsites.com "work"? That is, why is the Travelsites Web server able to send a file in response to this incomplete URL? Because, every Web server has a default file that it delivers. In the case of the Travelsites Website this file is one named index.html. So, the URL http://www.travelsites.com is equivalent to the URL http://www.travelsites.com/index.html and requests the Web server at Travelsites to send the client the index.html Web page. Nearly all World-Wide Web servers have a default file like this. That is why so many URLs can be advertised or used with just a few words and without the .html file at the end. Any URL like http://www.travelsites.com is really a "shorthand" way of saying http://www.travelsites.com/index.html or http://www.travelsites.com/homepage.html or http://www.travelsites.com/welcome.html. The decision is made by the person in charge of the Web server (called the "Webmaster") what file name will be sent if someone fails (as is often the case) to request a specific file name.
Sometimes a URL can be quite long ... such as
already mentioned above, or
What does a URL like this mean? Let us examine the latter in more detail. The "http" for HyperText Transfer Protocol looks ok. We can also tell that there must be Web server software located on a computer named www.travelsites.com. Finally, the file this requests will be named qantas.html. But, what do we make of the word InternationalSites. This is a directory (or folder) name. Here is the way this works: When this URL is sent to the Web server at www.travelsites.com, it proceeds to the standard directory (folder) where Web pages are stored. But, instead of looking for the page qantas.html, it looks next for the sub-directory (sub-folder) named InternationalSites, goes to that directory (folder) and retrieves qantas.html there. By making several sub-directories (sub-folders) Website designers can keep related material in separate locations. But, this can sometimes lead to silly looking URLs like
Luckily, most such URLs are available by clicking on a hyperlink and do not require typing the whole thing in to the Web browser.