Welcome to the Wonderful World Wide Web

• Click
• Get Started on the Web the Easy Way:
• Gopher vs. WWW
• Web Does Gopher, and a Lot More...
• Web Provides Marked Text
• Web Browsers Understand HTTP
• Some Browsers Use External Viewers
• There's a WWW conference in Chicago this fall:
• Reports of Gopher's Death are Premature
• WWW at the ADN
• How to Jump on the WWW Bandwagon

Click, Switzerland. Click, Seattle. Click, Sweden. Click, Paris and a tour of the Louvre. Click, Silicon Valley and the latest offerings. Click, click -- order flowers, see maps, hear sound bites, view movie clips, read electronic books, analyze the federal budget, search the world for whatever it takes to scratch that itch. And still be back in time for lunch.

The World Wide Web or WWW makes it easy like never before. The Web was developed at CERN (a high-energy physics lab in Geneva) to facilitate sharing of information. Its popularity, and the number of information providers, has soared over the past year due to some high-quality Web browsers, notably Mosaic from the National Center for Supercomputer Applications in Urbana. Mosaic and the Web have been written up in the popular press, and have attracted serious attention from some no-nonsense companies. There may well be some truth to the claim that Mosaic is the first "killer application" of the Internet, useful to beginner and expert alike.

Fundamentally, the World Wide Web is a worldwide hypertext system. When you read a hypertext document, certain words will appear specially marked. Point to a marked word, click your mouse, and additional information that the word pointed to is displayed. The additional information may be another location in the document you're reading or it may be in another document entirely. Most hypertext systems put all the necessary files on one machine. But the WWW documents are scattered on Web servers throughout the world; your next click could take you next door or to the next continent.

Hypertext is conceptually very similar to books with footnotes and bibliographic citations. But when you encounter a reference to a cited work in a book, you must physically travel to the library to check out the citation. In a Web document, on the other hand, the citation might include the electronic location of the referenced work, and a click of the mouse will bring it to you. If the author of the document has done a good job marking citations, this is a significant convenience.

As if hypertext among multiple servers wasn't enough, the Web's concept of a "document" includes mixing words with pictures, sounds, and video clips. In fact, it's possible to mark pictures as well as words with hypertext pointers to other documents. There are some truly amazing examples of what can be done on the Web in the "Best of the Web" selection at the bottom of the ADN Web home page. But do keep in mind that your appreciation of the dazzling display might be somewhat limited if your WWW client (usually called a "browser") doesn't do videos or sound clips.

Mosaic is also part of the Network Services Kit for Windows and for Macs. Call the CSO to arrange to have it installed on your desktop computer, and you can have the world on your desk.

Be careful -- it's habit forming!

Gopher started as a simple system to make documents available remotely. Its simplicity is a serious virtue. The ease of constructing gopher clients and servers led to its rapid proliferation, as gophers are wont, and today there is probably more useful information stored on gopher servers than on generic WWW servers.

The balance between Gopher and WWW is changing, though, and it will not be long before WWW is the preferred method for both clients and servers. The reason is that WWW browsers use four key ideas that retain a lot of Gopher's simplicity, yet are easily extensible in ways that will keep WWW current in technology for years to come.

How does a WWW client know what protocol to use in each instance? Each marked word or phrase in a WWW document is associated with a pointer that leads to another item. The pointer is called a Uniform Resource Locator, or URL. The protocol to be used for the new item is included in the URL, along with its name and exact location. (See "What is a URL?".)

Markup, at its best, is used to mark the function of the words. This leaves the rendering of the words to the client program, which can take best advantage of available hardware and selected user options. Thus, when you view an HTML file with a Web browser, the browser uses the markup to decide how to display the file. Titles are displayed in one font, headings in another; some browsers (Mosaic in particular) allow you to select the style and size of font it uses. Lists are formatted as lists; hypertext pointers are highlighted in some way underlined or maybe bolded; and the text is sized specifically for the window it is in. Often there are additional things the browser can do with the HTML file -- you can ask it to show you the original HTML file, to convert it to a plain text file that you can save on your disk, to generate a PostScript file for printing, or possibly something else.

But unlike Gopher, HTTP relies heavily on the MIME (Multipart Internet Mail Exchange) format used to send sound, video, graphics, binary files, and even regular text through electronic mail systems. HTTP first tells the client what kind of files it is about to send, and then sends them. Since the client knows, for example, that the first file it's receiving is really a sound file in .AU format, and the second file is really a graphic image in GIF format, it can properly display them.

HTTP goes beyond simply labeling the types of files that are sent. When a client makes a connection, it typically sends the name of an interaction "method" before sending a request for a particular file. The "get" method is most common, and it means to simply fetch a file. But another method called "put" can be used by the client to send information to the server. For example, by using the put method, a client can send a sophisticated database query to an HTTP server, and have the server act on the query by activating a special program. All kinds of custom interactive applications can be built this way, using the generic functions of HTTP.

To be fair to Gopher, there is an improved Gopher protocol, Gopher+, that addresses many deficiencies in the original gophers. But the advantages of WWW combine in a natural and extensible way that Gopher+ can't quite match. The fact that good WWW browsers are now becoming widely available has probably turned the tide in the long run.

• Good hypertext is hard and very labor intensive. A human brain has to decide how to mark the text, and that takes time. Sure, there are automatic conversion programs for many formats, but a good, clear document design still depends on real Homo Sapiens, not just software. It is much easier to make plain text available, and gopher has a natural simplicity in dealing with plain text.
• The Gopher menu format might seem rigid, but it forces every entry to have an associated title. Reasonable titles are clear and descriptive, and they can be searched by programs like Veronica. URLs embedded in HTML documents, in contrast, are not required to have any associated description at all. The item a URL points to might have a title, but you won't know what it might be until you fetch it.
• WWW documents tend to become collections of small subdocuments, all pointing to each other, whereas Gopher documents tend to be larger, more cohesive pieces. This is perhaps an overgeneralization, and neither scheme really forces a particular style of document organization. The multiple subdocument structure is well suited to hypertext manuals, where instant look up is valuable, but it is less useful when one wants to print out the whole set to get an overview. Also, the multiple document collection needs to be well thought out, because each document will be read in many different contexts.

Mosaic is part of the Network Services Kit for Macs or MS- Windows. Mosaic for X Windows is available from ftp.ncsa.uiuc.edu. Lynx for UNIX/VT100 is available for public use on tigger (enter lynx), or you can get a copy for your own workstation from the anonymous FTP site ftp2.cc.ukans.edu, in the directory pub/lynx.

How about making your own information available on the Web? It's quite easy to make and test a simple HTML file there's enough information in "A Small Sample of HTML" to get started, and "Do You Want to be a WWW Server Yourself?" explains how to make your HTML files public. We may be able to advise you on some tools to make larger jobs fairly straightforward, too. And, of course, if you find something interesting on the Web, please let us know. We're always on the lookout for more useful information to be added to our growing collections.

Comments are welcome; please send them to
Robert Goldstein, bobg@uic.edu