But what if such a "stack" is allowed to span many machines; indeed, what if it may be distributed over the entire Internet? That's what a scientist named Tim Berners-Lee envisioned in mid-1989 while at the European Laboratory for Particle Physics (CERN) in Switzerland. While addressing a local demand -- CERN's need for improved communications within the Laboratory -- he sought a solution to a more general problem. He initiated and headed what became known as the World Wide Web project, aimed at creating the basic architecture to support Internet-based distributed hypermedia -- truly global hypermedia.
World Wide Web (WWW) came to describe the subset of the Internet that serves hypermedia documents to the whole Internet. On the Web, information providers use server programs that comply with HTTP (hypertext transfer protocol) to serve information to client programs (Web browsers) elsewhere on the Internet. On the server side there may be documents of all kinds, including documents in a special text description format called HTML (hypertext markup language) that facilitates the expression of hypertext. Links within HTML files may reference other documents anywhere on the Internet through the use of text strings known as Uniform Resource Locators (URLs). On the client side, browsers display HTML files and possibly other document types, allowing users to navigate from site to site in search of information by following these links.
By January, 1993, a few Web browsers existed but were underpowered, and there
were only 50 known HTTP sites on the Internet; these were very related problems, because a large Web user base would motivate better browsers and vice versa. The infrastructure for the WWW was in place, but something was missing from the client technology that prevented people on both sides from taking full advantage of it.
A new browser emerges
Free browser software called NCSA Mosaic would soon fill the void. Mosaic was originally developed by the Software Development Group of the National Center for Supercomputing Applications (NCSA) here at UIUC last year. A version for Unix (running the X Window System) came first, followed by versions for Windows and the Macintosh. The key features that distinguished Mosaic from earlier attempts at browsers were a compelling graphical user interface, pioneering support of inline images, external handling of non-HTML data types such as Postscript and video, and smooth integration of other Internet services (such as FTP, news, gopher, and WAIS) into the program.
Mosaic's good looks and great personality made Web navigation easy and Web publishing a compelling option for information providers. It certainly played a big role in growing the Web: By the Fall of 1993 the number of HTTP servers had jumped by an order of magnitude since the beginning of the year to 500, and there are now at least 2,000; downloads of Mosaic from NCSA's FTP server number over 2,000 a day.
Past and present ACM members have been actively involved in the development and support of Mosaic: former SigMusic chair Chris Wilson, now at Spry, Inc., co-authored the initial Windows version with former PC lab manager John Mittlehauser, now at Mosaic Communications Corp.; ACM chair Alan Braverman supported the initial X version and now develops for it with former SigUUCP chair Dave Thompson; SigGraph chair Ryan Grant develops for the Windows version; SigNet chair Jay Kreibich develops for the Mac version; Social chair Jeff Thompson and Jason Marshall support the X and Windows versions, respectively; and vice-chair Ed Burns works in many capacities with the X group.
The refinement of the freely available Mosaic continues here, while about a dozen corporate NCSA licensees are working on or have released commercial versions of the program. NCSA has also released HTTP server software for various platforms and plans to release additional tools in the future (see the interview with Kim Stephenson on page 7.)
Getting onto the Web
You may access Mosaic for X from within X on all Engineering Workstation machines (execute "Mosaic" from a command line), or from the Instructional Lab Sparcs in 1240 DCL. The Macintosh and PC versions may be found at CCSO sites across campus, usually in a folder or Windows program group called "Mainframe Connections". (Note that some CCSO PCs, such as those at the Illini Union site, do not run Windows.) You may use all three versions of Mosaic at the ACM office in 1225 DCL if you are an ACM member.
Using Mosaic is quite straightforward. Upon running the program you will enter the "home page" as designated by your machine's administrator; the home page is a jumping-off point of sorts, and from here you can typically go to other documents on the Web by following links that branch from the home page. Links are represented as "anchors" -- underlined, highlighted text regions (and sometimes inline graphics) -- on Mosaic, and to follow them just click the mouse on the anchor. If you wish to jump to a Web document that does not branch from the page you're looking at, select "Open URL..." from the "File" menu and type in the URL to which you wish to jump; we'll mention a few interesting URLs later.
If you have trouble getting started, you'll find that Mosaic's online help is extensive and just as easy to use the program itself; in fact, the manual (accessible under the "Help" pulldown menu) is itself a cluster of the Web.
Carrying On
Perhaps the most common use of the Web to date is as a means for institutions, student organizations, and other groups to set up and publicize their own home pages, allowing for centralized retrieval of group-related information for members and non-members. The ACM at UIUC has its own Web server, for example; give it a try at the URL http://www.acm.uiuc.edu. You'll find information about the various SIGs and their projects, information about ACM members, and even online back issues of the Banks! Dozens of other organizations at the University have set up home pages. You'll want to start at http://www.uiuc.edu, the University's main Web server, if you wish to explore all of these.
Personal home pages for individuals have sprouted up, too. Some list interests, current project descriptions, and favorite quotes, hypertext analogs to what are provided through ".plan" files and the "finger" service; other home pages go further, providing political commentary, randomly generated quotes, photos of their subjects, and even links to résumés. You'll find a list of some student home pages at http://www.cen.uiuc.edu/~mg7932/uiuc.people.html.
Another frequent use of the Web is for staging multimedia exhibits. The Library of Congress, for example, has sponsored educational exhibits on topics ranging from the Dead Sea Scrolls to the archives of the former Soviet Union, combining the same text, graphics, and sound that you'd find if you visited their "real" exhibits in Washington, D.C. A Web page called the "Expo" serves links to a variety of multimedia exhibits; visit it at http://sunsite.unc.edu/expo/ticket_office.html.
This process of jumping from document to document may be called "browsing," but thanks to innovative Mosaic features such as "forms," the Web experience is increasingly an interactive one. Forms are simply collections of input items that can appear in HTML documents: editable text boxes, lists of things to choose from, buttons, and so forth. The reader may fill out a form and submit it back to an HTTP site, where it is processed by a program complying with what's known as the Common Gateway Interface (CGI) and a result is (possibly) returned. Forms on the client side and CGI scripts on the server side frequently act in tandem as a means by which the user's input can do meaningful things, such as allowing user feedback or processing credit card information Stop by http://cougarxp.princeton.edu:2112 /bpd/webweather.html
to see an interesting interface to weather reports for regions across the country.
This only scratches the surface of what can be done. For a list of links to what are, in the words of NCSA, "exemplary demonstrations" of Mosaic and the Web select "Demo" from Mosaic's "Help" menu.
Information Overload
On the Web, unlike other media, unsolicited information is also unobtrusive. There is plenty of "junk mail" out there, but you do not have to sift through it in the traditional sense: Unless you get sidetracked (which seems to be as much a feature of the Web as it is a drawback), you won't spend time reading pages of no interest to you.
On the other hand, it is not always easy (and sometimes it's not even possible) to reach the pages you do want to read, i.e. to find information that answers a real question or addresses a real need. Though "browsing" the Web by hand can lead to interesting answers to questions that you never asked, it is not often a good approach for the ones you do ask. Search engines to date aren't fast enough or exhaustive enough to put the entire Internet at your fingertips, and even after the inevitable leap in technology and improved organization schemes handle these concerns, there isn't enough content out there yet.. And it's not clear whether the truly useful content of the future will be free.
As for Mosaic itself, it is perhaps the best interface imaginable to what's out there, and as a Web client it is unmatched in its power and ease of use Its success can be measured by its pervasiveness: it has reached such a level of ubiquity that people instinctively associate it with more generic constructs such as HTTP servers (frequently called "Mosaic servers") and HTML documents ("Mosaic pages"), and it has received applause in periodicals ranging from Wired to The Wall Street Journal.
[[Sigma]] If you're interested in exploring novel uses for Mosaic and CGI scripts, see the SigUnix update on page 5 and stop by a meeting.
[[Sigma]] If you're interested in discussing issues such as security, privacy, and the availability of free information on the Web, check out the SigCAS update on page 3 and stop by a meeting.