Mosaic and the Web: the more sophisticated Internet

This article was prompted by John Koper and several other IEEE members who contacted me after the appearance of the first gopher article (The Institute, January 1994) and said - great article but have you tried Mosaic...

In this article, I take an introductory look at some of the newer developments that are descending upon us like wildfire. Some key words are: World Wide Web, hypertext, hypermedia, Hgopher, and Mosaic. What does this all mean? Should we get excited? Are there significant opportunities?

As we entered the '90s, e-mail seemed a big step forward in communicating and accessing information. It was mainly text and point to point. As we head for the middle of this decade, our information retrieval potential is exploding. The sheer volume of electronically stored information requires the ability to search efficiently. Mere simply-formatted text is being replaced by a mix of elegantly presented text, graphics, and sound. This material can be combined in virtually real time using resources available around the globe. New tools are being developed to handle this explosion of information. The web and mosaic are two of those tools.

SERVERS AND CLIENTS.   Let's start with the difference between servers and clients. A server is a package of organized information on a computer system. The client is the program you use to connect to the server. There are gopher servers and web servers. The web server is more complex than the gopher server. The client you need to connect to a web server has to have more features than one that you use to connect to a gopher server.

We will first discuss the web server as an extension of a gopher server. Later we will discuss clients that can connect you with gopher and web servers.

GOPHER LIMITATIONS.   Each gopher server contains a so-called root directory which is the top level menu of information. You connect to the gopher at this point. The root directory contains the names of the next-level directories. This process continues for any number of levels. Text files and other forms of information and resources are listed in the various directories. To go from one item of information to another, you may have to backtrack through this tree structure.

There are several difficulties. One is the problem of naming and organizing the directories meaningfully. Another is the reality that many items fit equally well in many directories. Also, as the size (and complexity) of the gopher increases, the number of steps to find a particular item increases. We can solve the latter problem, if we need to repeatedly trace the same path, by using "bookmarks". You add a bookmark beside an item (by hitting the "a" key). The next time you use gopher, you call the bookmark screen (by hitting the "v" key), and select one of your bookmarks.

The bookmark idea is particularly useful if the second item is on a different server from the first. Most gopher servers interconnect to other gopher servers for much of the information they list. This avoids duplication of information and simplifies the process of updating. In computer jargon - one gopher is "hooked" to another gopher. The transfer from one server to another is often transparent to you even though the two servers may be located on different continents (unless you look closely at the clues on the screen) but may introduce a time delay.

WEB SERVERS.   World Wide Web (or WWW or W cubed or simply the Web) is an extension of the gopher concept. In text files that you read on a web server, certain words are highlighted. If you select a highlighted word, you are transferred to another text file which is likely in a different directory and possibly on a different server (computer). This implies that the difference between text files and directories is blurred. Most important - duplicate entries are eliminated.

A document that contains text with highlighted words and hidden imbedded commands that cause the transfer to another document (file) is referred to as a hypertext document. We use a hypertext editor to produce the document. The next step is hypermedia. The hypermedia document contains, besides highlighted words, special symbols. These symbols, if selected, cause either the display of graphics files or the playing of sound tracks.

The word "hypertext" is attributed to Ted Nelson, the founder of the Xanadu Project, in the mid-'60s. The Web was started in 1991 at the European Center for Nuclear Physics (CERN) in Geneva, Switzerland.

Web servers can exist on their own or be imbedded in gopher servers. You connect to a web server by connecting to what is called a home page. At present, many so-called WWW home pages are in very early stages of development and some hypertext or hypermedia features are extremely limited or may not work. Also, there are often long delays while graphics or sound files are transferred.

WHAT YOU NEED.   The use of hypermedia features requires more system resources than those needed for gopher, and incredibly more than for e-mail. However, the potential is exciting and the availability of such tools will likely drive the development of higher capacity networks and more sophisticated software to make use of these tools. Which brings us to the question - now that we know what the Web is, how do we access it?

You need full Internet access and the right kind of hardware and software to explore or browse the World Wide Web. To access these features in the Web you need - for example - a PC with color monitor and sound card running Mosaic software under Windows, all connected via a high speed local area network to the Internet. NCSA recommends that your PC be a 486/33MHz with 8 Mb RAM.

I know that the "for example" is a bit of a copout, but I have learned that there will always be more variations in software and hardware than one person can ever keep track of. What I am about to discuss is my own experience. If some of you have other, very different experiences which you think our readers would find useful, please send me an e-mail message with the details.

CLIENT SOFTWARE.   You need a client to access a server. I am presently using three clients; Gopher, Hgopher, and Mosaic. The Gopher client is a program that runs on the host computer where I have my e-mail account. This client is a program that runs under the Unix operating system on our Silicon Graphics (SGI) workstation. Our system manager obtained the software from anonymous ftp at "boom-box.micro.umn.edu" - that's the University of Minnesota, birthplace of Gopher. I log on in the usual manner and type "gopher" at the Unix prompt. The default convention is that I connect to our on-site gopher. If I type "gopher gopher.ieee.org", I connect to the main IEEE gopher server. In my view, this is your basic gopher access package. You can access it remotely using a modem and dial-up line. It's simple and it works.

I have also installed (this is a euphemism for saying that I asked our system manager - Bruce Chiarot - to do this) an Hgopher client on my PC. This software was available from our computer systems support group at my company (a university). The client connects directly to the on-site gopher server by default. I had to get permission to have my PC authorized for off-site access since the gopher server connects to the Internet. I can not dial up from outside to use Hgopher. Hgopher runs in Windows, so I click on the Hgopher icon and see the root directory of our local gopher. This display is in color with several visually pleasing enhancements, but not operationally very different from my basic Gopher client.

MOSAIC.   We (meaning Bruce again) recently installed Mosaic, which allowed me to sample the intrigues of the web. Installation was not simple - even for Bruce. Since my Mosaic runs under Windows on my PC, I needed off-site access permission as in the preceding paragraph. Again, I can not use an external dial-up unless I am running an Internet access protocol such as SLIP. SLIP (Serial Line Internet Protocol) enables you to use a phone line and modems to connect your PC to the Internet instead of connecting as a terminal to another computer (workstation) that is directly connected to the Internet. We obtained the mosaic software from ftp.ncsa.uiuc.edu. The National Center for Supercomputing Applications at the University of Illinois in Urbana-Champaign is the developer of Mosaic. We found that we had to use the NCSA exclusive TCP drivers instead of the ones we were using (Beame and Whiteside). This means that I have to reboot with changed autoexec.bat and config.sys files whenever I want to use Mosaic.

We discovered that Microsoft does not provide the appropriate software to play the ".au" sound files that are sent from the web servers we tried. Then we discovered that NCSA has a zipped file "wplany.zip" which contains the exe file that we installed under Windows. We also had to carefully edit the mosaic.ini file and put it in the Windows home directory. The result was worth the effort in that we could view pleasingly presented hypertext with graphics and sound. The initial mosaic screen has Windows-based menus that allow the user to select from a variety of gopher or web servers.

Mosaic clients are available for Unix and VMS platforms, and MACs and PCs. There are also other browsers - client software developed to aid the browsing of web and gopher style servers.

Resources such as the web server and clients like Mosaic provide us with the opportunity to work cooperatively in any part of the world where the Internet exists, and organize access to information. Given the global range of the IEEE, our network of technical specialists, and our linkages between industry, education, and government, what an opportunity there is for IEEE members to work together, to share advances, and stay ahead in an economic environment that can be so damaging to those who fall behind!