From the time the National Science Foundation (NSF) assumed responsibility for the U. S. Internet backbone, they anticipated a transition to commercial use. There were a few commercial ventures in the 1980s, like the Clarinet News Service, CARL UnCover for scholarly documents, and the Computists' Communique electronic newsletter, but the NSF acceptable use policy and Internet culture were largely non-commercial. NSF is phasing out their support, and commercialization is taking off -- you can even order pizza [8]{footnote 1}.
Total U. S. retail sales were $1.5 trillion in 1993. Of that, $53 billion was catalog sales, $2.5 billion TV shopping, and $200 million on-line shopping on the Internet, Compuserve, and other services [12]. Delivered pizza is a $6.4 billion business. Today, on-line shopping is a drop in the bucket, but Forrester Research predicts it will grow to $4.8 billion by 1998. Many facctors point to increased on-line shopping:
Networks are expanding rapidly. More women are coming on-line. Hardware and software advances permit improved user interfaces and functionality. On-line presence is cheaper than CD-ROM catalogs, which is cheaper than print catalogs. On-line information is current, and interaction with customer representatives and immediate ordering are possible. Price comparison, comparative evaluation, in-depth information, and analysis tools promise more efficient markets. People are purchasing powerful home computers for games and education. Connectivity cost will fall as the market grows, and new options like cable, ISDN, and wireless become available.
If on-line shopping becomes substantial, we will need new distribution methods and channels. Many of today's warehouses, retail establishments, and last-mile delivery by housewives and other consumers, will not be needed. Information products like video, music, software, and news may be selected and delivered on-line with no retail or warehousing. Market-making businesses, like real-estate listings, travel agencies, and security and commodity brokerage may disappear. Staple items may be purchased on-line, and delivered via services like United Parcel or new local or manufacturer-owned delivery companies. Industrial and durable goods, where in-depth, comparative information and computer-based analysis tools will be most important, will also be examined and screened on-line. We may still go to the mall/mercado for fresh vegetables, movies, lunch, boutique clothing, and other items, but it will not be today's mall. The employment and resource allocation changes generated by on-line shopping during the coming century might be comparable to the shifts out of agriculture during this century. {footnote 2}.
Of course, we could have on-line shopping without the Internet. In earlier articles [14, 16], we compared the interactive TV and Internet communities, asking among other things, which might be best for interactive merchandising. The Internet has an early lead. There are many differences between the Internet and interactive TV or Compuserve and other on-line services, but decentralization is its key advantage. {footnote 3} The Internet grows from the edges, like a Tinkertoy model, free of centralized control and capital formation. An entrepreneur with a spare bedroom can connect a computer to the Internet, and open for business. Two-year market studies and mergers or stock issues are unnecessary. Similarly, software tool providers can develop products, and introduce them to the Internet community with little cost.
Internet commerce requires software. Merchants need server software and customers client software. At present, there are three major client-server packages for publishing on the Internet, WAIS, Gopher, and WWW.
WAIS, Wide Area Information Server [21], is focused on ASCII text documents. The server accepts queries in a standard protocol, and does full-text boolean search with relevance feedback. WAIS software is in the public domain, and Brewster Kahle, the inventor of WAIS, has also formed a commercial company to enhance the software and develop specific applications. For example, a multimedia version of Encyclopedia Britannica should be available on the Internet using WAIS by the time this article is in print.
Gopher servers [20] also publish ASCII text, but have a menu- oriented interface and limited boolean search capability. Gopher's most important feature is that menus can point to documents on any server, enabling you to transparently move from one server to another. A menu choice on a North American server may switch you to an Asian server with no change in speed or user interface. Therefore, a server may reference related information without duplication of effort, editorial control, or storage space.
World Wide Web (WWW) servers [1] can also point to other servers, and they are more general. While Gopher is restricted to hierarchical, textual menus, the WWW interface is a hyperdocument, a "home page." The home page can contain formatted text or images, and any part of it can be a hot link to another document on the home server or another server. Another generalization is WWW's ability to store formatted text, video, images, sound and other data types. WWW clients can also retrieve information from Gopher and WAIS servers and from file archives.
WWW began at CERN, the European Particle Physics Laboratory, as a tool for the physics community, but it has grown beyond that constituency. Today, WWW traffic is the fastest growing part of the Internet [1]. There was a time when an Internet address on your business card was avante garde -- today, leading-edge business cards have addresses of WWW home pages. The WWW project is now a shared by CERN and MIT, where the standard will continue to evolve, in a manner similar to MIT's X-Windows. Commercial WWW software is also in the works, and several packages will be out by the time you read this.
With public domain software widely available, and commercial products in the wings, setting up a WWW server will become progressively easier. The difficulty will be in being seen -- in distinguishing your home page from millions of others. One way to do that will be locating your home page at a popular place in WWW space. Meckler Publishing hopes their Meckler Web [7] will become such a spot.
Meckler Web is one model for Internet advertising and commerce. Their home page will be organized into subject matter domains like art, technology, education, law, medicine, and senior- citizens. One might think of these domains as similar to specialized magazines. Domains will be moderated by full-time experts, and Meckler sees professional societies (like ACM) as likely sources of unbiased moderators. Meckler is selling its service to Fortune 1000 clients with commercial interests in the various domains. The client will pay $25,000 annually, but there will be no charge to view Meckler Web. As with broadcast radio or television, free access seems to be an evolving custom on the commercial Internet. Pay-per-view may come later.
Clients are entitled to 25 Mbytes on Meckler's server. As they gain experience, and wish to exceed the storage limit, they will move their information to in-house servers. At that point, the client will be paying for only pointers to their server.
Does $25,000 per year sound like a lot to pay for pointers to your server? It does if only a few people visit Meckler Web, but if millions visit daily, $25,000 is a pittance. It is a pointer to your electronic brochure (e-brochure). An e-brochure can be multimedia, and as large and content-rich as you choose to make it. E-brochures will also be interactive. Using today's WWW technology, you can return an order blank or other questionnaire, or send email to people listed at appropriate points. Future WWW modifications will almost certainly include the capability of initiating a real-time audio or video conversation with a sales or service person.{footnote 4} With e-brochures, the distinction between content and advertising in a traditional newspaper or magazine may blur, and compared to a Super-Bowl minute or magazine ad, a Meckler Web pointer may turn out to be a terrific bargain.
Transaction processing would be an obvious extension to Meckler Web. Someone reading about a service or product could place an order, thereby authorizing payment and shipment (with Meckler getting a small commission). For now, purchases will be handled outside the system, as with traditional advertising. Meckler plans to add transaction processing, but they decided not to do so initially because of security risks. An unencrypted file of credit card numbers would be a tempting target for a hacker as would transactions routed to Meckler's server.
Several organizations are working on secure WWW software, including CommerceNet [6], another Internet commerce initiative. CommerceNet is a regional consortium in California's Silicon Valley, and they are adding value by developing software for secure transaction processing and WWW publishing aids. Subscriptions start at $150 per month plus a $1,250 annual fee, which includes a 128 kbps ISDN connection. Like Meckler Web, CommerceNet will maintain a WWW page with member directories and links to their WWW servers. Subscribers will create and manage their own servers using CommerceNet tutorial material, training, and software tools.
They hope for regional development through industrial collaboration. For example, a company might post an encrypted RFQ for some development work, and obtain bids from selected local consulting firms which receive the encryption key. Negotiations would be conducted using encrypted conferencing or email, and eventually a contract with an authenticated signature would be signed.
Since encryption and authentication are not part of today's WWW software, they are developing security extensions. They are also working on WYSIWYG tools for creating WWW pages. One piece of the transaction-processing pie that still seems to be missing is some form of digital cash, but others are working on this service, for example, [2, 13].
CommerceNet is a non-profit consortium with State and Federal funding. This funding might be a source of concern to small firms which provide connectivity or a WWW presence without public funding. As of this writing, at least 55 firms are offering WWW space and assistance [11]. The cost of entering the WWW-presence business is low -- an entrepreneur can start at home. It is reminiscent of the booming computer BBS business, and may grow at the same rate. Small providers could also spread through franchising.
As enhanced server software and good e-brochure design tools become available, many businesses will create their WWW presence without assistance. E-brochure design skills will also become widespread. For example, one could imagine WWW home pages for shopping malls. The mall home page would contain general information and pointers to store home pages, which would be created and maintained by store personnel.
Large firms will almost certainly have home pages. For example, DEC's home page [9] contains product specifications, order blanks, public domain software, new product announcements, and other material. Their customers will have no trouble finding it because a WWW client is bundled with DEC computers. DEC is also a member of Meckler Web, so they will get "drop in" traffic as well.
Product-specific home pages are another strategy. For example, O'Reilly Associates' Global Network Navigator (GNN) [5] was voted "best commercial home page" at the first International WWW Conference in May, 1994. While it covers many topics, there is an extensive travel center. If the GNN travel center becomes {ital} the {ital} place to visit when planning a trip, hotels, airlines, and tour companies will want pointers to their WWW home pages there.
In the long run, software agents may find information for us, eliminating the need for intermediaries like Meckler, but that will take a long time. Meanwhile, the Internet is delightfully and efficiently out of control and unpredictable, so I would expect to see a variety of paths to a presence on the WWW before the dust settles.
With a timesharing account, you are limited to using Lynx, a character-oriented WWW client from the University of Kansas. Lynx may be installed on your host machine, so try typing the {ital}lynx{ital} command after logging in. If your host has Lynx, it will execute, starting you out at a default home page (probably at the University of Kansas). If your system does not have Lynx, you can use a public client by {ital}telnetting{ital} to ukanaix.cc.ukans.edu, and logging in as WWW. Either way, the user interface is self-explanatory, and there is on-line help.
To take full advantage of the WWW data types, your PC or workstation must be directly connected to the Internet. If your computer at work or your university is not already directly connected, you will want dial-in connectivity using the SLIP (serial-line IP) or PPP (point-to-point) protocols. Commercial SLIP/PPP accounts typically cost more than timesharing accounts, but I expect their prices to drop as the market grows and technology improves. You will also need SLIP/PPP software for your PC. Such software is available commercially and in the public domain, and it will be included as part of the next version of MicroSoft Windows. See [4] for a discussion of sources of and experience with this software for the Macintosh or PC compatibles or [10] for Windows only.
If a university or employer provides your Internet access, ask about dial-in SLIP or PPP. If it is not available, perhaps you can interest them in dial-in access. You can reference the home- IP program at UC Berkeley. Berkeley students, faculty, and staff can request a home-IP account, which remains valid during their entire stay at the University. They are also given the appropriate software for their home computers.
In addition to SLIP/PPP connectivity software, you will need clients for WWW, Gopher, WAIS, and other Internet services. Let's look at WWW. Many companies are working on WWW client software. None are on the market today{footnote 5} (they will be when you read this), but there are good public domain options, see [4, 10].
The two public domain clients are Mosaic and Cello. Mosaic is available for Mac, Windows, and unix. Cello is less well known, and is only available for Windows. That is the bad news. The good news (for Windows users) is that Cello developer Tom Bruce plans enhancements which fully exploit Windows' capabilities -- he promises the integration of WWW browsing with other applications, and simplified WWW publishing.
Regardless of which client you choose, you will need viewers for each data type. Public domain viewers for images, sound, video, Postscript, and other data types are available on the Internet [4, 10]. If you have a slow modem, retrieving multimedia documents is slow, but you will get the idea, and in the future, faster communication will reduce this problem.
Obtaining and installing communication software, clients and viewers takes time and effort. It must be located on the Internet, copied, and decompressed. Initialization files must be created to identify network addresses and host-connectivity options, specify preferences for clients, point clients to viewer software, and so forth. The documentation may be terse. When commercial packages become available, they will be easier to install and have differentiating features. However, if you have a bit of hobbyist in you, and would like to know what is going on under the surface, public domain software is for you.
Writing this article reminds me of the early days of personal computing. For example, the first useful operating system was called CP/M. Installing CP/M was similar to installing public domain Internet software. Digital Research, the purveyors of CP/M, sent you a disk and manual in a baggie, and installation meant writing keyboard, screen, and disk drivers to interface with your hardware configuration. The manual had examples, and the drivers were simple (character-in, character-out, and so forth), but, like installing Internet software, writing them was non-trivial.
The parallels between the early days of personal computing and today's Internet go beyond software installation to questions of culture and value. In the 1970s, personal computing was largely a hobbyist movement which valued craftsmanship, information exchange, and sharing of expertise. There were early commercial ventures, but they centered around hobbyist publications and clubs. Apple Computer co-founders Steve Jobs and Steve Wosniak gave away Apple I schematics and sold vector-board machines on a card table at Homebrew Computer Club meetings in Palo Alto, CA. After the meetings, presidents of fledgling companies drank beer together at a local hamburger spot, and bragged about clever board designs.
The Internet is more sophisticated and complex than early personal computers, but it was also formed in an atmosphere of craftsmanship and information exchange, which persists today. Leading-edge software is distributed on the Internet, and valuable databases, document libraries, and indices are also published at no cost.
Perhaps more important, the Internet culture supports open communication. People answer questions, make suggestions, and freely discuss a myriad of topics for the satisfaction of participation and perhaps some enhancement to their reputation -- the payoffs are not explicit [3, 15]. This barter/gift-exchange arrangement makes for a more comfortable society than one in which every information transaction is explicitly compensated, and no accounting is needed. This open culture is subject to abuse, but it has persisted for years on the Internet. Will increased commercialization end openness? Must it? Can we find policies that balance openness and marketplace efficiency? Social predictions are difficult at best, and the global nature of the Internet makes them even more difficult.
I understand that volunteerism and information exchange are difficult to sustain, and commercial enterprises are very effective. The capitalization and blazing growth of the personal computer industry would have been impossible by any other means. But, something has also been lost in the Microsofting of personal computing.
Internet Connectivity Options -- Timeshared Host vs. Direct
Connectivity
Timeshared Host
+ can use a terminal or (much better and almost universal in the
US today) a low-cost personal computer emulating a terminal
+ works satisfactorily with a low-speed connection {footnote 6}
- limited to a command-line, menu-based, or character hypertext
user interface
- limited to text and numeric data types
- adds the complexity of a second computer between the user's
computer and the Internet, requiring a beginner to build a more
complex conceptual model of the system, and to learn commands
for the editors, application programs, communication programs,
and operating systems of both computers
Direct Connectivity
+ allows a graphical user interface
+ simplifies the user's view of the system since information is
transferred directly to and from the Internet to his or her
workstation
- requires a higher speed connection -- at least 14.4 bps modem
and a fast serial interface.
- requires a faster personal computer with more memory and disk
space
Direct connectivity requires more hardware, but is more flexible
and easier to use. Direct connectivity is more attractive as
hardware evolves. A character-oriented timesharing account was
appropriate for a PC with a 286 CPU running DOS, but when a
typical PC is a 486 running Windows, it is less so. A new
program, The Internet Adapter [18], promises to provide some of
the advantages of direct connectivity with a timesharing account,
but I have not yet tried it.
1. Schrage [17] suggests that Pizza Hut extend their service to
allow for interactive, drag-and-drop custom pizza design. He
points out that purchases of boxed chocolates, t-shirts and
other clothes, flowers, and many others items could be
customized by the user.
2. The impact could be even more radical in developing nations,
where many city dwellers do very small retail work.
3. AT&T will also be an important player. They are working on
many communication services including shopping using General
Magic's agent technology, interconnection of Novell LANs, and
on-line servers for Lotus Notes.
4. Cornell University's CU-SeeMe is a creative prototype of such
ad-hoc video conferencing [19]. With CU-SeeMe an Internet
users with low-cost video cameras participate in conferences
with one or a few others. Speeds are just a few frames per
second, but this is a low-cost harbinger of interesting
things to come.
5. Netcom, a major connectivity provider, has WWW browser called
NetCruiser, but it is bundled with their accounts, and tied
to proprietary server software. The next release will be
compatible with standard WWW servers.
6. My first Internet account was on a timeshared host using a
Teletype bundled with an acoustic coupler, giving me 100 bps.
Even that low speed was satisfactory -- in fact it was
amazing.
Two excellent monthly consulting newsletters centered on issues raised in this column are: The Internet Letter which is devoted to corporate Internet use, internetworking, and information services. It is a source of timely Internet news and analysis and practical information. Net Week, Inc., (301) 229-6693, netweek@access.digex.net. The Forrester Report on People and Technology analyzes the impact of technology on consumers. This is the newest of Forrester's "Strategy Service" reports. These monthly reports are based on interviews with Fortune 1000 companies and comprehensive analysis. Forrester Research, Inc., (617) 497- 7090. Other newsletters with a broader focus, including outstanding coverage of Internet commerce are: Release 1.0, monthly, (212) 924-8800, 5113705@mcimail.com. Computists's Communique, weekly (electronic), laws@ai.sri.com. Computer Letter, 40 issues/year, (212) 696-9330, 3712412@mcimail.com There are many sources of information to help you get get WWW and other Internet client software running on your personal computer, including: Jayne Levin (publisher of the Internet Letter) has written a booklet [10] showing how to find and install WAIS, Gopher, and WWW clients for Windows-based PCs. (301) 229-6693, netweek@access.digex.net. Frank Hecker maintains an evolving document on SLIP/PPP and client software for Macs and PCs on the Internet [4]. It contains pointers to many other resources. There are also a number of book/disk combinations to get you started, including: "Internet Starter Kit for Macintosh" by Adam Engst and "Internet Starter Kit for Windows" by Adam Engst, Corwin Low, and Michael Simon, $29.95, Hayden Books. "The Windows Internet Tour Guide" and "The PC Internet Tour Guide" both by Michael Fraase, $24.95, Ventana Press. "The Mosaic Handbook for Microsoft Windows," by Dale Dougherty and Richard Koman and "The Mosaic Handbook for the X Window System," by Dale Dougherty and Richard Koman, and Paula Feruson, O'Reilly & Associates. You will also need a fast modem (14.4 kbps or above), and if you have an old computer with slow serial ports that cannot keep up, you need a serial port accelerator such as Hayes' ESP. I have tested the ESP, and it works well in a 16-Mhz, 386-SX PC. Hayes Microcomputer Products, (404) 840-9200. For an excellent example of a combined shopping and information service on a commercial network, see the Scientific American area on America On Line (AOL). AOL has a simple point-and-click interface designed for low-speed modems, and offers free 10-hour test accounts. (800)-827-6364, Ext. 3853.
1. Berners-Lee, Tim , Cailliau, Robert, Luotonen, Ari, Frystyk,
Nielsen, and Secret, Arthur, "The World-Wide Web,"
Communications of the ACM, vol. 37, no. 7, pp 76-82, August,
1994.
2. DigiCash, ecash-beta@digicash.support.nl.
3. Dyson, Esther, "An Explicit Look at Explicitness," Release
1.0, October 31, 1993, EDventure Holdings, New York.
4. Hecker, Frank, "Personal Internet Access using SLIP or PPP,"
ftp from ftp.digex.net, pub/access/hecker/internet/slip-
ppp.txt.
5. http://nearnet.gnn.com/
6. http://www.pizzahut.com/
7. http://www.service.digital.com/home.html
8. http://www.mecklerweb.com/demo.html
9. http://www.commerce.net/
10. Levin, Jayne, "Internet Software Guide for the PC," Netweek,
Inc., Washington, DC 20045, netweek@access.digex.net.
11. Morris, Mary E. S., "List of WWW Service Providers,"
marym@finesse.com.
12. Modhal, Mary A., Colony, George F., and Chowdhury, Seema,
"Interactive Shopping," Forrester Report on People and
Technology, Vol 1, #1, May, 1994.
13. NetBank, send the message "netbank-faq" to info@agents.com.
14. Press, L., "Compuvision or Teleputer?" Communications of the
Association for Computing Machinery, September, 1990.
15. Press, L., "Systems for Finding People," Journal of
Organizational Computing, 2(3&4), 303-314 (1992).
16. Press, L., "Interactive TV and the Internet," Communications
of the ACM, December, 1993.
17. Schrage, Michael, "On-Line Pizza Idea is Clever but Only
Half-Baked," Los Angeles Times, August 25, 1994, pp d1.
18. --, "About the Internet Adapter," ftp from marketplace.com:
/tia/docs/tia.info.txt.
19. --, "CU-SeeMe Frequently Asked Questions," ftp
gated.cornell.edu: /pub/video/CU-SeeMe.FAQ.7-6.txt
20. --, "Gopher Frequently Asked Questions," ftp from
rtfm.mit.edu: /pub/usenet/news.answers/gopher-faq.
21. --, "WAIS Frequently Asked Questions," ftp from rtfm.mit.edu:
/pub/usenet/news.answers/wais-faq/getting-started.