Personal computers and the world software market.
Personal computers are also helping to shrink the world. Since they are affordable and appropriate to the needs of people in both industrial and nonindustrial countries, personal computers have spread throughout the world faster than mainframes. For example, in surveying Indian computing, software-engineering authority Ed Yourdon tells us, "The history of the Indian computer industry can be neatly divided into two phases: a stagnant, IBM-less, pre-PC period, and a new, thriving era that began with the introduction of the PC and Gandhi's push for modernization of the information technology sector" .
Computer-mediated communication networks are also helping to shrink the world. As John Quarterman says, we are developing a worldwide matrix of connected computer networks and conferencing systems . The matrix is being built by the academic community, corporations, and vendors of commercial services. The matrix depends on personal computers as well as communication technology. We access the matrix with computers, not dumb terminals.
The worldwide spread of personal computers and the growth of the matrix may have dramatic effects on the software industry. On a recent trip to Chile I had a firsthand look at personal computer proliferation and the Latin American software industry. I will begin by describing what I saw in Chile and then discuss software export activities in other nations.
I went to Chile to present a paper at the Tenth International Computer Science Conference, hosted by the University of Chile. The technical program was varied, and researchers from Latin America, North America, Europe and Asia presented papers. The academic conference was interesting and of high quality, but it was not the most significant event of the week.
The Computer Science Conference was at a large convention facility, and it was held at the same time as Softel '90, a major Latin American trade show. I was reminded of the old national computer conferences, which also featured scientific programs in conjunction with commercial exhibits.
Softel is like a scaled-down Comdex conference: Comdex fills the Las Vegas Convention Center and about half a dozen major hotel convention areas, while Softel was spread over nine exhibit rooms with a total of 9,000 square meters. Softel booths are smaller than those at Comdex. Comdex has hundreds of talks and seminars, Softel has 30 sessions. Comdex has thousands of exhibitors, Softel 300. Comdex has over 100,000 attendees, Softel around 30,000. The Comdex exhibits take three or four days to see, Softel takes less than a day.
As with Comdex, Softel is predominantly a personal computer show. Registration is handled by clerks using networked Macintoshes. The vast majority of the exhibitors show personal computers, personal computer software, and supplies. Many of the personal computing companies that exhibit at Comdex are at Softel-Apple, Microsoft, Epson and Novell to mention a few, along with their Latin American distributors. There were even half a dozen Spanish language personal computing magazines like PC Tips, Computer News International (bilingual), BITS, Microbyte and PC Magazine (a Spanish version of the United States publication). The magazines look like thinner versions of their U.S. counterparts. About the only large systems I recall seeing were from communication companies (about 20% of the exhibitors were from the telecommunication industry).
Although there are fewer exhibits and the magazines are thinner, much of Softel's for-sale technology is the same as at Comdex. Apple features multimedia and desktop publishing applications in Santiago just as it does in Las Vegas. There are Novell networks, 486-based servers, and the same sorts of peripherals and accessories you see in the U.S. Softel was held just two months after the introduction of Microsoft Windows, but Windows was running in many booths. More significantly, software houses were working on Windows applications and development tools. The same was true for Unix TM. Many vendors sell Unix machines, and software companies are working with Unix windowing environments and C++ . (There were software houses from ten countries at Softel.)
There is still some lag in the movement of personal computer technology to South America, but it is much less than in 1982, when I was last in Chile. That visit was nearly a year after IBM had announced the PC, and it was still six months away from introduction in South America. In 1982, a handful of small companies were selling 8-bit systems, which were rapidly disappearing in the U.S., and the major vendors I visited-IBM, DEC and NCR--were happily selling minicomputers, oblivious to the personal computer revolution. In less than 10 years the time gap has closed somewhat from an initial year or two to a month or two.
Of course, you see things at Comdex that you do not see at Softel. Softel shows items that are for sale today. At Comdex you also see demonstrations of tomorrow's technology. The conference sessions also have a different emphasis. Comdex sessions lean toward marketing and market projection, while Softel sessions are oriented toward practical applications of communication and computing technology. There were sessions on applications in business, agriculture, construction, and textile production.
Two other interesting events were taking place at the same time and conference facility as Softel and the Computer Science Conference. One was the Thirteenth Annual Systems Engineering Workshop organized by the University of Chile.
This actually consisted of about a dozen simultaneous workshops on practical topics such as pollution reduction and the forestry industry. The workshops ran for the entire week, bringing together engineers and researchers from around the world.
The fourth, and most interesting, event of the week was the first Meeting of Iberia-American Software Exporters (abbreviated EIES in Spanish). As are their counterparts in many other nations, Iberia-American software companies are eyeing the world software market.
EIES was attended by 126 people, representing 53 companies in 9 nations. It was sponsored by the Chilean Software Association and the Committee of Software Exporting Businesses. The meeting was chaired by Pablo Palma, president of ARS Innovandi, a Chilean software house.
In his opening address, Palma predicted explosive growth in the world software market during the coming decade.  He stated that today only 2% of the world's software is produced in Latin America and Africa, but he feels the expanding world market offers a significant opportunity for software companies in those areas. Palma believes that in order to compete with global companies, Iberia-American firms will have to cooperate; EIES was convened to inaugurate that cooperation.
This four-day meeting featured speakers from Latin America, Spain and North America. I only heard the opening-day talks, which included speakers from the United States, Spain, and Chile. They included a variety of topics--from a government economist speaking about the strength of the Chilean commitment, to free trade to a somewhat technical presentation on OSI. There were sessions on communication infrastructure, practical experience with software export, and even a presentation on the United States software market which was lifted from the annual survey published in Jeffery Tartar's newsletter, So Letter.
I thought the most inspirational talk was given by Dan Mapes, a United States marketing consultant. Mapes foresees the growth and spread of the software industry as contributing to a more just, equally developed world. He reasons that since innovative software is developed by small groups with little capital equipment, companies from nonindustrial countries will do well in the world market. He is optimistic, believing that different cultures, each having a unique worldview, will invent and develop different sorts of software. The world community will be enriched by a flow of software and ideas from Latin America and other regions. In a more practical vein, he feels that companies with software products that are successful in local markets can export them to the U.S. and other countries with the help of marketing and distribution partners.
I interviewed Palma after EIES to learn what conclusions had been reached and found that an association of national software associations has been formed. As of today, Spain, South American countries (with the exception of Columbia and Venezuela), and several Central American countries are members of the association. By next year it should include all of Central and South America, Spain and Portugal. While the participants agree that thinking of a worldwide market is inevitable, during the coming year, they will concentrate on fostering trade within Iberia-America. They will produce a database of products and companies, and meet again next year. I will keep Communications readers posted on next year's developments.
Adios American Programmer?
Howell, et al. estimate that the U.S. accounts for about 52% of world software consumption and U.S. software producers hold roughly 70% of the world market. Another study, by International Data Corporation estimates that in 1989 the U.S., Europe and Japan accounted for 81 % of the world software market . While such figures are not precise, there is no denying that today's software industry is concentrated in North America, Europe and Japan.
It is difficult for me to believe the software industry will remain so concentrated. The world communication infrastructure is improving daily; salaries and overhead are very high in Europe, North America and parts of Asia; and powerful personal computers have drastically lowered capital costs for software companies. Take Chile as an example. Chilean universities are excellent, and many of the people I met have advanced degrees from the U.S.; we communicate via MCI Mail and The Internet; their salaries, office space and other ancillary services are much cheaper than in the U.S.; and they use the same commercial hardware and software tools as we do.
Ed Yourdon puts it more dramatically. He believes that international competition will put U.S. programmers out of work, that "the American programmer is about to go the way of the dinosaur and the dodo bird. During the next five to seven years, I foresee massive unemployment among the ranks of American programmers, systems analysts and others in the data processing industry [6l."
If Yourdon is correct, we had better start looking for second careers. Even if he is only partially correct, we will be affected, so I decided to examine to what extent personal computer software companies are turning to off-shore programmers.
For a start, I learned that Ashton-Tate, Claris, Lotus, Microsoft, Retix and other software companies have offices in Ireland. Each company localizes its software in Ireland. For example, according to Stuart Kazin, vice president for manufacturing, Lotus produces versions of its programs in eight European languages, and translation and testing of the software and manuals are performed in Dublin. (Software translation and conversion for the Asian market is handled out of Japan.) Claris's Dan McClarren told me they do the same, overlapping it with U.S. beta testing, so software is released at the same time throughout the world. These companies also manufacture European software and support and distribute it from Ireland.
According to co-founder john Stevenson, Retix, a vendor of OSI networking software, has gone further than the others, moving significant development to Ireland. It began by recruiting programmers from Irish universities and bringing them to the U.S. Since that went well, Retix opened a 15-person office in Ireland three years ago. Doing so put the company close to European customers who needed support and software customization. It eliminated the eight-hour time difference and made travel cheaper and easier. The office has grown to 60 people, with roughly 25% working on support and 75% on development. (The entire company consists of 400 people.) The Irish office is growing rapidly, and it might eventually have as many developers as in the United States.
For now, most of Retix's product-marketing people are in the U.S.; consequently the decision to develop a product is typically made here. If it is decided that development be done in Ireland, the requirements documents, functional specification, writing, testing and maintenance would all be done by an Irish-managed and -staffed team. A product is developed in whichever office is the center for expertise on the relevant OSI protocols.
I asked each of these companies if there were any major drawbacks in working with Irish offices, and all said "no." They eliminate potential cultural clashes by employing Irish managers, find the employees well-educated and industrious, like being near European markets, and save money on salaries, overhead and taxes.
This satisfaction is no accident--it was planned. The Irish government noticed that they had a brain-drain problem and a 17% unemployment rate, so they made a decision to become a software leader. Computer science programs in universities were enhanced, a 10% corporate tax rate was established, $2.5 billion was invested in telecommunication, and the Industrial Development Authority of Ireland opened offices around the world. Their chief asset is well-trained, hard-working employees, who were praised by everyone I interviewed.
India's effort to export software has been well publicized, and soon after I began researching it, I discovered that Ed Yourdon had been to India to study the software industry. Yourdon reported on his experience and impressions in his newsletter, The American Programmer. Let me tell you a little of what he found
Since 1985, when Rajiv Gandhi began a push for modernization and the personal computer was introduced, the Indian information technology industry has grown at an annual rate of 59%, compared to 12% in the U.S. Current Indian software exports are approximately $10-15 million; the government hopes to expand software exports to $100 million (some predict $500 million) by 1995. The primary reason for believing they can succeed is that the typical programmer earns $2,400 per year, and overhead is low. While Yourdon feels India bears watching, there are many obstacles to overcome. Since the information technology industry languished until 1985, it is weighted toward small and personal computers. There is little experience with large systems or projects; there are few sophisticated end users (an important prerequisite to a sophisticated software industry); and the domestic software industry is only.25% of GNP, compared to between 3 and 5% in Western industrial nations. Yourdon believes that to have a viable software export industry, India must also develop a strong domestic industry. Finding qualified people is also a problem. The projected need is for 275,000 programmers by 1995, but the country has only 50,000 today. (The U.S. has around 1 million.) Yourdon also found the telecommunication network woefully inadequate.
There are several ways that software can be exported. Companies can station programmers at a client's site for the duration of a project, write turnkey systems from a specification developed in consultation with a client, or develop packages for worldwide marketing. In 1989, 80% of Indian software exports were from on-site programming, turnkey projects accounted for 15%, and software packages only 1% (the remainder was data capture).
While on-site programming is somewhat profitable, the return is low compared to those writing turnkey programs and developing packages. In addition, many of the on-site programmers find ways to join India's significant brain drain. For these reasons, the industry hopes to move away from on-site programming, and a recent report from Forrester Research  indicates that the time may be right. In a survey of 50 Fortune 1000 companies, Forrester found that 38% have decentralized application development and 22% are evaluating doing so. Some of this decentralization will leave the country. As consultant Herb Halbrecht told me, "The role of a CIO is becoming like that of an international investment broker. Companies scour the world looking for pockets of excellence."
Yourdon agrees that India should move away from on-site programming, and suggests a five-stage strategy:
1) Build a reputation by providing inexpensive on-site programming.
2) Shift the services back to India with well-specified projects.
3) Shift emphasis from low-cost to high quality.
4) Shift from service to product orientation by finding niches or producing cheap clones.
5) Find applications that encapsulate
India's unique expertise. This transition is illustrated by the plans of database software vendor Ashton-Tate. According to vice president David Proctor, Ashton-Tate has ten programmers coming from India. It is their plan to use them on projects and get to know them. In the long run, two or three will remain here as liaison with programmers living in India. Proctor estimates that the programmers in India will cost Ashton-Tate $1012,000 per year, including salaries and overhead. Yourdon concludes with a discussion of social differences between the U.S. and India. For example, in India there is a pervasive attitude that labor is cheap and machines expensive and there is a strongly rooted tradition of dominant public-sector companies in all key industries. In spite of cultural barriers and other problems, $12,000 per year for a university-trained programmer proficient in C and Unix cannot be ignored.
I am writing this on German Unification Day. This milestone and other recent changes in the former Eastern Bloc nations make one curious about the prospects of software export from Eastern Europe. Industry watcher Esther Dyson has traveled extensively in Eastern Europe during the last year, so I turned to her newsletter "Release 1.0" for an overview .
Dyson writes that the Eastern European countries are heterogeneous in many ways, including their cultures, economies, copyright laws, technical sophistication, and installed computer bases. Together, they have an installed base of a little over one million DOS personal computers, compared to 40 million in the U.S., but "PCs have been first seeping and now flooding in through less official channels." (A student of mine called last week to tell me that his firm had received an order for 29,000 PC clones from the Soviet Union.) Still, the local market for software is relatively small, and a robust, sophisticated local market will be necessary for software companies to compete internationally.
Development of a strong local market is hampered by a lack of hard currency and, more subtly, by a value system that undervalues intangible services. The idea of making a profit by merely working as a programmer or software distributor is foreign to many. Dyson also states that custom software is more prevalent than are packages in Eastern Europe. This is due to the small market and social factors such as a lack of copyright laws and enforcement, and little emphasis on productivity enhancement, due to salaries being low and computers being expensive.
While I did not come across examples of Eastern Europeans doing custom development in the U.S., a few companies are beginning to sell software packages. The most visible Eastern European software import is the game Tetris [TM]. Developed by Alexey Pazhitnov of the Soviet Academy of the Sciences, Tetris won four Software Publishing Association awards in 1988, including best entertainment program. There have been some disputes over royalties, but the Soviets seem to have learned to do business in the West, and Pazhitnov has become a minor celebrity. Tetris initially made its way to the West through Hungary, and its success is reminiscent of the Hungarian Rubik's Cube. Yourdon and Mapes both point to software that exploits unique world view--sperhaps we will get engrossing games from Eastern Europe.
PC software companies Autodesk and Ashton-Tate have also been active in Eastern Europe. They have paid their dues by accepting soft local currency, which they used to finance local offices. Both established training centers and dealer networks several years ago, and they are now beginning to receive hard currency payback. These relationships are also beginning to yield software for export. Autodesk plans to market three AutoCAD-related software tools which were developed by their Soviet partners, and Ashton-Tate will be marketing database development tools, beginning with a localization aid.
Graphisoft, a Hungarian firm, has taken an even more independent approach. They have opened their own offices in Canada, the U.S., and Western Europe, and market their internally developed CAD software for the Macintosh. The six-year-old company now has 30 programmers, so it appears to be viable.
The Far East
Remove the cover of your personal computer and you will see how well Far Eastern companies have been doing in hardware. The names printed on the chips read like a United Nations rollcall, and most (but not all) are from the Far East. Furthermore, there is a good chance your personal computer was assembled in the Far East. Still, for the time being, your Far Eastern hardware is probably running American software.
According to Carl Chang and Mikio Aoyama, software has been one of the fastest growing industries in most Far Eastern countries during the last few years. In surveying the software in Japan, Taiwan, South Korea, and Singapore , they observe a high level of activity in every country. They also note that while developments in each country are unique, there is one common thread--that "the effort to industrialize software is likely to be made through a national government-directed, and publicly funded initiative."
One company, Visible Systems, has complained that this subsidy and coordination is unfair. Visible Systems markets a CASE tool in competition with Picture-Oriented Software Engineering (POSE), a program developed in Singapore. Visible Systems filed a claim with the Department of Commerce asserting that POSE had been developed with a $15 million subsidy from the Singapore National Computer Board. The Department of Commerce initially decided in their favor, imposing a 15.25% punitive duty on POSE, but they subsequently reversed the decision, citing fear of retaliation . The Department of Commerce still holds that software is merchandise--not intellectual property--and that they therefore have the right to charge a duty.
While some U.S. companies, like Visible Systems, are asking for protection from Far Eastern competitors, IB M has tried joining its competitors. Nearly three years ago, IBM, in a strategic alliance with Taiwan's government-supported Institute for Information Industry, formed International Integrated Systems, Inc. (IISI), to develop software exclusively for IBM. Although IISI was required to compete openly with other software companies, it has delivered over 100 programs to IBM. The projects were divided fairly evenly between translation of IBM programs and documentation to Oriental languages, double-byte character set conversion, and new development. Once specifications were established, all of the work was done in Taiwan.
Having learned from this experience, IBM exercised an option to make an equity investment in IISI, and it now hopes to market software services to other customers. A retired IBM executive, Rusty Scheuer, who is heading the U.S. marketing effort, has found it is not an easy sale. Many potential customers are U.S. companies selling packages that they would like to convert for marketing in the Orient. Even at Taiwanese rates, however, the cost of conversion is steep, and there is also the reluctance of a U.S. company to give its source code to someone so far removed, in a country not known for its copyright enforcement.
This survey has cited examples of early efforts to export software to the U.S. It remains to be seen whether Yourdon's prediction of the demise of the American programmer will come true, but the American programmer should surely take offshore competition seriously. The quality of today's offshore programmers is high, and their salaries and overhead are low. On the other hand, while cultural and language barriers are diminishing, they still exist. Software, particularly custom, in-house applications, requires communication with and sensitivity to the users.
In the long run, education may determine software success. As long as a country like Chile or India is educating only a few elite programmers, these programmers will be very good; however, it remains to be seen if those countries (and we) can build the infrastructure to educate masses of average programmers needed by a growing software industry. Personal computers also affect education by cutting the cost of hardware to support teaching and research.
In any event, I hope our reaction to the possibility of foreign competition is not to seek protection, but to seek excellence. While offshore competition may be bad news for American programmers, it is good news for the American software consumers, and they are more numerous than we are.
Larry Press welcomes questions and comments from readers. His address, phone number and email address are:
10726 Esther Avenue
Los Angeles, CA 90064
For overviews of developments in Western Europe, see the April 1990 issue of Communications, and the June 1990 issue of IEEE Spectrum. The two complement each other, providing insight into the likely impact of the 1992 economic unification on the computer, communication and electronics industries.
For coverage of software development in the Far East, see the March 1989 issue of IEEE Software. It has articles on Japan, China, South Korea, Singapore, and Taiwan.
To learn what is happening in Eastern Europe, turn to Esther Dyson's newsletter Release 1.0. Dyson writes about far more than Eastern Europe, and wins my award as the most interesting industry watcher. She is an articulate writer with an understanding of computer science and the world of business. EDventure Holdings, Inc., 375 Park Avenue, New York, NY 10152, (212) 758-3434.
The American Programmer is a newsletter on software engineering and management published by industry veteran Ed Yourdon. In addition to presenting the latest in CASE tools, Yourdon travels extensively, and reports on his findings. He presents relevant cultural information along with facts and analysis, and writes with humor. (The copyright notice in the newsletter prohibits "eating the material.") Contact American Programmer, 161 West 86th Street, New York, NY 10024-3411; (212) 769-9460; fax: (212) 769-9458 or MCI: eyourdon or Compuserve 71250,2322.
ADAPSO, the Computer Software and Services Industry Association, publishes reports and holds seminars on international software developments. For example, their well-researched report on Japanese software was referenced above. Contact ADAPSO, 1300 North 17th Street, Suite 300, Arlington, VA 22209; (703) 522-5055.
If you are interested in exhibiting at or visiting Softel 91, contact Feria Internacional de Santiago, Tenderini 187-Casilla 40 D, Santiago, Chile; (562) 557-7096; fax: (562) 557-6923.
For information on the organization of Iberia-American national software associations, contact jose Fabio Marinho de Araujo, Director, IBPI, Rua Lauro Muller, 116/ Grupo 1207, 22290, Rio de Janeiro, Brazil; (021) 275-6594; Fax (021) 295-6993.
The Industrial Development Agency of Ireland has offices in 17 cities in North America, Europe and the Far East. For more information, contact Emmanuel Dowdall at 1821 Wilshire Boulevard, Suite 317, Santa Monica, CA 90403; (213) 829-0081; Fax: (213) 829-1586.
If you are not yet a user of the worldwide Matrix of computer-networks, you might take a look at either of the following books. If you are a user of the Matrix, you may enjoy them even more.
Frey, Donnalyn and Adams, Rick, "!%@:: A Directory of Electronic Mail Addressing and Networks," O'Reilly and Associates, Sebastopol, CA, 1990.
Quarterman, John S., "The Matrix: Computer Networks and Conferencing Systems Worldwide," Digital Press, Bedford, MA, 1990.13
1 Mr. Palma's optimism is widely shared. For example, the International Data Corporation estimates that the worldwide software market will grow from 36.7 billion in 1989 to 80.7 billion in 1994.
1. Chang, Carl K, and Aoyama, Mikio. Software in the Far East. IEEE Softw. (Mar. 1989), 11-12.
2. Davis, Ludlum A. Commerce department reverses Singapore ruling. Computerworld (Apr. 2, 1990), 1 1 9.
3. Dyson, E. Eastern infrastructure. Release 1.0. (Aug. 21, 1990), 1-34.
4. Howell, Thomas R., Noellert, William A., Ohri, Bonnie, and Wolff, Alan W. Japanese Software, the Next Competitive Challenge. ADAPSO, Arlington, Va., jan. 1989.
5. Quarterman, John S. The Matrix: Computer Networks and Conferencing Systems Worldwide. Digital Press, Bedford, Mass., 1990.
6. Yourdon, Ed. The decline and fall of the American programmer. Am. Prog. (Mar. 1988), 1-8.
7. Yourdon, Ed. India. Am. Prog. (Oct., 1989), 3-26.
8. Woodring, Stuart and Colony, George F. How software will be managed. Forrester Software Strategy Report, june, 1990.
9. Worldwide Information Technology Spending Patterns, 1989-1994. An analysis of opportunities in 30 countries. International Data Corporation, Framingham, Mass. Sept. 1990).
[TM] Tetris is a copyrighted trademark of Academy Soft-ELORG. It is also sublicensed to Sphere Inc. and Mirrorsoft Limited.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Personal Computing|
|Publication:||Communications of the ACM|
|Date:||Feb 1, 1991|
|Previous Article:||ACM forum.|
|Next Article:||Quick, where do the computers go?|