Back to the Future: Java Goes Mobile

Back to the Future

By D. Tweney
Sun first touted Java as a universal client-side platform—and even went so far as to develop brain-dead network computers (NC) that relied on Java for their operating system and on servers for their storage and smarts. But NCs flopped, Java applets on webpages are a dying breed, and client-side Java now seems all but dead, especially now that Microsoft has pulled Java support from Windows XP.

Yet Java is quietly undergoing a renaissance on the client—this time as a platform for applications embedded in cell phones, PDAs and other mobile devices. One reason is the announced intention of major cell phone manufacturers to start selling Java-enabled mobile phones. Gartner estimates that 40 percent of PDAs and 68 percent of mobile phones will be Java-enabled by 2006. The prospect of hundreds of millions of Java-enabled mobile devices has many application developers drooling, and by this past June’s JavaOne conference, more than 150,000 developers had already downloaded Sun’s toolkit for mobile Java—the Java 2 Platform Micro Edition.

In the enterprise, Java-enabled cell phones and PDAs present an opportunity to extend feature-rich enterprise applications to mobile workers, such as traveling sales staff, field service personnel and delivery people. That becomes even easier as enterprises move toward XML-based Web services architectures, which make it easy for developers to extend applications to a variety of client devices.

United Air Lines, for instance, is building a Java-based middleware architecture, with the aim of making it easier to deliver data through a variety of client channels—including, ultimately, wireless devices, says CIO Eric Dean in Chicago. “Web services will be a great way for embedded Java to communicate back with the server somewhere,” says Mark Carges, president of BEA Systems’ e-commerce application components division.

“Of course mobile and wireless environments are still very immature,” cautions Mark Driver, research director for Internet and mobile technologies at Stamford, Conn.-based Gartner. For now, few major corporations have actually deployed Java-based mobile applications; most are just testing the waters. But stay tuned: Once there are hundreds of millions of Java-enabled cell phones in the world, it’s only a matter of time before enterprise applications start reaching out to those devices.

For other resources at :
CIO Home

Link: Back to the Future: Java Goes Mobile

Link broken? Try the Wayback Machine.

Back to the Future: Java Goes Mobile

Strong Java

Strong Java

Despite Microsoft’s best efforts, Java is well-established in the enterprise. Can it hold its ground?BY D.F. TWENEY

THIS YEAR, the programming language-cum-development platform called Java turned 5. It now stands as one of the world’s most popular computer languages—and it continues to grow. The number of Java programmers is increasing by 10 percent per year, according to research company Evans Data.

Yet Java’s ascendancy hasn’t happened quite the way Sun envisioned back in 1996. In stark contrast to the swarm of Java applets populating the Web during its first years, client-side Java is almost nonexistent today. Instead, the language has moved behind the scenes, within the application servers that drive corporate websites—and increasingly, companies’ line-of-business applications.

During the past year, enterprises have taken Java to heart like never before. The language has matured. Tools for developing and deploying heavyweight Java applications are readily available from Borland, IBM and Sun. And developers now have a wealth of experience with the language.

“Java today has become mainstream,” says Mark Driver, research director for Internet and mobile technologies at Stamford, Conn.-based Gartner. As a result, Driver says, Java applications are turning up everywhere from mainframes to mobile phones—and thanks to improved Java development and management tools, companies don’t necessarily need Java gurus to benefit from the language anymore.

Seeking Stability
At Detroit-based Ford Financial—the financial services arm of Ford Motor Co.—Java is central to the company’s

Back to the Future: Java Goes Mobile
Java is quietly undergoing a renaissance on the client—this time as a platform for applications embedded in cell phones, PDAs and other mobile devices.

Read More

migration away from a two-tier client/server model toward a three-tier thin-client architecture. While maintaining the company’s longstanding big-iron back end (IBM mainframes running DB2 databases), the company is now developing Java-based middleware applications that run on BEA Systems’ WebLogic Java application server. Ford’s applications, which handle such core business tasks as loan origination and account management, now have HTML client interfaces, eliminating the need to support client-side software in the company’s eight service centers and 150 dealer locations—and making it possible to extend these applications to consumers on the Web.

Ford has no regrets about basing its IT infrastructure on what five years ago was a brand-new technology. “We selected [Java] because it met our scalability, flexibility and value needs; and it has really proven itself,” says Marcy Klevorn, director of customer branch and dealer systems for Ford Financial.

What’s more, Java now has only one serious competitor—Microsoft’s .Net framework—but that competitor is just getting out of the starting blocks. (See “.Net Gain,” July 1, 2001.)

Enterprise OS-1
What catalyzed Java’s corporate growth was the release in early 2000 of Java 2 Enterprise Edition (J2EE). Not a product, but a set of standards and procedures, J2EE formalized a framework for building multitier Java applications, using technologies such as servlets (Java applets that run on a server), Enterprise JavaBeans to exchange data and application objects, and Java server pages (JSP) to generate HTML for Web-based applications.

J2EE caught on quickly, with developers lured in part by a well-stocked toolkit. “Even with just the J2EE environment provided by Sun, I’ve got a ton of my application already built,” says Ted Shelton, senior vice president and chief strategy officer for Scotts Valley, Calif.-based Borland, which sells Java development tools as well as Java application servers.

The standards provided by J2EE provide, in effect, an operating system for enterprise applications, handling low-level programming issues such as data access, file management and interoperability among application components. “Java is great by itself, but once the operating system galvanized—and that is J2EE—that’s what really made it go,” says Mark Carges, president of BEA Systems’ e-commerce application components division.

Wide Industry Support
Once J2EE appeared, enterprise software vendors, with the exception of Microsoft, quickly lined up behind it. As a result, Java’s biggest asset now is the wide range of middleware based on J2EE—BEA Systems, Bluestone, Borland, IBM and Sun’s iPlanet all offer Java-based application servers.

For enterprises, standardizing on one platform and language reduces risk, because the single standard makes it easier to replace software if necessary. It also simplifies integration issues and lets the same Java experts work on a variety of projects.

At Ford Financial, a central team of 25 Java gurus works with application development teams in the company’s various departments, providing help by evaluating vendors, assisting with integration and implementation, maintaining the underlying J2EE infrastructure, and looking for opportunities to reuse components and code among departments. “You’re only going to have so many people in an organization that really have strong [object-oriented] development skills. We try to isolate some of the nuts and bolts from the application teams so they can concentrate on the business logic,” says Jeff Lemmer, manager of the e-commerce and application architecture team at Ford Financial.

In some cases, business issues, rather than the language’s technical merits—lead a company to Java. “The majority of enterprises are not choosing between Microsoft or Java—they’re choosing Microsoft .Net, IBM WebSphere or BEA Systems WebLogic,” says David Chappell, principal of San Francisco-based IT consultancy David Chappell and Associates. “In choosing an enterprise Java product, the most important thing is that the vendor who makes your product will still be selling it five years from now.”

Eric Dean, CIO of Elk Grove Township, Ill.-based UAL Corp. (the parent company of United Air Lines), makes a similar point, noting that WebLogic and its Java structure are convenient tools, “but there’s not a religion around Java.” Instead, the important feature is the middleware layer it provides, he says.

Power Productivity
In its earliest days, Java was touted as a “write once, run anywhere” language. It didn’t quite work that way—applications usually need some tweaking to run on different platforms. But what is truly transferable are people’s programming skills. Stu Stern, who heads Sun Java Center, the Palo Alto, Calif.-based Java arm of Sun’s professional services division, quips that it’s a “learn once, write anywhere” language.

In addition, IT managers appreciate the increased productivity of Java developers. Java provides a good balance between rapid development—thanks to its object-oriented nature and the wide availability of Java components—and the ability to access low-level computing processes. According to Richard Monson-Haefel, an author and programmer, productivity under Java is typically 20 percent to 40 percent higher than when using C or C++, thanks to built-in features such as automated memory management and the ease with which components can be reused. Ford’s Lemmer has seen application developers’ productivity increase as much as two to three times when they move from C++ to Java.

Java developers are still in high demand, says Gartner’s Driver, but their numbers are increasing fast. While Java Developer Connection estimates that there are currently about 2 million registered Java developers, Gartner puts the number of “qualified” developers at about half that. But the company predicts there will be nearly 3 million experienced Java developers by 2005—forming a rich talent pool for enterprise IT departments to draw from.

The Race Is On
Still, it’s too early to say that Java has won completely. The corridors of computer industry history are littered with the corpses of companies that underestimated Microsoft. Although .Net is still brand new, one thing all commentators agree on is that Microsoft will continue to improve it until it becomes a serious threat.

“Over the next five years, we see two de facto platforms for a vast majority of e-business apps: Java and Microsoft,” says Driver. “We don’t see a clear winner. We expect a 40-40 split, or perhaps a 50-30 split in favor of Java,” with the remaining 20 percent divided among a variety of legacy and other platforms.

For large companies with a wide variety of platforms and hardware in their data centers, Java will probably remain the platform of choice, thanks to these companies’ existing relationships with Unix vendors and to Java’s cross-platform strengths. For small and midsize companies, however, it’s much easier to standardize on a single platform—and that’s where Microsoft may enjoy an advantage, Driver says.

Regardless, the next five years are likely to see a lot of light and heat generated over the Java versus Microsoft issue. Take it all with a grain of salt. Both platforms are technically robust and will likely remain around for a long time, says Chappell. “Both are good enough. If that weren’t the case, one would be crushing the other one.” Whether or not that statement still holds in five years remains to be seen—but for now, Java is going strong.

D.F. Tweney ( is an award-winning writer and editor covering business technology and the Internet.


Link: Strong Java

Link broken? Try the Wayback Machine.

Strong Java

Common Language for the Next-Generation Internet

So-called Web-services applications will soon talk to each other without human intervention. But first everyone needs to agree on how they’ll communicate.

In my last column, I wrote about Web services, which are business applications that share data with other programs over the Internet. Instead of people initiating every communication, computers could have virtual conversations with one another, taking care of routine transactions quickly and efficiently (see “What’s Going On Down at the Plant?“).

For instance, when your production system detects that you’re low on a critical part, it could automatically post a purchase request on a private extranet site that can be accessed only by your suppliers. Those suppliers’ computers could, in turn, initiate a buy order and feed that request into their own manufacturing system so the replacement part gets made and is shipped to you — all automatically.

Right now that kind of integration is difficult to achieve. Applications vary widely in how they handle data, so techies need to link up each application with others on a case-by-case basis, carefully converting data formats one line of code at a time. It’s a tedious process, and companies such as Tibco, WebMethods, and Vitria have staked out a decent business doing this kind of legacy integration.

With Web services, by contrast, applications would make their data available to the world, along with instructions for how to access and control the application itself. If this “interface,” as the programmers call it, gets built in a consistent manner, then applications of all sorts can easily communicate with one another unaided, with no case-by-case integration needed. With Web services, for example, an airline could easily share data with online travel sites, describing not only flight information but also fare availability, seating choices, and even in-flight movie options.

“It’s like putting up a webpage,” says Barry Morris, CEO of Iona, a company that sells technology to enable Web services and application integration. “Suddenly the webpage is available to millions of people. The same thing goes for Web services, only instead of a person with a browser accessing it, you’re going to have other programs accessing it.” Naturally, companies will provide security restrictions on their Web services to limit entry, just as they do with intranets and extranets today.

The key to making this scenario work is instituting a set of widely accepted protocols outlining exactly how Web services process information and communicate with each other. Right now three main protocols look promising: SOAP, WSDL, and UDDI. This is how they work:

— SOAP, simple object access protocol, defines how applications talk to one another to exchange information and commands. Think of it as a vocabulary list for the language these computers speak.

— WSDL, Web services description language, lets applications describe what services they offer, like a virtual menu.

— UDDI, universal description, discovery, and integration standard, creates a worldwide directory where businesses and the applications they use can locate other Web services; it functions like a phone book for Web services.

All three protocols are based on extensible markup language (XML), which means they’re relatively easy for Web programmers to learn and use. But they are still relatively new and haven’t been officially recognized by any major standards organizations. Only one, SOAP, is currently under active consideration by the World Wide Web Consortium, the most influential standards-making body for Web technologies. (Microsoft is also a backer of the SOAP standard.) But the ratification process for SOAP could take years. In other words, all of these emerging standards are still subject to change — and to possible subversion by large companies with enough clout to create their own variant versions of the standards (remember the “browser wars” between Microsoft and Netscape?).

For that reason, most information technology departments are waiting until the protocols become more universal before going full-bore into Web services. It’s just too risky to bet the farm on a few emerging standards. But keep an eye on these three, because they’re the leading candidates right now. And once a common language does emerge, Web services will be a central part of the next big wave of corporate IT investment.

For more information on SOAP and the XML protocol (a related standard), see the W3C’s page.

Also useful:
SOAP, WSDL, UDDI: Digesting the Alphabet Soup (Information Week)

For details on WSDL, see the W3C’s note on the topic.

For more on UDDI, visit the homepage for the UDDI community.

Link: Common Language for the Next-Generation Internet

Link broken? Try the Wayback Machine.

Common Language for the Next-Generation Internet

Think Globally, Act Locally

By: Dylan Tweney
Issue: November 2001
Adding foreign language versions of your website can easily pay for itself in fresh leads and revenues.

Most websites take a Model T approach to overseas marketing: You can access them in any language you want, as long as it’s English. Eighty-five percent of the Web’s pages are in English, but only 45 percent of the Web’s users are native speakers of that language, according to research firm IDC, and the percentage is dropping year by year.

Many non-native speakers can read English, sure, but making potential customers speak your language — instead of soliciting them in theirs — is a backward way of doing business. English-only might suffice for routine transactions between established business partners (or for American tourists in Paris), but if you really want to sell something to someone, you’d better speak that person’s language.

At Otis Elevator, based in Farmington, Conn., monthly online sales leads shot up from 130 to more than 1,000 after the company launched local-language versions of its website — with 70 percent of those leads coming from new customers. Even simple gestures help: Travelocity (TVLY) vice president Ned Booth says that whenever his company adds country-specific travel services or information to its site (such as a local customer-service phone number), sales in that country typically double. As for Travelocity’s new online operation in Germany, Booth says, “We couldn’t market at all if we didn’t do it in German. We’ve really got to be able to communicate clearly and make the experience comfortable enough that they buy.”

Such successes are little surprise, but what inhibits many companies is the price of developing a multilingual presence online. It involves translating webpages (which can cost 25 cents per word), maintaining several different sites (one for each country or language), coordinating content and branding between the sites, building business systems that are capable of handling international e-commerce, and so forth. According to IT research company Aberdeen Group, the typical cost of producing a website in another language is $50,000 to $100,000, and large projects can run as much as $2 million per language.

It is possible to keep the costs under control. My advice: Think globally (when it comes to setting up site infrastructure and online branding) and act locally (when it comes to content, prices, fulfillment, and customer service). In other words, build a centralized Web infrastructure that’s capable of delivering websites in a variety of languages. That way you don’t have to reinvent the wheel for each new site, and you can make sure that your company’s core branding remains consistent. Then you can leave it to your employees in local markets to produce and oversee content suitable for their respective regions.

The key to making this work smoothly and trimming costs is globalization technology — specialized content management software that will let you run a multilingual and multinational website with relative ease. Using globalization software, webmasters and site designers at company headquarters create templates that control the overall appearance for a variety of sites in different languages. With the templates in place, content managers publish information on the site using simple webpage forms. Globalization systems also typically include work-flow tools that allow documents created in one language to be sent electronically to translators. When the translations are done, they are delivered to local content managers for approval and then publication — again, all using Web forms.

When Nintendo of America (NTDOY) created a “micro-site” to promote its new handheld gaming platform, the Gameboy Advance, it was with an eye toward internationalization. The site’s designers — Blast Radius, a Web design firm based in Vancouver, British Columbia — used a proprietary content management system that would be capable of handling multiple versions of the site down the road.

Blast Radius estimates that Nintendo’s international-minded approach added about 10 percent to the initial development cost but figures that the investment paid off when Nintendo of Europe decided to launch four localized versions of the Gameboy Advance site. Instead of re-creating the whole site, the company simply translated the content and deployed it on country-specific sites using the existing content management system. The European sites are now garnering about a million hits per week, says Nintendo of Europe online coordinator Marcus Brown, and were launched faster — in a few weeks — and at a much lower cost than would otherwise have been possible. Even so, the complexity of launching four websites at once was considerable, Brown says: “When you look at the site, it doesn’t appear to be that big, but to get it off the ground required an enormous amount of work and coordination.”

The leaders in selling globalization software are Idiom, GlobalSight, and Uniscape, but the overall market for their products is still small, and no vendor dominates. The software typically costs $80,000 to $500,000 per license, depending on how many versions of a site you want to create. Mainstream content management vendors like Vignette (VIGN) have also begun to add multinational and multilingual capabilities to their products. Service providers like Bowne Global Solutions, Berlitz GlobalNet, and Glides offer globalization services on an outsourced, or application service provider (ASP), model.

Ultimately, the biggest value of website globalization is that it can turn foreign browsers into buyers, says IDC analyst Rob Rosenthal. “[It] can have an enormous effect on top-line revenue,” he says. Add to that the burgeoning market of non-English speakers online — and the fact that fewer than 20 percent of websites today offer multiple languages — and you have a massive opportunity waiting to be tapped.

Link: Think Globally, Act Locally

Link broken? Try the Wayback Machine.

Think Globally, Act Locally