Change happens

"The dotcom market was slipping," recalls Steve Stinton, E-botz.com's CTO. "The investors wanted to see something now, and we had nothing to give them." The business plan Stinton had developed with CEO Lonnie Wills and COO Tim Shanahan called for E-botz.com to provide a health-care supplies comparison-shopping service. The Vancouver, Wash.-based venture aims to help clinics, doctors' offices and similar facilities find the best deals on rubber gloves, needles and other widely used medical products. Great idea. But like zillions of other dotcom startups, Stinton and his partners had put the cart in front of the horse, polishing their business plan before creating a functional technology. Now, facing increasingly skeptical investors, they had to build the service's basic architecture -- fast.

After considering and rejecting several e-commerce-oriented development tools, the team settled on NQL Solutions' Network Query Language, a scripting language designed to simplify the creation of intelligent agents, bots and Web applications. "NQL is a solid company, and its technology is easy to use and superior to the competition, particularly its pattern matching and recognition tools," says Stinton. But the language was also very new -- still at the beta-test stage -- which meant that E-botz.com found itself risking its future on a product that had no commercial track record. Yet Stinton felt it was more important to use a technology that provided the capabilities E-botz.com required rather than settle for a less capable but tried-and-tested tool. "Sometimes you have to cross your fingers and hope for the best," he says.

As new programming languages and development tools roll across the IT landscape, more than a few CIOs are tossing dice and praying to whatever idols are conveniently available. These CIOs are finding themselves in a pinch, says Larry Perlstein, research director of Gartner Group, a technology research company based in Stamford, Conn. "In order to stay competitive they're forced to use tools that they and their programmers don't yet fully trust or understand."

CIOs tend to have a love-hate relationship with emerging programming technologies, says Vito Legrottaglie, former vice president of IT systems and operations for Programmer's Paradise, a Shrewsbury, N.J.-based software reseller that specializes in languages and programming tools. "On the one hand, new languages and tools enable programmers to develop applications faster than before. On the other hand, they also require significant investments in research, training and other resources."

For E-botz.com, the gamble with NQL is paying off. After some four months of planning the service's design and scope, E-botz used the product to develop a working model of a product-pricing service in a mere two-and-a-half weeks. Then E-botz was able to present that model to both current and potential investors. "NQL helped by providing the technology and by reducing our learning curve to the point where we can bring a developer on board and have that person writing pretty decent stuff in a couple of hours," says Stinton. After some moree refining and a period of beta testing with clinics, the E-botz.com service is scheduled to begin operation by January 2001.

Yet, while things are working out well for E-botz.com, not every encounter with a new programming technology is trouble-free. Jim Ruggiero, chief technology officer at Novo, a San Francisco-based e-services company, recalls his organization's recent experience with a less-than-perfect development tool. Late last year, Novo selected San Mateo, Calif.-based Blue Martini Software's e-tailing architecture Customer Interaction System to help it create an e-commerce website for San Francisco-based cosmetics retailer Gloss.com. According to Ruggiero, the toolkit wasn't quite ready for prime time, although he had entered into the venture fully realizing that Novo would be the technology's first user. He attributes the problems he encountered to the increasing complexity of development tools. "The latest products are extremely useful and powerful. Yet they are so complicated and have so many different layers of technology that it's virtually impossible to get things right without calling on the vendor for help."

As it turned out, Ruggiero and his staff needed a lot of help. "In some ways we were the guinea pigs," he says. "There were lots of things that didn't work the way they were supposed to or didn't work at all." Fortunately, Blue Martini proved to be very cooperative, even to the extent of bringing experts to Novo's site so that they could work shoulder to shoulder with the company's developers. "They were fixing the bugs as we were finding them," recalls Ruggiero, who feels the experience left him wiser and generally satisfied. "At day's end, the technology worked very well, and we had the benefit of nonstop vendor support. Not a bad deal."

Whenever a CIO considers a new language or programming tool, vendor support, stability and viability are critical concerns. "Vendors often go under or change direction," says Gartner's Perlstein. "With a new development environment it's very possible that only a single vendor is supplying and supporting the tools, so one has to look for all of the available options and choose wisely."

Yet simply keeping up-to-date on new and potentially useful programming technologies can be a major challenge. At Hewlett-Packard Co.'s OpenView business unit, in Fort Collins, Colo., technology brainstorming is a team effort. Programmers, researchers and other IT experts are invited to participate in brown-bag meetings where they discuss and debate emerging programming languages and tools over lunch. "People who are fired up about a new technology like to share their thoughts," says HP OpenView Product Engineering Operations Manager Dwight Schletter.

Once a new technology has been selected, retraining current programmers or finding qualified new developers looms as a major challenge. While XML and many other frameworks are relatively easy to understand, more complex technologies, such as Java, require a much longer learning curve. "It can take anywhere from six to 18 months before an individual becomes proficient in Java," Perlstein says. Yet hiring programmers already skilled in a particular technology -- potentially a quick fix -- can also be expensive. "A new technology, combined with high demand, equals a shallow talent pool and costly, hard-to-find workers," he adds.

Even so, the latest, greatest tool -- and its promised potential -- can be pretty appealing. But Legrottaglie urgees CIOs not to get carried away by enthusiasm for a new technology. "It's quite possible that once you've analyzed your situation, you'll realize that you'll be able to get the same results by using a less risky methodology." But if using a new technology is inevitable, the best thing a CIO can do is to thoroughly research the field and brace yourself for a rollicking, yet educational, roller-coaster ride, says Legrottaglie. "As they say in the gym," he quips, "no pain, no gain."