Double Down

While the lights of the casino flash hypnotically and a barrage of electronic bleeps and cascading coins assaults your ears, you try to focus on the lightening fast blackjack dealer at your table. The two cards in your hand add up to eleven. In a split second, you must decide whether to double down -- a doubling of your bet in an attempt to win the hand with a single card. If the dealer slides you a winner, you’re a hero. If you come up short, don’t look to the dealer for a second card -- your doubled bet is already gone. Many enterprises have been making double-down bets as they try to Web-enable existing applications. Ignoring leverage and revenue for now, a major result of these Web bets is the doubling of software maintenance costs. We are beginning a product cycle where companies have to maintain multiple, distinct, and interrelated products. Keeping developers interested in maintenance, paying for double maintenance, or avoiding double maintenance will be the millennium’s first major technical challenge.

If your company’s only goal during the past few years was to get your technical products moved to the Web, you thankfully had many tools to assist you. From 3GLs to 4GLs, from HTML layout tools to XML data manipulation tools, from screen-scraping Internet terminals to sophisticated Java clients -- the software development industry has been rife with the means to Internet-enable your company. If you chose to bolt a Web front end onto a legacy product, rather than embracing a Web-oriented architecture, you may open your wallet twice to pay for this year’s software maintenance.

SAP AG, an ERP provider, is a good example of how Web-enabling a product can cut both ways. After creating an initial marketing stir with its campaign for mySAP.com, the company’s stock price hit a high of about $85. Once analysts determined mySAP.com was simply a bolt-on HTML interface to the old SAP architecture, SAP starting taking a pounding. That pounding included a 43 percent net income drop, a stock price drop to the mid $40s, and the departure of its American chief. Because mySAP.com isn’t a complete replacement for SAP’s existing applications, SAP developers will be pulling double duty -- supporting client-server and Web applications -- as they try to recover.

PeopleSoft, my former employer, created a number of Web strategies and products before landing on its current Internet architecture. PeopleSoft ran the gamut from basic HTML page creation -- too limited -- to a fat-client Java version of their Windows client -- too bandwidth-intensive. To its credit, the company didn’t stop thinking after the first few attempts. PeopleSoft's latest architecture is very thin, leverages much of its existing three-tier architecture, and requires only HTLM on the client. Most importantly, PeopleSoft has made it a goal to replace all of its existing Windows applications with an Internet equivalent. If it succeeds, the company will have sidestepped double maintenance.

In blackjack, you can double down on any hand. Of course, the clever gambler will typically double his bet only if he’s confident a single card will beat the dealer. In blackjack, that usually means you’re holding eleven. It’s important to look at the hand your company holds before deciding how to Web-enable your assets. If you are only holding a six -- a rickety old mainframe system that provides nothing but aggravation to its current users -- you probably shouldn’t double your bet that adding a Web interface will turn things around. It might be time, instead, for a Web-oriented redesign.

On the other end of the spectrum, if your product is graphically intensive or uses hierarchical tree navigation as a critical part of the interface, you may want to think carefully about how well your product will work on the Web. Graphics are getting better on the Web thanks in large part to Macromedia’s Flash and Shockwave products. Once Microsoft and Netscape agree on an XML and JavaScript standard within their respective browsers, creating hierarchical presentation without making tons of server roundtrips will become much easier. You don’t want to tarnish your reputation with a weak Web product, or have to maintain custom code once the rest of the world is getting native functionality for free.

Regardless of your decision, a few guidelines will help increase your odds of winning the Web bet without spending your winnings on maintenance. When you evaluate a vendor’s development product, go beyond the floorshow of developing an application from scratch and make sure you understand how their tool will ease your first revision. Lean toward XML tools -- they are more likely to capture your business’ rules and structure than HTML tools. Consider tools that target Windows and Web products from the same definition, substantially reducing maintenance costs. Finally, if your user interface was never a big selling point, consider changing to a browser-only interface, which removes double maintenance entirely. Once you make your decision, rub your rabbit’s foot, throw salt over your shoulder, and place your bet. --Eric Binary Anderson has led projects at a number if enterprise software companies and is currently the senior architect at IBT Financial, an Internet-based training company in Bend, Ore. Contact him at ebinary@columnist.com.