The Next Big Thing

Will Web services do anything for consumer software?

by Barry Parr

The Internet has convinced a lot of people that every five years or so there will be a Next Big Thing. But what happens when that Next Big Thing doesn't show up to save the industry? (For an example, take a look at the telecom business.)

Today, most vendors, businesses, and media outlets are betting on Web services to be that Next Big Thing for the software industry. While Web services may be a helpful solution for legacy integration, there's danger in believing that consumers will favor distributed applications, with components running on and communicating from the servers at major corporations.

The concept has the potential to be a lot more like WAP—a once-hyped technology that faded quickly. If you believed the talk when WAP and wireless computing were all the rage, you would have thought that we'd be checking our email, ordering books, and getting Starbucks coupons on our cell phones by now. Notice how many of the WAP proponents disappeared as soon as the economy started moving downhill and CTOs began focusing on technology that actually mattered. Today, the WAP discussion has been replaced by similarly grand theories about distributed applications, like word processors that have a thesaurus hosted on one server and a spellchecker on yet another.

Before rushing in to rewrite desktop applications, we really need to figure out how Web services will be different from previous failures to move computing off the desktop. Remember client/server models, network computers, and application service providers? In each case, the personal computer remained at the center of the desktop. The big difference with Web services is that, instead of being promoted only by the enemies of desktop computing (like IBM, Oracle, and Sun), this time, even Microsoft is pushing for the concept.

Who Really Needs It?

The push for consumer Web services is coming simply because we now have the technology to distribute applications, and most importantly because it serves the strategic needs of vendors. Neither case has anything to do with end-user demand. This is the classic formula for technology-driven failure, not market-driven success.

Vendors need Web services more than consumers do. Consider the current problem with the PC business. Weak sales are due not only to the recession, but also to the fact that the desktop computer has reached the flat part of its evolutionary curve. For the last five years, Intel has been struggling to find applications that can tax the ever more-powerful processors to maintain market share and margins. Dell, on the other hand, has admitted that higher gains these days will come from a more efficient internal process, rather than an innovative new product.

For Microsoft, Windows XP was the last chance to sell a must-have upgrade for at least another five years. As we're seeing, the plan hasn't worked and XP adoption is low. But Microsoft has another plan: the company wants desperately to sell subscriptions to its software.

What the folks in Redmond would really love to do is rent our customers' data back to us once they've gained control of it. That's the dream behind Passport and other .Net services. Once Microsoft or any other organization convinces customers to store all of their information at one location, other companies will have to pay Microsoft and upgrade their software to access that data.

What To Do

How should we as system builders respond to this? The way we always have: by focusing on solving problems, and not on implementing technology for the sake of the technology, or the demands of our vendors' strategies.

It sounds obvious, but we need to start with the problem and then seek out the best solution. That's the only way we'll wind up with the right technology. Unfortunately, the current lather over Web services suggests that many people are missing the obvious.

As I said earlier, there are clearly instances where Web services make sense. These are applications where there is a single source of authority, real-time information is needed, communication is a major component, and automation is both doable and desirable.

The hype is hot for Web services on the consumer side, but that's due to run its course this year. The industry still hasn't made a case for which end-user problems the new technology will solve. For that reason, any software vendor that plans to rewrite its applications to use Web services should first examine whether it wants to build a solution without a problem.

Barry is an Internet media consultant and the former director of e-commerce research at IDC. He was also co-creator of News.com and a key developer of the first online newspaper at the San Jose Mercury News. Contact him at [email protected].