/var/opinion

They're Ba-ack

Nicholas Petreley

Issue #157, May 2007

The Network Computing revolution rears its beautiful head once again, thanks to Ajax.

What is Ajax, really? There's Ajax the technology, and then there's the fact that Ajax makes it possible to provide a rich client experience on a Web browser. Think about the latter for a moment. Wasn't that the basic idea behind Network Computing?

Does anyone recall when IBM and Oracle pushed the concept of Network Computers? Network Computing was all about delivering a rich client experience without the price tag of a PC and a commercial PC operating system. It also was about centralized storage and client management, which would bring the total cost of ownership way down.

The concept was so logical and compelling that it struck fear in the hearts of Microsoft and mainstream computer journalists. Microsoft had the most to lose. The Network Computing environment promoted by IBM and Oracle was, by necessity, platform-neutral. The goal was to base everything on browsers and Java, making the hardware platform and operating system irrelevant. You could participate in the revolution with a powerful Windows PC equipped with Java, but the fact that you ran Windows was incidental. It wasn't a necessary component.

The revolution self-destructed, however. Despite how sensible the concept may have been, there were two things wrong with it. First, hardware and Java weren't ready. Java applications were buggy, and most Network Computing appliances walked software, they didn't run it.

But here's what really killed the movement. IBM, Sun and Oracle discovered it was incredibly hard to make good money by selling truckloads of cheap computing appliances and only a handful of powerful servers. It's much more profitable to sell people massive computing power at the client as well as the server end, even if the average user never takes full advantage of the client machine's power.

The economic impact is also the reason why most computer journals hated Network Computers. The success of Network Computing would scale down the computing economy so much that advertising revenues would plummet. As a result, the mainstream computing press printed reams of anti-NC propaganda and hung on Microsoft's every word about the NetPC and Zero-Administration Windows. Remember those? Right, these reactive “initiatives” by Microsoft vanished the moment it became obvious that the Network Computing revolution wasn't going to get off the ground.

Wasn't going to get off the ground—yet. I used to go show-hopping with a presentation about Network Computing. I repeatedly predicted that Network Computing was so sensible you could count on the success of a Network Computing revolution, whether it happened that year or in decades. I also predicted that it would be based on Java, but I was careful to add that Java wasn't necessary. If Java flopped, some other platform-neutral technique of delivering applications and content would emerge in its place.

Hello Ajax. Ajax-based office suites are popping up everywhere, some free as in FOSS, some free as in service, some nonfree and some free with upgrade options. You can get a taste of the experience if you sign up at www.ajax13.com for free access to a suite of Ajax-based office applications. Or, you can try out Google's Docs and Spreadsheets at docs.google.com. Better still, you'll find out why I still prefer Java over Ajax by trying out the ThinkFree office suite beta at www.thinkfree.com. ThinkFree lets you choose between a lightweight and heavy-duty application. The lightweight applications are Ajax-based, and the heavy-duty applications are Java-based. Both types of applications are terrific, but the Java-based applications, such as its heavy-duty word processor, is much more slick and polished than the Ajax equivalent.

Here's why these efforts are much more likely to lead to a successful Network Computing revolution. They take advantage of the relative platform neutrality of browsers, but the success of these Web-based suites is not tied to any hardware platform. In other words, the NC revolution as pitched by IBM, Oracle and Sun expected you to buy a truck full of cheap clients. These Ajax and Java Web-based applications will work on a cheap client, but that's purely coincidental. This approach to the Network Computing revolution doesn't hinge upon changing what you buy, thus enabling hardware companies to keep selling you faster boxes with decent profit margins.

That's where Linux comes in. If this Network Computing revolution succeeds, OEMs will have one less reason to pay more to sell a Windows box than a Linux box. If people begin to depend on Web-based office applications, why pay Microsoft an OS tax on every unit when people can get the same experience with Linux and Firefox?

Many people will voice most of the same fears and objections as they did during the previous attempt to push Network Computing back in the late 1990s. If Web-based office suites pick up enough steam, you'll see these fears dissipate.

I'm still a bigger fan of Java than Ajax, and the fact that Java is going GPL may change the future of Web-based suites. But, even if we end up with Java, we probably will thank Ajax for getting it started. Regardless, I maintain that we will see a Network Computing revolution, whether it's today or decades from now. And, when it happens, sooner or later, it will be great for Linux.

Nicholas Petreley is Editor in Chief of Linux Journal and a former programmer, teacher, analyst and consultant who has been working with and writing about Linux for more than ten years.