Not fade away
Before you buy the line that peer-to-peer and Web services are passing fads, look again at ways they're converging
Follow @infoworldON WALL STREET, the canary in the mine shaft for a turnaround is capitulation. It's the tipping point between despair and salvation, when enough people believe the worst is yet to come to fuel the upturn they've given up on.
Fundamental to capitulation is the leading negative indicator. Lee Gomes of the Wall Street Journal is one of my favorites in the technology sector. "It's starting to look as if the end may be near for the 'peer-to-peer' fad," Gomes pronounced in April 2001. The influential journalist's comments sent many p-to-p players scampering to the hills, downplaying the peer elements of their strategies.
Now, in a recent column, Gomes has his eye on Web services, which he defines as "when companies use the Internet to do business with each other." And then out comes the hammer: "But that was the great promise of the b-to-b fad of three years ago, and Web services seem like just a reheated helping of the same dish."
Sure enough, everybody is stepping up to the plate like Barry Bonds looking for a pitch to hit out. Typical is Intel CTO Pat Gelsinger, who told me at Agenda that we may look up a few years from now and realize that the dynamic discovery vision of Web services was just a fantasy. As John Patrick reported (See " Off the record "), "Pat put forth the idea that Web services is fundamentally flawed because it optimizes to the network as opposed to direct computing resources, e.g., microprocessors."
The only thing better would be for Bill Gates himself to sit down in a one-to-one interview and soulfully admit, "I'm superashamed to say it, but Web services just didn't pan out the way I thought it would." But that won't happen -- just ask Marc Lucovsky, if you can find him.
Ironically, Lucovsky's Hailstorm technology was the inspiration for Groove Network's forthcoming edge services initiative (now called Groove Web services). "I hadn't really heard of SOAP yet," Groove's John Burkhardt recalls. "Then we heard about Hailstorm and Web services. They've got these messaging services -- MyNotifications, MyMessages -- and isn't the [Groove] relay server sort of a message storing service, and couldn't it be a MyRelay? A Hailstorm MyMessageQueue server."
Two years later, Hailstorm is gone, absorbed into the .Net Framework and servers. But Groove is being transformed by edge services. "[Microsoft] is trying to figure out how to build massively scalable applications that they're going to run on these megaservers that can handle millions of clients hammering on this thing," Burkhardt says. With Groove, he says, "now you have your own little, personal SOAP server."
Meanwhile, a few weeks ago at the Gartner conference, I was treated to a demo of Microsoft's XDocs, a dynamic forms generator that sits atop an XML Web services platform. The product, slated for release in mid-2003, is billed as an Office family member, yet it can operate with or without Office and/or .Net.
It's hard not to drop into market-speak about XDocs. How's this: XDocs forms are easy to develop and easier to use. Sorry, but it's true. Domino designer developers are already familiar with its dynamic rows, scriptable visibility, date pickers, rich text fields, and validation tools. You can wire up all types of data inputs: embedded lookup tables, external databases, and -- that's right, ma, I'm only bleeding -- Web services.









