In his speech at BEAs eworld conference in San Francisco, open-source activist and publisher Tim OReilly described Microsofts defunct HailStorm project as a good idea from the wrong company.
HailStorms notion of a massive in-memory cloud of XML data and metadata was doomed, not by the daunting mechanics of schematizing a broad set of generic business processes or assembling a mesh network of server clusters, but by the fears of many that Microsoft would commandeer personal information to achieve network domination, as it once seized the desktop.
Today, those same concerns are focused on a new suspect, as Google threatens to move beyond its search bastion to Gmail e-mail and the emerging RSS (RDF Site Summary) platform. Basic to Googles rise is its Tom Sawyer-like success in harnessing others work to its benefit—corralling their requests for information to create an authority map based on page rank. The resulting rank becomes the coin of the realm, leveraged for intelligent search responses, targeted advertising and services designed by Google and its users through Web services API calls.
At the core of Googles dynamic is implicit metadata, made on the fly by users as they reveal their interests by browsing, messaging, filling out Web forms and creating documents. Separating the metadata from the underlying data it describes has let Google initiate a conversation in which users effectively trade general personal data for access to services derived from the aggregated requests. This scoped contract between users and the cloud sidesteps most privacy and political concerns. Where HailStorm required a Passport account to enter the network, Google merely requires a willingness to view additional information—advertising—that has been dynamically generated based on user metadata inputs to the system.
The implicit metadata contract between users and services is creating what open-source journalist Doc Searls calls “mutant” companies, where startups services mutate to exploit users wishes and desires.
In his eWorld keynote, BEAs chief architect, Adam Bosworth, cited the similar transformation around the GUI, which gave procedural control to users. Now RSS is creating another shift, away from the Web request model to user-controlled aggregation. TiVo-like user metadata can be harvested to offer services in return for access to group and trend data. And as RSS containers become more intelligent about applying authority filtering to feeds, the signal to noise improves. Bosworth showed eWEEK Senior Writer Darryl K. Taft and me an intelligent RSS router, built atop his Alchemy extended browser project, a framework that uses declarative metadata dynamic caching to create a rich conversation between a thin client and the server cloud. Sounds like HailStorm, I told Bosworth, who didnt disagree. But a HailStorm based on a framework BEA will open-source. Paging Tom Sawyer.
Contributing Editor Steve Gillmor is editor of eWEEK.coms Messaging and Collaboration Center. He can be reached at [email protected].