At the end of my Wired Windows column in this week's issue of Network World, I promised I would explore deeper the concepts of SMBmeta, so here we are.
SMBmeta, or Small to Midsized Business Metadata, was launched by Interland, "...to try to help small and midsized businesses provide information to search engines and directories in an open and distributed way so that customers can do more complete searching."
SMBmeta is the brainchild of Dan Bricklin who helped foster the PC revolution in the late 1970s and early 1980s. He is the co-creator of VisiCalc as well as other tools and applications that bridged the gap between the pure potential of personal computing in 1980 to the worldwide phenomena it is today.
Steve Lohr wrote a fascinating biographical piece on Bricklin for "The New York Times" a couple of weeks ago that grabbed the attention of the Burton Group's Jamie Lewis. It reminded Lewis that there were aspects of SMBmeta that he wanted to explore (see links below) which in turn led both Network World Fusion Executive Editor Adam Gaffin and myself to consider the implications and possibilities (see "Distributed Web directories" link below).
In many ways SMBmeta is a rational, easy to use, UDDI-for-the-rest-of-us kind of solution to the online problem of locating people, companies and services.
Lewis calls it "the self-organizing, loosely coupled directory" but I'd like to think of it going all the way towards becoming the universal, self-publishing, loosely-coupled personal directory.
The original idea is simplicity itself: companies wishing to participate would create a file, smbmeta.xml, and place it in the root directory of their Web server. So information about company ABC, which owns the domain ABC.com could be fetched as http://abc.com/smbmeta.xml (of course, FTP or any other file transfer protocol could be used).
The file would consist of a series of XML tagged data defining the company, its business and its location. Both standardized attributes such as the business type taken from the same list used by the U.S. Census, the North American Industry Classification System, as well as free-form descriptions are allowed.
The full specification and explanation can be viewed at http://www.trellixtech.com/smbmetaspec09proposal.html but the typical small business doesn't have anyone on staff who's generally handy with XML (and I count myself in that group) so there's also an application that guides you, step-by-step, through creating the file (http://www.smbmeta.org/tools.html).
Sounds similar to Universal Description, Discovery, and Integration (UDDI) so far, doesn't it?
There are some similarities, but SMBmeta lacks the centralized repository that UDDI requires, and which some think is why UDDI has been slow to be adopted for Internet use. Lewis, in fact, feels that UDDI will be very successful within the enterprise to be used as a way to locate applications and services. I think that's a possibility in a decentralized enterprise but a traditional, hierarchical structure will use standard directory technology and Service Locator Protocol because they're easier to implement and maintain.
But meanwhile, back on the SMBmeta ranch...
Bricklin is proposing more than just the creation of smbmeta.xml files, though. Without ways to use them, they amount to only a small collection of data. So Bricklin is proposing the SMBmeta Ecosystem. We'll explore that in the next issue.
Learn more about this topicDistributed Web directories
Network World Fusion, 05/08/03