I have logn felt that the work being done by Dan Bricklin on defining the SMBmeta spec is very useful. Some time ago, Dan released a sample directory and sample code.
A Burton Group article gives the wider context:
SMBmeta is a very simple set of XML tags that allow a business to describe itself. Using the tags, a Web site can provide a structured XML document that includes basic information like the business name, address, phone number, industry type, and so on. The structured XML document, which is very short and simple, lives a the root of the Web server, which makes it easy to find, parse, and process. Bricklin, who came up with the idea as a way to create value-added services for small businesses in his hosting business, thought that search engines could pick up the SMBmeta data, building a directory simply by retrieving and indexing the SMBmeta data.
In other words, SMBmeta is a self-organizing directory. In contrast with X.500 and its descendants, SMBmeta is very decentralized, pushing responsibility and data ownership all the way out to the edge. It makes data aggregation a loosely coupled operation that anyone can perform. Pretty cool.
What needs to happen is that the SMBmeta interface and directories need to be combined with an RSS-based publish-subscribe mechanism to create an information marketplace. SMBmeta is too static; it needs a dynamic mechanism (which can be built via RSS publishing) to enable SMEs to find each other and do business.