The problems of information overload, search engines, privacy, customization, dynamic interfaces, access control, trust, permissions, etc. are all connected to a central point.

The point is: what information does a person, process, or group need to know, show, and hide, and what informational activities do they need to perform, avoid, or be aware of.

With the advent of XML, the tagging of information content will progress, providing fine grained elements of information that are maintainable and distributable. XML-related mechanisms will continue to evolve to process that information and deliver it.

The question is - what information is to be delivered, in what form and when, to whom (person, process, group), under what circumstances?

This is where profiling becomes most important. Recognizing that "profile" is a loaded word, having many different meanings to many people, what I mean by "profile" is: the assocation of one object to another. When an object's (person, place or thing) profiles are combined together, whether the profile is relevant in past, present, or future, the result is that I call the object's context. The context shows the object taxonomic, derivative, container, and component associations, in the past, present, and future. Another way to describe context is that it is the collection of categories/groups to which an object is assigned.

I've organized the groups to which an object can belong into six dimensions: location, organization, workforce, function, process, and resource. If you want to know, hide, or show information or perform or avoid informational activity, you would probably describe this need in terms of one or more of these six dimensions. Each of these dimensions represents what is called a hierarchical namespace, such as a tree of categorized information, a multilevel catalog, an organization chart, a map, a financial account structure, a biological taxonomy, a strategic plan, etc.

For example: I need to know what's happening in location A, which contains my organization B, which has assigned me to workforce (office/role) C, where I perform function (executive, production, resourcing) D, by applying process E, in producing (or using) resource F. The string of these profile pairs could be arranged in any order within the six dimensions to answer questions, provide or hide information, or enable or block activity.

This type of overlapping group membership is at the heart of what is now being called policy-based management, which I have previously called Context Management.

While XML allows us to tag the universe of information, hierarchical namespace management technologies, such as LDAP and XML, allow us to 1) manage the resultant millions of nouns (objects) and their attributes and taxonomies; 2) manage the associations/profiles and their attributes; 3) manage the changes within objects and profiles to enable syncrhonization, history and planning functions; and 4) manage the internal behavior of the named objects and their associative interactions.

A new company (http://www.bowsoftware.com) has recently announced a product (combining XML and LDAP) to provide what appears to be a subset of the capability I've described here. They call it "Internet Channel Management", and are getting major support for their efforts (IBM, Microsoft, Novell, etc.).

The concepts presented here are covered in more detail, at no cost, at /rer/owis/. I specifically propose this Context Management approach for use in Enterprise Integrated Systems, Services, and Applications Management (ISSAM) as shown at /rer/owis/ldap-ca.htm.