The New Work of Information Managers
At Northern Lights, Inc. we spend a lot of time talking to executives, deans, directors, vice presidents, provosts, etc. about information services. Specifically, we talk about how to create information services that are more relevant, more tailored, more value-added. We talk about how to accomplish this within the range of tight budgets, limited personnel and with maximum return on the institution's investment. Managing information is expensive. Not managing it can be even more expensive. Dollars are hard to come by and how they're spent, especially when it comes to technology, is important to many people who must produce results.
Our discussions center around how to plan and design cost effective, value-added versions of information services as well as the mechanisms for accomplishing their creation and implementation. Most recently, much of our work has emphasized approaches like web services, software solutions like content management systems, and notations like XML (for a common language of integration and data exchange.)
I want to mention a point about XML that has emerged as critical in all these discussions. In fact, I believe it may be critical to the future of information management services in general and to the education of all information managers in particular.
XML is important to current technology trends for many reasons. Whole websites and entire shelves of bookstores are devoted to it. In the context of our work, we're always looking for ways to discuss with our clients (and their constituents) how to add value to information services. XML is certainly one of the ways value can be added. One of the most basic is also, I believe, the most important.
Whenever information is managed, whether it's a single document or an entire body of content, it is modeled. In a content management context, for example, that modeling is handled by analyzing the body of content within a given content management system's domain and creating a content model, based on that analysis. I'm making something very simple that isn't, but what you're basically doing is creating an abstract version of the information. This abstract version is often very complex and sophisticated, especially if it's prepared by experienced information managers in concert with experienced users.
To be useful in a machine-processable environment - in other words, as part of some kind of information management system - that abstract model must be formalized and created in a way that allows it to interact with software. To use a content management example again, you'd formally notate the various content types in your content model, using XML notation. Each one of these content types would contain elements, subelements and attributes that would describe it. Of course, a set of files filled with XML versions of content types don't begin to capture the richness of the abstract content model.
To go further toward capturing more facets of the content model, you'd also have to add in XML versions of access structures. Access structures are the formal way the content is made discoverable and include a wide range of things like indexes, cross-references, associations, hierarchies, etc. Metadata can be created to add all kinds of value to the content. It can describe what elements should be used depending on what form of output the content takes. It can describe the level of user sophistication best suited to a given set of elements, like novice, intermediate or advanced. There are also a series of rules that describe how metadata is to be created and applied. All of these mechanisms and approaches combine to try and realize the fullness of the abstract content model.
What this means is that one of the good things XML facilitates is the transformation of a rich, sophisticated abstract model into a formal set of structures that make the abstract model live as a fully machine-processable entity. This is very important since it's the ability to machine-process that directly affects the per unit cost of information management services and that leverages the results of those services as far as a given system can be distributed. If we're talking about distributing a set of information services over a campus intranet or over the Internet, we're talking about a very wide range indeed.
It seems to me the future of information management as well as the future training of its practioners may well be directly related to continually enriching the ability to translate the inherent richness of abstract models of information into machine-processable ones.
Comments