Convergence – Enterprise Applicability
For those of you (all three of you) who check in on a regular basis to this site, you may notice a running theme of Convergence. I am actually in the prcoess of writing a whitepaper on the parallel paths of the configuration management, knowledge management, collaboration and social networking worlds.
What is amazing to me is that the analysts and technology media have not yet picked up on this convergence, and the potential impact to the enterprise. Without giving away the store, just think about the problems with KM, for example. Traditional apps have concentrated on the storage of data, however trying to find data again is a painful process. So, the logical developmental step is to look at search capability, right? Wrong. The fundamental architecture of these systems is broken.
On the software configuration management side, they have articulated the concept of the “artifact” – which can be anything: a document, a line of code, a requirement, an mp3. It doesn’t matter. But in these systems, through branches and labeling you define the connections between these artifacts as you move along in the development timeline.
Now combine the power of KM storage and retrieval (for an individual or team) and CM tracking. What if every artifact within the system understood its relationship with every other artifact within the system? An artifact could not exist in the system without a relationship to another artifact. For example, a user of the system creates a set of requirements. This document is an artifact. Each individual requirement is an artifact. A product spec is then created based off of those requirements. The doc has a relationship to the spec, as does each individual requirements. Eventually a product is created, and customers buy that product. The system understands the chain, and every artifact connected to that chain.
Look at the potential impact of understanding this web of artifacts. The customers request a change. The change request polls the data, and through the data model, the product manager can understand the potential impacts across the entire product of making that change. If the request is to modify the design, the PM knows the impacts to the physical components inside. Change the capability, and the PM knows the impacts to the software.
Yes, yes, the object models could be enormous. But the DOD and other large bodies have been dealing with models in the multi-terabyte range for years. That aspect is not new ground. What is new ground is how we think about our data connections, and how making these connections will promote the coveted but never actually achieved concept of reuse.
More on this topic is coming…