I recently took an introductory course on Six Sigma which was both fun and educational. The course introduced SIPOC diagrams, used to model processes undergoing improvement. This experience had synchronicity with my post on ERP and Collaboration Systems, which discussed how existing systems don’t allow the modeling and enforcing of processes. So, a collaboration system that represented processes in a way compatible with SIPOC diagrams might provide a path to process improvement via Six Sigma.

A collaboration system that enabled and tracked the creation and modification of information artifacts by multiple users could in fact automate the representation of a process by detecting patterns in these operations. Since some collaboration systems allow rating of information artifacts, good processes (i.e. those resulting in highly-rated artifacts) or useful processes (i.e. those whose results are widely used by teams) could be distinguished from poor or unused processes. One practical challenge is that information artifacts are not today enabled and tracked only in collaboration systems, so such a system might have to integrate with e-mail, office applications, etc.

Once a collaboration system could represent a process, it would have to enable the definition or characterization of suitable inputs to the process and expected or guaranteed outputs from the process. The definition or characterization would have to be sufficiently general that outputs from one team’s process could become inputs to another team’s process, but sufficiently specific that the collaboration would not match all possible outputs with all possible inputs. Somewhere in the continuum from specific to general lies a region of productivity.

I imagine a user characterizing the input of his team’s process and receiving immediate feedback on how many outputs from other teams are a suitable match. The user could then express eagerness for cross-team collaboration by characterizing the input generally (accepting large numbers of outputs from other teams) or express unwillingness for cross-team collaboration by characterizing the input specifically (accepting few or no outputs from other teams).

The definition or characterization of an information artifact means the assignment of attributes. Tagging is a way to assign attributes in a manual fashion. File and document properties are automatically-generated attributes.

The envisioned system exhibits characteristics of the semantic web.

Database table and column names and the ranges of values in database columns all serve to characterize the table. However, some constraints may not be obvious from examination of column data types or of the values in the database. These values are positive examples only, and negative examples may be required.

Associated with each process represented in a collaboration system may be ways to determine whether a given information artifact is a usable input or not.

Examples of what might be possible in such a system:

• Two people create separate tables of data about customers. The system discovers similarity between the structure and content of the two tables. The system automatically fills missing values in one table from another, thereby enriching what each has already gathered.
• Two people create information artifacts whose structure and content are similar. The system introduces them to each other. This promotes collaboration, division of labor, and efficiency through non-duplication of effort.