The philosopher’s curse: a “breakthrough” is anything but. Needless to say, after yesterday’s blog post I questioned whether I really came up with anything original or exciting. Perhaps this is what separates philosophers from gurus. My process for cutting through the epiphany-generated haze of enthusiasm is to state an idea as simply as possible and test it against other examples.
The thought process yesterday took me to two assertions:
-
We express content management processes (ie: workflows) as stories, a collection of actors, objects, and events. This mechanism is most appropriate because it draws upon standard cognitive tools, according to Mark Turner. Transforming reality into stories makes it easier for us to represent those stories as workflows in content management systems.
There is a tacit understanding among people who participate in a business process that the process does not necessarily happen the same way every time. We make little adjustments to the process based on specific circumstances. Conceptual blending is a cogntive ability that allows us to see similarities between the prototypical case (the one documented as a workflow) and the present situation. That ability allows us to say things like:
Sarah’s out today and she usually does the final approval, but we’ll combine the final approval with the prior step, proofreading. Since James usually does the proofreading, he can take care of approving the document as well.
In this blend, the present circumstances (Sarah’s absense) are mapped to the prototypical workflow and we can make an adjustment to the process to accommodate. This idea builds on Turner’s notion that the mind is literary because we can create new stories out of existing stories and a new set of circumstances. Computers do not have literary minds, and can not therefore create these blends that allow them to envision new workflows on the fly.
This is the conclusion of the first assertion: that computers can’t do conceptual blending, and therefore must be programmed to account for every possible exception to a process. Probably not earth-shattering, but it offers a good foundation for analysis.
-
The second assertion applies this idea toward a better content management system. The second assertion suggests removing the computer’s burden of identifying an appropriate exception to a process, eliminating the “conceptual blending” responsibilities. Instead, we simply ask the computer to capture the circumstances and report them to a human, who can make a decision about what to do next.
To build a CMS on this idea would require two features:
- Some sort of reporting mechanism that makes it easy to report process disruptions like, “Sarah is out today” or “James’ responsibilities changed” or “The deadline has moved up” or “Management has determined a different course of action for this content.” Whatever.
- A flexible workflow mechanism that allows a user to make changes in response to process disruptions.
Tomorrow (if I have a chance) I’m going to look at how these two assertions can apply to a different content management concept. That will be the real test of their value.
June 14th, 2005 at 4:50 pm
I spent the last two days at a conference for records managers. They have some interesting perspectives on all this. I’m still sifting through it in my head, but here’s an initial gut reaction to what I heard. [full disclosure: I’m a trained records manager].
First of all, let me do a twenty-second intro for people who are not records managers: records are documentation of business processes, policies, practices, and key transactions. They are usually documents, but any media can be a record. A photograph or a whole database can be records too. A consultant’s deliverables, an employee’s annual review, a corporation’s 10K filing, and a citizen’s 1040 tax form are all examples of records. Records enable legal defense and continuity of operations, among other things. Our concepts of “government”, “citizenship”, “real estate” and “corporation” are based entirely on records. Records are literally the fabric of society.
Since records managers are concerned primarily with documenting process, there is a new (radical and ironic) idea that the documents themselves are not actually crucial to that mission - the process, if it is automated, may be documented directly in the workflow application. They are thinking that most processes have a few crucial steps that need to be documented, but the rest of the process is basically immaterial so there is no reason to document it for the long term. That is to say, they are proposing a sort of hybrid workflow model, where there are important/mandatory steps and not-important/optional steps. For example, assume a government agency produces a report that gets reviewed by a committee, commented by the public, and signed by a director. In this example, the records managers may only care that the committee review occurred (i.e. distribution list and date), but the results are not important. The public comments, on the other hand, are considered worthy of capture and storage. And the final signature makes the draft an official record. In document management, we would have three separate publication events, but records management looks at it as one process - simpler and more purely linear, in a way. Perhaps we have two records in this example - one automated aggregate to document the creation process itself, and another to document the final version that resulted. This (radical) concept puts a whole new spin on document management and workflow, because as far as records management is concerned (in this interpretation), all the draft versions are trash the second that final draft is signed. Not only that, but one of the two permanent records in cases like my example would be an automated aggregate accreted from the process itself rather than a “document” in the sense we normally mean.
Anyway, I don’t know how clear all that was. But it is interesting stuff.
June 14th, 2005 at 5:56 pm
Conceptual blending in content management
Dan Brown has written a blog entry on the interesting concept of conceptual blending in content management. To quote: There is a tacit understanding among people who participate in a business process that the process does not necessarily happen the…
June 14th, 2005 at 5:58 pm
Hi Dan, this parallels some of my thinking on this topic:
http://www.steptwo.com.au/papers/cmb_noworkflow/index.html