From: jcthomas@us.ibm.com (John C Thomas)

A couple of other pointers….First, Christopher Alexander's newest books, The Nature of Order. In these, he outlines fifteen “properties” of good form. These apply, according to CA, to beauty in nature as well as good architecture, but further to any design whatsoever. Now, compared with his “Patterns”, these properties are at once more abstract, require more creativity in application, more interpretation in even knowing whether a property is “there” and more generative. Nevertheless, I think they are good guides for visual representations. I've tried applying these properties to the design of stories, organizations, human-computer interaction, and excellent educational experiences and they always seem to “work.”

For example, one of the properties is “Levels of Scale.” (In somewhat different terms, Tufte talks about this too). Basically, this says (in the current context) that you should be able to look quickly at a representation and get an overview, a general feeling for main trends, and then “zoom in” (either mentally, visually, by walking, or via computer) to get more detail, and keep zooming. Ben Bederson at the University of Maryland has developed some software to allow such “zooming” quite nicely, but one shouldn't be trapped into thinking that “levels of scale” require a computer. Tufte has some nice examples in some of his books, and CA's point about traditional beauty in architecture is that as you approach the building, you see more and more detail.

Here are the fifteen properties:

1. Levels of scale.

2. Strong centers.

3. Boundaries.

4. Alternating repetition.

5. Positive space.

6. Good shape.

7. Local symmetries.

8. Deep interlock and ambiguity.

9. Contrast.

10. Gradients.

11. Roughness.

12. Echoes.

13. The Void.

14. Simplicity and Inner Calm.

15. Not-separateness.

Here is an attempt to apply these to the socio-technical domain itself. I hope this doesn't confuse the issue; I'm just trying to give people a better idea what the properties are. They can purportedly be applied BOTH to the socio-technical domain itself (as below) AND to the design of visual representations to help understand complex systems — such as social systems.

Socio-technical patterns should have a wide variety of levels of scale. There are patterns about how companies might be organized, how large-scale projects should be organized, about how to run hour long meetings, and about conversational turn-taking.

In terms of IT support, it would be nice if one could “replicate” up and down scales in some appropriate fashion. “Appropriate” would probably not mean exact duplication but intelligent duplication. If you spent time organizing a CSCW conference, e.g., it would be nice to be able to “pull out” the structures from that and apply some of what you did to a CHI conference or a IUI conference. The system might give guidance about what was likely to change as you change the size of the enterprise. Similarly, imagine someone is promoted from first line manager to second line manager where they are now managing managers. How could their IT infrastructure “respond” to this appropriately?

This could be related to goals, purpose, vision, corporate culture etc. But note that just as in architecture, it isn't enough to simply “declare” something a center, it surely isn't enough for a collaborative group to simply “state” their vision, principles, etc. For the center to be strong, it must be reinforced by behavior, ritual, visual stimuli, etc. Perhaps we could even measure the extent to which an organization had a “strong center.” Just to take one example, suppose we ran “Value Miner” across the e-mail and websites of an organization. Suppose we had this normalized profile of values: Innovation 60; Efficiency 15; Customer Service 24; Misc. 1 Would this be a more integrated organization than one whose profile was: Innovation 10; Efficiency 10; Customer Service 10; Saving money 10; Misc 60. ?

Aries de Geus, in his study of large companies that had longevity found four properties of such companies:

  • Financial Conservatism
  • Tolerance for exploration and innovation at the edges
  • High levels of mutual trust
  • Strong boundaries to get into the organization — but once in, you

could stay.

Dave Snowden talks about the importance of boundaries in some of his papers as well.

The rhythm of work runs deep – except in electronic collaboration. From pulling fishing nets, to French polling of canoes up river, to dance, people have traditionally coordinated their activities temporally. Huge hole and opportunity if it can be done in an acceptable, elegant, non-hoakey way. Not only would things work more “smoothly”; I strongly suspect people would feel more energetic at the end of the day.

My cousin, a clinical psychologist told me of a study of videotapes that showed a family interacting. When slowed down, it showed the family moving more or less in synchrony except for a schizophrenic daughter. Whenever she changed to get “in synch” with the family, they “cooperated” to change the rhythm and leave her out again.

This I interpret as the opposite of: “not my job.” While it may be true that too many cooks spoil the soup, it's better to have people (and systems?) that are always on the lookout for how to expand what they are doing outward to fill in all the possibilities rather than shrinking away from responsibility.

It is a bit abstract or metaphorical to see how this applies to organizational design. My best guess is this. The suborganizations themselves all have a distinct “shape” – meaning a distinct purpose and function and character that “holds together”conceptually in an elegant rather than an arbitrary way. At the same time, looked at from a larger perspective, these suborganizations participate in forming a coherency at the next level. An example of trying to do this (whatever the underlying reality) is the map that Nicholas Negroponte draws of the Media Lab and its parts and constituencies. An organization that has “good shape” has a balance and splits that are based on something more fundamental than accidents of history, friendship, or politics.

Could we imagine an IT tool that would help people “visualize” the relevant socio-technical systems? Certainly, some hints of this have been proposed such as Xerox PARC's hyperbolic trees for organizations and Tom's (and Harry Potter's!) office social proxy showing location and activity. But, could we show the activities so that, at a glance, one could tell whether people were tending to expand outward or shrink back?

Reciprocity. Mutual trust. Mutual respect. Lack of symmetry, in human relations, or in computer systems or their interaction can lead to difficulties. Is there any way we could “instrument” such things? For instance, would it be worthwhile simply to have a “map” of how much e-mail you sent person X versus how much they sent you? Would we want to show the number of e-mails, the specificity (are you just one of a hundred recipients) or the length of the content?

Ambiguity? Surely, this is something that an organization cannot want. But is that true?

Is IBM Research supposed to be doing long term research or helping the company solve today's problems? Both is the answer. And, in fact, to the extent that one can achieve a deep interlock between these apparently contradictory goals, the better. Is a sales person's job to maximize sales revenue or satisfy the customer? Both. And, again, to the extent that conditions are created that make for a deep interlock and ambiguity so that the salesperson themselves feels that they are doing both; that these are parts of one whole, we have a well-functioning sales organization.

Unfortunately, the thinking behind most current IT tools is that things must be dichotomized and linearly prioritized. The idea of supporting ambiguity or paradox is lacking but it could be a huge win if we could wrap our mind around it and actually develop an architecture capable of dealing with such a concept.

I take this to mean that an organization must create within it dynamic tensions of opposites. In animals, there are pre-existing, well-defined, and opposite tendencies of behavior. The contrasts can be shaded by events but it is much better to have an animal that sometimes sleeps and sometimes is awake than one that is always half-awake. It is better to sometimes fight and sometimes flee than to always fight half-heartedly.

Similarly, an organization needs contrasts of people and of function and of activity. A healthy organization should have people who are complete optimists and believe anything is possible – and complete pessimists who question everything. An organization should have an organization (or process) whose purpose is to expand the company in every possible direction and an organization (or process) whose purpose is to contract the company as much as possible. When brainstorming, to be as effective as possible, no real-world constraints should be allowed. When choosing which brainstorming ideas to pursue, every real-world constraint should be applied.

Helping individuals and organizations do some meta-cognition about the appropriate times, places, roles for various activities would be quite helpful.

Taken together, Contrast and Interlock, as well as Gradients and Boundaries, would seem to push design in opposite directions. Yet, there are architectural examples that seem to provide both of each property pair simultaneously. Living organisms also simultaneously exhibit both properties. An unresolved issue is when, how, where, and in what degree do we push Gradient more versus Boundary more. When should Contrast be emphasized and when Deep Interlock and Ambiguity?

It would seem off-hand that traditional command and control organizations have tended to emphasize Contrast and Boundaries to the detriment of Interlock and Gradients. One would hope that in an adaptive organization, people would pitch in more and help each other out “across” organizational boundaries.

Some examples of where gradients might be effective might include the following. One could imagine a gradient funding system wherein projects would not be either “in” or “out” of a plan, but gradually get (or lose) more funding as the benefits and costs became clearer. Organizations already use a gradient market introduction system where successive “trials” allow for increasing commitment to a product with favorable results. One can also conceptualize summer employment as a chance for company and individual to examine the suitability of longer term employment. Often cross-organizational activities that result in mergers and acquisitions begin as much more limited partnering arrangements. A natural example of gradient might be that very large customers get very large account teams while progressively smaller customers get smaller account teams.

In some cases, it would be nice to “associate” e-mail items, documents, presentations etc. with various folders in various degrees. I can picture a pretty smooth way to do this even more quickly than the current NOTES implementation of moving things to folders.

I think this relates directly to Austin Henderson's notion of “pliant systems” — you will never be able to figure out a priori exactly how to “optimize” any large complex system ahead of time. (Detailed centralized planning like the Russian “five year plans” don't work). Leave some flexibility and local control.

This seems most naturally construed as the organization pulling toward an overall vision so that many different aspects of the activities have the same “flavor.” If Customer Service is paramount, that should manifest itself in a thousand small ways. Each of the individual acts that provide excellent customer service is different, but each is an echo of each other and an echo of the larger whole – the vision.

With IT tools, such as the “Value Miner” we've been working on, one could imagine having the IT infrastructure actually check on the degree to which their are “echoes” of the vision in everyday activities.

Another sense of “echoes” is in the idea of inheritance. It's still the case in many current tools that you have to “respecify” things in a large number of cases. It would be nice, e.g., if, once I filled in the destination and purpose for a trip in an e-mail to my manager, that that information would automatically be entered into an expense account form. In fact, my calendar should note when the trip is over and send me e-mail with the attached expense account to be filled out and the system should know who the account has to be okay'd by (with potential human override).

I take this to mean that in order to be adaptable, organizations, team, work practices, etc. need to have some unfilled space. If every moment is taken up by assigned and defined tasks, there is no time and resource for adaptation. We should work in time and place and perhaps IT support for reflection. Hard sell in today's world.

Policies, procedures, practices have a way of becoming nightmarishly convoluted and self-protective over time. I think of this as organizational “scar tissue.” A system to help teams, groups, organizations and individuals identify and remove (now) unneeded bureaucracy would be a huge enabler of increased (quantum) productivity. Of course, this was somewhat the idea of “Process Engineering”; unfortunately, that mainly failed because in practice people never bothered to understand how work was actually done but did their redesigns on the basis of management fantasies of how it was done.

This relates both to your recognition that work/life has more interpenetration and that the organization does not live in a vacuum but must relate to the environment.

– John C Thomas ( ttp:www.research.ibm.com/knowsoc \ \ http://www.truthtable.com/

  • nature_of_order.txt
  • Last modified: 2007-06-13 07:33
  • by nik