How would you design a distributed decision-making system?

This sounds like it might be interesting if I knew what it meant.

Can you be a more concrete? Give an example of the type of DMS you are talking about? What kinds of problems do you want to solve? What kind of data? Is this aggregated human decision making or something like an expert machine? Are we talking some sort of learning algorithm? What do you mean by visibility?

@roundsquare I added some more context. Any kind of data, but you can use an example if that helps you. Decisions could be computer-aided. You definitely have to crunch statistics of say an affected community (as in someone doing something disagreeable affecting a community).

In terms of implementation I suppose I’d start with a Drupal application, mostly since I have a Drupal install already set up. Undoubtedly this would involve coding up one or more custom modules. Sadly PHP is the only language in which I have any measurable competence so while Drupal isn’t the best framework possible for such an endeavor, it’s the best one for me right now were I to implement such a thing.
I haven’t checked out your earthengine link, but I expect I’d crib shamelessly from it.
Basically, I am just barely competent if that to design or implement such a thing, and therefore my answer is essentially useless. Oh well.

Off the top of my head, I could see constructing a system for a closed group in something like Sharepoint WSS. Problems/decisions could be documented in discussion format, allowing for shared imput. Documents could be reviewed and edited by multiple users. Wiki formatted pages could used to share past history and context setting information, but also a history of decisions while keeping them editable. A multipart web page format would allow for embedding links to third party sites containing information relevant to the project. A workflow would allow for the routing of decisions/sign-offs/delegated tasks assigned to group members.

I am not a computer person, but I am have worked in a variety of real-world situations where such an application would be beneficial. The physical model we have used is that of the Incident Command System and I think it might be practically applied to the question (not to mention that automating ICS functions would have practical benefit). There is a free computer course at FEMA.gov called ICS 100 that explains the basics pretty well.

Essentially ICS was developed by the National Fire Service when companies responded from across the country to address western wild fires. Their radios didn’t always talk to each other and they each had their own command structure which made it insane to coordinate efforts and track accomplishments. We have implemented the very same design at our state health department to address outbreaks of concern (food-borne/water-borne illnesses, etc.) with great success. ICS does a great job of making all the tasks visible to all the members of the group and dependent tasks can be identified—it just all happens on paper in the settings I have used it.

The only exception is a card-based system using little pockets to track deployed resources (people, firetrucks, ambulances, search dogs, etc.) and log when they return, track communications, etc.

At the risk of providing a completely misinformed answer: wasn’t Haskell a language that would be quite suitable here?