Summary

A developer asks:

We've completed the requirements and defined the class structure of our first project.  I'm about to run the Application Accelerator since there is only one process flow in the application. 

So far, the project team has deferred discussion of the organization structure.  I wish to understand the impact of deferring this part of the requirements: that is starting to build the application now, and designing the organization structure later.  What will be the impact of that?  For example, the initial setup wizard asks me to enter at least one silo of organization info. Operator IDs also require that information.  What other areas may become a problem later? 

We have deferred the org structure design because ours is a huge organization. People have different views about org structure design - along management hierarchy, operational groups, line of products, etc.  Also, the org structure is dynamic, incurs frequent changes. 

• Can we accommodate that dynamism within Process Commander on a continuing basis? 
• If I define an org structure for one application, can I define separate, others for other applications we build? 
• We have heard that we should look at LDAP or Active Directory configurations, and base at least the first two levels of the Process Commander org structure on that, to simplify any later integration for single sign-on and other features. Is that true?

Suggested Approach

Keep in mind the things that the org structure is actually used for.  You have three levels of org structure: Organization, Division, and Unit.

Typically, the organization structure is used primarily by routing functions. If more needs to be done with your institution's organization structure, your organization probably has an LDAP server that already records all the needed relationships.  

You can define separate organization structures for each application.. Note that the GenerateID activity (that derives work object IDs) is sensitive to the organization level of the org structure.

Applications typically do not rely heavily on the org structure, especially if an institution's org structure is stored in an external repository such as LDAP.  Instead, when a user is authenticated through LDAP, you also use some of the information in LDAP for authorization data, what access group to give the person, what OrgUnit the person belongs to, etc.  So in that case you really do rely on LDAP information, not the Organization, Division and Unit structure in Process Commander.