How to use 'output' section - new use case needed

The output can be used in two diff. way:
1. To display output to the use at the end of template deployment
2. Pass the output to a new template (i.e. use output as an input to a new template)

Matt Rutkowski (IBM): In v1.0, output parameters are associated with Plans or operations. The semantics of having outputs defined here in at the top of a topology, should be clarified: are we saying that these outputs are automatically settled after some initial deployment/provisioning of a new instance? Who is supposed to use these outputs? Can a single node define its own outputs, available as soon as this node is deployed? Could these outputs change e.g. after some reconfiguration of the instance?

Matt Rutkowski (IBM): Compared to v1.0 those outputs kind of correspond to the Properties in the BoundaryDefintiions of a ServiceTemplate.
It represents data that you want to expose to the user of the template to avoid him having to scan the internals of the template in order to get information important to him (like a web URL).
This features becomes important also for nested templates (I am working on it).
Single nodes cannot expose outputs. It is a decision of the template author which ones shall be exposed outside of the template.
And yes, the values could change, since they are linked to node properties.

Chris Lauwers: Make a statement in-line that lists "uses" of outputs: 1. primarily used for displaying feedback to the user. 2. Also used for substitution mapping. 3. Service chaining.

Chris Lauwers: Make statement that output values can change as attribute values change during operation.