Proposal: Follow traits schema to specify how app scopes are referenced by components

[mikkelhegn]

In the operational (application) configuration, it's not specified how the applicationScopes are referenced: https://github.com/microsoft/hydra-spec/blob/master/6.application_configuration.md#component

Suggesting that we follow the schema defined for traits: https://github.com/microsoft/hydra-spec/blob/master/6.application_configuration.md#trait

Follow-up to #119

[hongchaodeng]

Suggesting that we follow the schema defined for traits: https://github.com/microsoft/hydra-spec/blob/master/6.operational_configuration.md#trait

Thanks for the suggestion. But in the Configuration definition, it already contains the Scopes definitions in the same level as Components. I think the reason is that Scope could cross Component boundaries. In plain words, involve multiple components. So my conclusion is that we will keep Scopes definitions as is today, but define references in Components.

[hongchaodeng]

Let's move discussion to https://github.com/microsoft/hydra-spec/issues/144 ?

[mikkelhegn]

The problem is that we have only specified how scopes are created as part of an application configuration, not how components are added to a scope, unless we assume every component in an application configuration will be a member of all components in that one application configuration. That however goes against the whole idea of the application scopes.

[hongchaodeng]

I think the whole idea is to reuse and reduce definitions. In fact I'm fine with either solution. But in last weekly meeting, that was the unanimous decision. @MikkelHegn Could you join the weekly next time and discuss your concerns?