A few points from the presentation/discussion:

• This proposal provides more granularity around opening a new application or raising an intent to find a specific instance of an application that may already be open 
• Covers implementation for a multi-platform agent
• What is a platform (something like an inhouse container –> e.g., Tick42, OpenFin, Eikon, etc.) and is it the same as desktop agent?
  
• The group requested more time to review and discuss offline, specifically to understand the real or perceived deficiencies of the existing spec and this proposal addresses those. Gaps mentioned include 1) missing the concept of application vs application instance and 2) the way intents are defined does not provide enough exposure of functionality

There was continued discussion around what the first version should cover, e.g. how specific it should be in the first instance vs. what should be deferred to a subsequent version following experience and insight gained from initial use.

Aim is to determine if the WG is ready to bring to vote.

The group reviewed the remaining issues with the core one revolving around intents and the expected response. There was discussion around the expectation of what will happen once an intent is raised (e.g. immediate response or the potential waiting), including challenges dealing with UX and interoperability.

Some examples of points raised include:

• Riko raised a concern with raiseIntentPrompt and implication that there is a user action
• Colin agreed → what never seems to come up is context?
• Nick asked a question about whether work around launchers was a center piece or whether it was more of an edge case.

The group did not reach consensus and questioned whether a larger sample size was necessary. The point was raised that it might be useful to see implementations and make decisions following that. This prompted consideration of moving to “validate” rather than “ratify” in the first instance and request people to implement what is in the spec in order to get real feedback.

The group agreed to use the mailing list to agree next steps as they were at an impasse and there were not enough people to make a decision.