Rechartering Input from the TD TF

[egekorkan]

No big changes for the TD TF related deliverables in content but its management has changed in an important way. Below are the inputs we have from the TD call of 07 August 2025:

Normative Documents:

• TD 2.0
  • Alignment with other specs (arch, discovery)
  • New features: Initial Connection, Manageable Affordances, Data Mapping
  • Establishing a User Story driven feature development
• WoT Bindings Registry on the W3C Registry Track
  • Establishing the registry document
  • Establishing the mechanical process to accept entries

Informative Documents:

• Individual Group Notes each binding written by the WG
  • Working on adapting the documents to the registry requirements
  • Doing the testing needed for each document

[relu91]

Establishing a User Story-driven feature development

I think we worked enough on organizational topics. In the past months, we established a good consensus on how to write use cases and user stories. I would keep it as it is and move on with the actual features. A plus would be that while we work on features and use case/user stories definitions, we can even figure out how to lean the process even more.

The TD features mentioned above are definitely something we need to bring to reality. Another one is the discovery process for additional protocols (e.g, how a binding could extend the discovery process of the WoT by adding additional mechanisms).

Finally, even if the train has already passed, I'd like to keep an eye on the work of AI agents. It would be nice to continue to keep feedback and use cases from that world so that Thing Descriptions could be used not only to describe devices but also AI tools or Agents.

[benfrancis]

@relu91 wrote:

I think we worked enough on organizational topics. In the past months, we established a good consensus on how to write use cases and user stories. I would keep it as it is and move on with the actual features. A plus would be that while we work on features and use case/user stories definitions, we can even figure out how to lean the process even more.

I agree.

The TD features mentioned above are definitely something we need to bring to reality.

I agree.

Another one is the discovery process for additional protocols (e.g, how a binding could extend the discovery process of the WoT by adding additional mechanisms).

Is it a good idea for each protocol binding to have its own discovery mechanism? What is the motivation for this? Can you give an example?

Finally, even if the train has already passed, I'd like to keep an eye on the work of AI agents. It would be nice to continue to keep feedback and use cases from that world so that Thing Descriptions could be used not only to describe devices but also AI tools or Agents.

I think AI agents could make great Consumers of Things. I don't think Thing Descriptions should be used to describe AI agents, that seems like dangerous scope creep. There are lots of other standards in development specifically intended for that purpose.