I appreciate your initiate in working to improve the straw man document. I agree with some of your suggestions and disagree with others. One key difference in our thinking is that I view the definitions if "type" as a logical architecture, while your perspective is that we need to address the physical architecture. Perhaps we need both. The group will have to decide.
Below is my feedback on your proposed changes:
- Intro bullet 1 device vs. instance. I can accept that "device" might be misleading and cause people to only think in terms of hardware. I understand your point about much of the heavy lifting of a processor being performed by software, but I believe that the term "instance" is so generic that it will create more confusion. Whether an application is on a chip, a dedicated device, a shared device, or a device in the cloud, it must run on something. The processor is really a combination of the application and the device it runs on. We might want to change the word "device," but I am uncomfortable with substituting "instance". Perhaps we should merely call it a "thing" -- one of many components on the Internet of Things. But, personally, I think that the word "thing" doesn't sound precise enough.
- Intro bullet 2 adding the concept of identify relationships to the discussion of ownership. Identity relationships are important, but I think that they are "second order" characteristics. Because identity relationships are likely to be in continuous flux and new roles may arise for many "things," I don't think that they are as fundamental as type, ownership, and accessibility. Perhaps they represent a fourth category. But I don't think that they should be mixed with ownership.
- Figure. 1 changing the image of processor in Figure 1 to something more "up-to-date". I have no objection to this recommendation. Have at it.
- 1.2 Data Processor vs. Data Processor/Processor Application. I would give the same argument here as for the device issue in #1 above. Admittedly all processing requires an application, but I don't know that the distinction is important enough to complicate the picture. Just because processing takes place in the cloud rather than locally changes the physical architecture but not the logical architecture.
- 1.4 Adding gateway as a device type. I see "gateway" as an element of the network, not of the "things" on the network. The "Internet of Things" isn't just a collection of things; it is a collection of things connected by a network infrastructure. I see "gateway" as part of this network infrastructure, rather than one of the "things" being integrated.
- 1.5 additions to definition of "combination". Based on my discussion regarding gateways above, I do not believe that the addition is appropriate.
- 2. Relationships description. I don't disagree with any of the points made about the different relationships described here. I just think that they are independent of the notion of ownership and should be addressed separately. The various "relationship" roles all need to be allocated by the "owner" -- though the owner may choose to delegate some of his authority.
- 3.1 additions to the definition of "discoverable". I have no objection to the additions
- 3.2 additional to the definition of "not discoverable". I would argue that a "thing" not connected to the network is not a member of the "Internet of Things" and, therefore, falls outside of the scope of the document. For this reason, I would not support the addition.