writemultiple has difficult semantics, how about read only? Best effort approach? Fail, if one of the writes is failing?

[mlagally]

No description provided.

[sebastiankb]

I'm not sure if I understand this issue correctly. Please not that writemultipleproperties is already part of TD 1.0.

how about read only?

We have readmultipleproperties and readallproperties.

Best effort approach? Fail, if one of the writes is failing?

This is a question that depends on the protocol used and should be clarified in the protocol binding TF.

[mlagally]

@sebastiankb
My comment was probably a bit brief. The question is how writemultipleproperties should behave, when somebody tries to write a set of properties of which some are read-only. Should the entire operation fail? Should it succeed without notifying the caller that he made a mistake? Is it implementation dependent?

[benfrancis]

@sebastiankb

This is a question that depends on the protocol used and should be clarified in the protocol binding TF.

For example, this is being discussed for the HTTP Baseline Profile protocol binding in w3c/wot-profile#162

[sebastiankb]

My comment was probably a bit brief. The question is how writemultipleproperties should behave, when somebody tries to write a set of properties of which some are read-only. Should the entire operation fail? Should it succeed without notifying the caller that he made a mistake? Is it implementation dependent?

This is a good point but this is also true for other operations. E.g., so far the TD have not predefined how the servient should behave when, e.g., a write call is executed on a property that can only be read. TD offers the opportunity to define multiple responses which can be used to express if writemultipleproperties was successful or partial successful or not successful with specific reasons. The Profile can take advantage of this and establish certain guidelines.