Adding terminology "WoT Profile" #865
Conversation
mlagally
requested review from
egekorkan,
mmccool,
sebastiankb,
chachamimm,
k-toumura and
mryuichi
Terminology entries for "WoT Profile" and "Profile"
| <dt> | ||
| <dfn>Profile</dfn> | ||
| </dt> | ||
| <dd>A technical specification which provides a set of assertions such that any <a>Consumer</a> |
There was a problem hiding this comment.
Two comments here:
- do we need another definition what "out-of-the-box" mean?
- I think, the definition can be more clear that the Profile is a subset of the TD specification by providing additional assertions.
Proposal: A technical specification that is a subset of the Thing Description specification which provides additional set of restrictive assertions such that any Consumer which conforms with the those assertions is out-of-the-box interoperable with any Thing which also conforms with those assertions.
There was a problem hiding this comment.
@sebastiankb What does it mean to be a subset of a specification?
A profile is a collection of assertions which are neither a subset nor a superset of the assertions in the Thing Description specification, they are a separate collection of assertions with may define additional constraints and assumptions for conformant Consumers and Things.
I suspect what you're trying to say is that a Thing Description conforming to a profile can only use a subset of the features defined in the Thing Description specification, but that isn't really an accurate description either. A profile may define a subset of TD features (such as security schemes, protocol bindings, and semantic contexts) which a conformant Consumer must support, but it may also define additional assumptions (such as protocol binding defaults), some of which can't even be described declaratively in a Thing Description. That could be described as a superset of the features described in the TD specification.
This current definition of a profile has already been reviewed and landed in the WoT Profile specification, so whatever lands here will need to replace or be consistent with that definition. I prefer the current definition.
There was a problem hiding this comment.
Note that I've started a discussion about a "Profile Mechanism 2.0" which could be much more rigidly defined, but that is a different approach to profiles than is taken in the current specification.
|
I agree with Sebastian that we should keep the definition simple and not add extra embedded definitions. In fact I would not "guarantee" oobi because it's difficult to prove. The INTENT to improve it is enough. May I suggest: WoT Profile: A named subset of Thing Descriptions defined by a set of assertions. The intent of a profile is to improve interoperability between Consumers and Things whose Thing Descriptions conform with those assertions. |
|
@mmccool Please see my comment at #865 (comment) |
|
Arch call on Oct 27th: |
Implement the proposal from TD issue #1719
Preview | Diff