Alan said;
>
>The definition of a function seems to indicate that a function is a unit.
>
It CAN be considered that way in a ceratin implementation, but we CONNOT
say that is the ONLY way we have to see it.
(Especially in the case of FDS, where we have to start thinking
about other protocols.)
At one point in the discussison, Brian had mentioned that some people may
consider a service as a function, and visa-versa. I agree to this idea, and I guess
we are not allowed to give an answer (restriction) to the relationship between a
function and a unit.
We agreed a unit corresponds to a driver (in the case of the PWG profile).
If so, the relationship of a driver to a function is implementation specific.
The only restrictions I think I can make for "function" are :
1) A part of the unit architecture (not meaning a part of the unit, but meaning
a part of this architecture i.e. part of a node, module, or whatever)
2) A collection(meaning 1 or more) of services
3) User identifiable
I cannot think of more resrtictions to add and still make the definition consistant
within 1212
ats
>
>by Atsushi Nakamura
>----
>My shot at definition of "function":
>A function is part of the unit architecture and is an instance of a device
>model, e.g., mass storage, CD-ROM or printer. A function is usually a
>collection of services that are user identifiable.
>-----
>
>The definition of a function seems to indicate that a function is a unit.
>