Re: Is behavior of a single element in scope

Some suggestions inline!

"Rosen, Brian" wrote:

> I have a question to pose:
>
> Is a discussion of the behavior of a single element in
> scope for discussion? For standardization?

The access part (request/reply) and the format of access should be
standardized.

>
> For example, suppose we have an element called a Location Server,
> which knew the location of many "targets".
>
> Can we talk about what the Location Server can or
> cannot do with the data?

This would restrict the implementation of Location Server. There are
many proposals of Location Framework (LIF, WAP etc.) but they all talk
about the mechanisms of location access, to retrieve the target location
object, but not the implementation. The Location Server should be free
to do/manage whatever it wants to with regards to location data.

>
>
> In the IETF we do protocols. Even in a case,
> like geopriv, where we are defining an object,
> it is a protocol object, which goes between (at least)
> two elements. We can talk about what information
> is passed, and how. We can talk about how it is
> protected in transit.

Yes, this sounds like the right idea.

>
>
> We can clearly talk about how an element MIGHT
> use the information as background for why we
> might need specific information elements or
> protocol mechanism.
>
> However, we don't standardize what a single
> element does with the data.

True and it limits the implementation which needs to be avoided.

regards,

Bobby Sardana.
sardana@obsoft.com

>
>
> Specifying behavior is a slippery slope,
> and I'd like to stay away from it.
>
> Brian
Received on Wed Mar 20 22:08:21 2002