Re: [Ecrit] AD Evaluation: draft-ietf-ecrit-lost-servicelistboundary-04

[Karl Heinz Wolf <[e-mail]>]

Robert,

On Tue, Nov 23, 2010 at 7:58 PM, Robert Sparks <[e-mail]> wrote:
> I've requested IETF Last call of this document.
>
> A question:
>
> The examples use a civic address boundary (of Lower Austria). Do you expect automata to be able to recognize when they have
> left these regions? (Would a phone be able to turn "Lower Austria" into a polygon so that it would know when it left?). Would it have
> made more sense to have constructed these examples with geospatial regions?
>

For clients knowing their civic address a civic boundary makes sense,
no geocoding desired.
In order to ensure that each client gets suitable boundary information
Section 3.3 says:

   The server ... MUST use at least one profile that the client used in
   the request in order to ensure that the client is able to process the
   boundary information.


> ----
> Some suggestions:
>
> Since the introduction is long, it would help to move the main point of the introduction (currently in the last paragraph) to the front of the section.

ok

> On page 4, 2nd to last paragraph, the sentence "Nevertheless, the client does not detect this, because only the mapping of the initially discovered services (police, ambulance, fire) are refreshed." is hard to parse.
> I suggest "Since the client is only required to refresh the mappings for the initially discovered services, the new service is not detected."

ok

> At the top of page 10, the phrase "so no boundary information is reasonable" is ambiguous. Is it trying to say
> "there is no boundary information that is reasonable to send" or "it is reasonable to send no boundary information"?

Would putting the sentence this way make it clearer:
The <listServices> request is purely for diagnostic purposes and does
not contain location information at all, so boundary information
cannot be calculated.


Thank you for your suggestions, I'll note them.

Karl Heinz