On 19/03/2013 20:39, Michael Thomas wrote:
> On 03/14/2013 01:43 AM, Brian E Carpenter wrote:
>> On 13/03/2013 20:54, Ted Lemon wrote:
>>> On Mar 13, 2013, at 4:01 PM, Michael Thomas <[email protected]> wrote:
>>>> All of this is lacking in section 3.7. If I were a contractor using
>>>> this "architecture"
>>>> I wouldn't know what to build.
>>> The architecture document acts as a basis for the work that you, as a
>>> contractor, would build from.   You can't build from an architecture
>>> document, unless you are a lot smarter than we are.
>> OK, but that doesn't mean that the *architecture* should punt on
>> principles
>> of the namespace.
>>
>> I'll put down a marker: local namespaces are ambiguous and therefore
>> should
>> not be part of an architecture.
> 
> I agree.
> 
>> However, there is a way to reconcile this
>> with pragmatic use of .local - stipulate that when a global name is
>> needed
>> for a local resource, it is <name>.local.<subscriber>.<isp>.
>>
>> That requires some namespace knitting but it is architecturally clean.
>>
> If you had a place to hook into the global namespace, why ever use .local
> at all?

No particular need, but since .local is out there already, it seems harmless.

   Brian
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to