Re: CARD review - Minor issues

Vijay Devarapalli wrote:

> Marco Liebsch wrote:
>
>>
>>
>> Vijay Devarapalli wrote:
>>
>>>>   L2 type:       Indicates the interface type (optional).
>>>>
>>>>                  If the L2 type indicator is not used, this field MUST
>>>>                  be set to 0x00.
>>>>
>>>>                  The following types are initially defined:
>>>>
>>>>                  Technology    |  L2 type
>>>>                  --------------+---------
>>>>                  IEEE802.11    |   T.B.A.
>>>>                  CDMA2000      |   T.B.A.
>>>>                  WCDMA         |   T.B.A.
>>>>
>>>
>>> I think IANA does not have assign the L2 type. it can be done in
>>> this draft itself. I know I was the one who raised this issue
>>> earlier. but now I realise IANA's role is not needed for the L2
>>> type. my mistake.
>>>
>>>
>>> Vijay
>>>
>> I am not sure about this. If we want type identifiers to be valid beyond
>> the scope of a specific operator's domain or a specific implementation,
>> the control on these numers should be assigned to somebody, ensuring
>> their use in a unique manner.
>>
>> Why do you think this is not necessary, or, more precisely, who,
>> if not IANA, should take over the control function on these identifiers?
>
>
> define them in the CARD protocol document itself. anybody who implements
> CARD will use the same type values from this document. there shouldnt be
> any interoperability problems.


This works for some selected access technologies. What about future ones or
some we might forget? Preferably would be to use something existing or
at least something that can be re-used by others when specified once.
Wouldn't be reasonable if every protocol defines its own types, or?

marco


>
> Vijay
>