Re: GIST Q/D-mode issue

Greetings,

On Sep 22, 2008, at 9:37 PM, Teemu Huovila wrote:

> Hello
>
> On Wed, 17 Sep 2008, Robert Hancock wrote:
>> short summary: my preference at the moment is to
>> use the flag bit method, but to put the flag at
>> the start of the message type field. (saves a bit,
>> and brings out the relationship between flag and
>> message type.)
> Roberts reasoning for putting the information in the message field  
> made
> good sense to me. Though I can not say I fully comprehended all the
> implications, which Robert mentioned, of the original issue  
> pertaining to
> NAT boxes. I also endorse the "message type division" option.
>

A flag seems the most straightforward solution to the problem from an  
implementation's
point of view, so I also fully support it. Dividing the message type  
space into stateless
and stateful messages seems a good idea from extendability and  
philosophical point of view.

BR, Nuutti