Re: [Insipid] Reviews for INSIPID Session-ID solution draft Version 11

[Paul Kyzivat <[e-mail]>]

On 1/22/15 1:47 AM, [e-mail] wrote:

>>> Section 4.1 (and Section 5)
>>>
>>> Would it be worth to mention if the Session-id is case sensitive.
>>> Since when we will need to implement it we have to explicit state it if
>>> it is case sensitive or not.
>>
>> Per RFC 3261, "field names are always case-insensitive". I don't think
>> that needs to be repeated here.
>
> [RJ] I looking here from Operator point of view. We have so many discussions about the case (in-) sensivity of fields with vendors who should know. I'm partly OK with your proposal. But nevertheless it will led to wrong implementations seen from my experience with vendors. And the point the following point will confuse a little. On one point we say UUID lower case but RFC3261 points that header filds are case insensitive. Thus I would at least prefer a note.
>
>>>
>>> I think it would be worth to mention that the UUID is case insensitive
>>> as defined in RFC 4122
>>
>> The text currently states that the UUID characters are lower-case.  The
>> syntax also enforces this.  The RFC 4122 text is funny, because it says
>> "lower case characters and are case insensitive".  I don't know what
>> that means, except that "F" is not a valid character.  Rather than
>> create confusing language, I'd rather just leave the text as it is in
>> saying the characters are lowercase.
>
> [RJ] OK

The *safest* (belt and suspenders) way of dealing with this would be to 
require that senders always insert lower case, but require that 
receivers properly process either upper/lower case.

IMO it *should* always have been case-insensitive. IIRC the reason it 
isn't that way is because Kaplan was case sensitive and there are 
implementations in the wild that depend on it.

	Thanks,
	Paul