Re: [Iptel] rfc 3966 BNF queries

[Henning Schulzrinne <[e-mail]>]

I don't recall how and why this happened, but RFC 2806 (the  
predecessor) had

phonedigit            = DIGIT / visual-separator

This would have allowed multiple separators in sequence.

I also checked the draft history, and the original version in

http://tools.ietf.org/html/draft-ietf-iptel-rfc2806bis-00 and 01

has

honedigit            =  DIGIT [ visual-separator ]

which is close to what's being proposed. This changed to the current  
format in -02.

The only change log entry is

Changes from ietf-00 to ietf-01

         o Editorial changes suggested by Francois Audet.


I also checked my mail folder and the IETF mailing archive around the  
time 02 was issued, and there's no discussion on this feature that I  
can find.

Personally, I don't see the need for multiple consecutive separators.

Henning

On Apr 16, 2007, at 7:32 PM, Paul Kyzivat wrote:

> It strikes me that if this is to be judged as a "bug" then we need  
> to ascertain what the *intent* was and then come up with a valid  
> ABNF that realizes the intent. Otherwise we are simply devising  
> another extension.
>
> I don't know what the intent was. I can only guess at what seems  
> "reasonable" to me.
>
> 	Paul
>
> Eugene Seah wrote:
>> On 4/13/2007 8:43 PM, Henning Schulzrinne wrote:
>>> I agree. I guess the next step would be filing an RFC erratum.
>>>
>>> On Apr 13, 2007, at 8:00 PM, Paul Kyzivat wrote:
>>>
>>>>
>>>>
>>>> Eugene Seah wrote:
>>>>> Hi, two queries here on RFC 3966 BNF issues.
>>>>> Firstly, I was wondering if the optional BNF square brackets  
>>>>> around "visual-separator" in RFC 3966 was correct.
>>>>> i.e.
>>>>> phonedigit = DIGIT / [ visual-separator ]
>>>>> There's a discrepancy here with extension BNF, since phonedigit is
>>>>> essentially optional when visual-separator is made optional,  
>>>>> meaning the 1*
>>>>> prefix to phonedigit there doesn't enforce that anything goes  
>>>>> after ";ext=".
>>>>> see:
>>>>> extension = ";ext=" 1*phonedigit
>>>>
>>>> You are right - this is really messed up. And the same is true  
>>>> of phonedigit-hex.
>>>>
>>>> (Because of the way they are used this is only technically a  
>>>> problem with 'extension'.)
>>>>
>>>> I think a correct way of stating this would be:
>>>>
>>>>    phonedigit           = [visual-separator] DIGIT [visual- 
>>>> separator]
>>>>    phonedigit-hex       = [visual-separator] HEXDIG / "*" / "#"
>>>>                           [visual-separator]
>> Thanks for your replies, these BNF changes to 'phonedigit' and  
>> 'phone-digit-hex' do change the permitted BNF for the 'global- 
>> number-digits', 'local-number-digits' and 'extension' constructs,  
>> I just want to clarify that this change is intentional.
>>  global-number-digits = "+" *phonedigit DIGIT *phonedigit
>>  local-number-digits  =
>>      *phonedigit-hex (HEXDIG / "*" / "#")*phonedigit-hex
>> e.g. previously, a uri such as "tel:+()()1" would be allowed under  
>> the BNF, but with the change in 'phonedigit', visual-separators  
>> must be neighbored by as least one DIGIT (or  HEXDIG/"*"/"#" for  
>> local numbers).
>> Does that new limitation sound right or are there valid examples  
>> of telephone numbers where visual-separators shouldn't need to be  
>> neighbored by these symbols?
>> Thanks.
>
> _______________________________________________
> Iptel mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/iptel


_______________________________________________
Iptel mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/iptel