Re: [sipcore] Draft new version: SIP Push -26

[Ben Campbell <[e-mail]>]

> On Feb 28, 2019, at 1:41 PM, Christer Holmberg <[e-mail]> wrote:
> 
> Hi Ben,
> 
> Thanks for your comments! Please see inline.
> 
>>   §4.1:
>> 
>>   - "For privacy and security reasons, a UA MUST NOT insert the SIP URI
>>   parameters (except for the pn-purr parameter)”
>> 
>>   The “except for the pn-purr parameter”part is part of the normative requirement. Please don’t bury it in a parenthetical phrase.
>> 
>>   - "in non-REGISTER request"
>> 
>>   Plural Disagreement
> 
> Can you suggest better wording?

either “in non-REGISTER requests” or “in a non-REGISTER request”.

> 
> ---
> 
>>   §4.1.4:
>>   - "The indicator value indicates
>>   the minimum time (given in seconds), prior to the binding expiration
>>   when the UA MUST send the REGISTER request.”
>> 
>>   That MUST is a statement of fact. (The normative requirement was in the previous sentence.)
> 
> I suggest s/MUST/needs to.

WFM.

> 
>>   - "request that
>>   a push notification is sent”
>> 
>>   s/is/be  (note that this pattern occurs many times in the draft.)
> 
> As I indicated to Jean, the wording is the outcome of the previous discussions associated with Ekr's IESG review.

I’d be highly surprised if ekr objected to “be”.  it is more grammatically correct than “is”. The subjunctive “be" form is indicated by the fact that we talk about a request for something to happen, not a statement that something is true.

I’m not going to block things over it, but I would bet a beer that the RFC editor changes it. :-)

[…]

> 
>>   - "A binding expiration interval MUST be considered too short if the
>>   binding would expire before the proxy would request that a push
>>   notification is sent to the UA”
>> 
>>   I’m not sure what to make of that sentence. Presumably the proxy will request the PN when it needs to
>>  do so. Perhaps the “would” in “proxy would request” should be “can”? (That is, the interval is so short the
>>  proxy is unable to request a PN in time?)
> 
>>   §5.6.1.2 and §5.6.2: Are the lists of procedure steps supposed to have bullets or numbers?
> 
> No __
> 
> I think the procedures are very clear thanks to the paragraph separation.

Do you mean the indentation? Traditionally indentation has been used in RFCs to show parenthetical explanations or sidebars. Also, the use of the indented list removes the white space between the paragraphs, which makes them harder to read.

Bullets would make them easier to read, and would be more consistent with similar procedure lists earlier in the draft.

[…]