[Sipping-emergency] Emergency-req: 13-

> Further, I believe it will be a burden on the Proxy to properly route 
> the call regardless of how many destinations there can be. Having the 
> location contained within the UA helps this. marrying some 
> cross-functional code between SIP and either Geopriv or the available L2 
> (wireless) technology to determine the UA's location will greatly help 
> this effort.
> 
> I believe jurisdictions will demand this functionality - and it will not 
> be easy.

I don't think this is a big issue. We have running code for that, at 
least for the US (using a company that has a proprietary interface 
database that offers a geo-to-jurisdiction mapping). Indeed, existing 
telematics services such as OnStar use such mapping services, so this 
seems like a manageable problem on the user end. It's much harder on the 
backend if you want to support several competing providers of such 
translation services - you essentially get back to the 
registrar-registry problem for DNS. I suspect that this is far beyond 
where we want to go in this effort.

> #15 - From I8: Call setup latency
> 
> Comment: This requirement is the first to elude to the concept of the 
> Proxy having to know what to do with UAs and their coordinates (and the 
> look-ups).
> 
> This leads into a new requirement that I haven't seen yet from section 
> 5: If the UA has its location, regardless of where the Proxy is, there 
> is a need for the (or any) Proxy to know which PSAP to route the call to 
> based on this location information.

Yes, I considered that implied in the mediated model, but I amplified this.

> 
> #17 - From 5.3, 2nd paragraph, last sentence
> 
> "Emergency numbers are generally not meant for anonymous tips. [TBD: Are 
> there any exceptions?]"
> 
> This is a viable and often used mechanism from pay phones within urban 
> cities to quitely inform the police about something the call wants them 
> to be aware of, yet have no record of who called (like calling in where 
> a drug deal occurred or is about to occur).

If we do get SIP payphones, they'd convey the same information as today: 
the identity of the payphone and its location (since there's no need for 
personal authentication), so I guess this doesn't seem to raise any new 
issues.

My TBD referred to the question as to whether there are jurisdictions 
that allow terminal selective anonymity, but I guess this question is 
unanswerable.

> #19 - observation at this point in the document
> 
> I've made a few comments about Location of the caller, and this is the 
> first real text that covers the subject. Therefore, I'd suggest that the 
> Introduction briefly go through what section each general topic will be 
> covered in (instead of having the reader read the whole ID before 
> knowing a certain topic is indeed covered).

I've added a bit of verbiage up front, but the beginning of the 
End-to-End section does point to the 'who' and 'where' sections.

> 
> #20 - A new Requirement that I haven't read here (and one that should be 
> considered for section 5.2, else section 5.4): If the UA provides its 
> location (coordinate or civil), or the first (or any) Proxy knows (or 
> learns by some other means) the location of the UA making the emergency 
> call, it MUST route the call to the appropriate ECC or IECC.

Noted.

> If you've already published this ID, I'll reproduce this list of 
> questions on the main SIPPING list for general discussion.

The I-D editor picked it up at 12:55 on Friday, so I'll create a -01 
version. It's at 
http://www.cs.columbia.edu/sip/drafts/draft-schulzrinne-sipping-emergency-req-01.pdf

Thanks again for your comments.

Henning

_______________________________________________
Sipping-emergency mailing list
Sipping-emergency@ietf.org
https://www1.ietf.org/mailman/listinfo/sipping-emergency