Re: [Ecrit] country specific emergency URNs

[Henning Schulzrinne <[e-mail]>]

I suspect, despite differences at the margins, we have general agreement on
a few points:

* Service URNs are meant to reflect reasonable, common understanding of the
service and lead to reasonable fall-back behavior if the specific service
is not supported. For example, if there's a service URN sos.fire.wildfire,
the behavior is as expected: if it doesn't exist, forwarding the call to
whoever handles sos.fire is probably right.

* We already have the 'counseling' tree for non-emergency advice services.
We seem to converge on having a non-emergency services tree (or possibly
more than one) that reflect services that yield a response (more than
talk), but are not considered emergency services.

* URNs don't encode policy (e.g., priority) or law.

* We have a light-weight process for adding URNs on an as-needed basis. I
suspect this discussion will yield a slew of new registrations that will
cover most ECRIT-using services (and many that never will), but we don't
need to strive for completeness now.

As I have mentioned, I have made an effort, now updated, to collect the
emergency and similar widely-known services for almost all of the OECD
countries at

*https://docs.google.com/presentation/d/18t1siLVrPTwyYWUm8TQK7EnwC9Y4hwkYZmEa846Pzig/edit?usp=sharing
<https://docs.google.com/presentation/d/18t1siLVrPTwyYWUm8TQK7EnwC9Y4hwkYZmEa846Pzig/edit?usp=sharing>*

If you see any mistakes or significant omissions, please let me know.

I'm hoping that the list can help avoid pointless discussions about
services that could exist, but don't actually do. Also, as mentioned, we
don't need to encode every possible widely-known service unless it is
likely to use ECRIT-style (service URN + LoST) resolution. Thus, for the
time being, we don't have to worry about, say, phone repair,
call-before-you-dig, directory services, talking clock services, weather,
traffic (511 in the US) or dial-a-prayer services, even though they may all
have short code numbers.

Henning

On Thu, Jul 13, 2017 at 2:38 AM, Christer Holmberg <
[e-mail]> wrote:

> Hi,
>
>
>
> >>RFC 5031 defines semantics for the sos sub-service, and at least my
> >>understanding is that sos is to be used for emergency services, and
> >>nothing else.
> >
> >There is no limit expressed, or implied, that sos is restricted to
> >³flashing lights² emergencies.  It¹s suggestive of emergency, but not
> >exclusively used that way.
> >The counter example was the non-emergency ambulance service.  Using
> >sos.ambulance for a non emergency ambulance service, unless a country has
> >both an emergency and a
> >non-emergency ambulance service that needs urns, is appropriate I think.
> >I don¹t think it would be wrong to create a non-emergency ambulance
> >registration, but I also don¹t think we
> >need to.
>
> I think we should have separate URNs for emergency and non-emergency
> services - no matter if the emg and non-emg services are provided in the
> same country or in different countries.
>
> >There is no MUST language at all.  Anything is acceptable, although the
> >expert review gets to offer an opinion, and take it to the list if
> >needed.
> >
> >The review is merely supposed to make sure the service is distinct from
> >an existing registration.  Registrations are cheap.
>
> Not if we are going to have list discussions about the name whenever a URN
> is to be registered. It would prolong the registration and, depending on
> who participate in such discussions, the outcome would be very
> inconsistent.
>
> So, I strongly think that if the service is an emergency service, we use
> sos. If the service is not an emergency service, we don¹t use sos.
>
> And, *URNs* are cheap, so we are not going to ³run out of them² if we use
> different ones for emergency services and non-emergency services.
>
>
> >>If we start using non-sos for emergency services, and sos for
> >>non-emergency services, I think we will end up in a messŠ
> >
> >The above example is how I think we don¹t get into a mess.  I don¹t think
> >there is a good reason to create an ambulance service that isn¹t in the
> >sos tree unless
> >we have an incidence of two services and we need to differentiate.  Even
> >if the ambulance service in a country is not considered an emergency
> >service, using sos.ambulance for it is fine in my opinion.
>
> I have a different opinion :)
>
> Regards,
>
> Christer
>
>
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Ecrit mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/ecrit
>