- From: Sean Lin <selin@mozilla.com>
- Date: Fri, 10 Apr 2015 18:01:32 +0800
- To: Paul Higgs <paul.higgs@ericsson.com>
- Cc: TV Control API Community Group <public-tvapi@w3.org>
- Message-ID: <CAO=RbvmKYmkH5ggjwCsg1-Mw8M8hphnhX=zYFhTU9Ym6eRijAA@mail.gmail.com>
Hi Paul, Alright. If you still insist, we may add the following method to the event. sequence<DOMString> getRegions(); But the returned array is not guaranteed to be non-empty because not all emergency warning systems would expose sufficient geographical info about the alert to the endpoint. Sean On Thu, Apr 9, 2015 at 6:35 PM, Paul Higgs <paul.higgs@ericsson.com> wrote: > Hi Sean > > > > If that is how you expect **all** emergency warning systems to operate, > then your proposal is appropriate. > > I do not, however, believe your high-level theory is correct > > > > Paul > > > > *From:* Sean Lin [mailto:selin@mozilla.com] > *Sent:* Thursday, April 09, 2015 12:16 AM > > *To:* Paul Higgs > *Cc:* TV Control API Community Group > *Subject:* Re: tvapi-ACTION-26: Contact sean and add emergency alert > requirements to the spec > > > > Hi Paul, > > > > I still don’t believe we can overload the severityLevel in this way. > > The same content is broadcast to all receivers and the transport stream > contains alert events for all regions where the broadcast could be > received. I, living in North Fulton county, Georgia USA and watching the > regional TV channel do not want severe storm warning alerts for the coastal > parts of the state (some 300 miles away) interrupting my viewing. > > > > It looks there might be some differences between our expectations about > how the emergency alert system works. The following is my high-level theory > and hope it would make us on the same page. (Please correct it if I'm > wrong.) > > > > * For a given emergency situation, the authority may only broadcast alerts > with different levels to affected regions. So if a hurricane is about to > hit NYC, an alert of hurricane warning may be broadcasted to NYC and its > neighboring areas; whereas another alert with lower threat level, like > watch or information, would be broadcasted to farther and less affected > areas. And for other places where no effect is expected, such as Chicago, > no alert is sent there at all. > > > > * The regional service provider, such as Comcast or Time Warner, may also > help only forward the alerts to the right places because they appear to > have more accurate geographical knowledge about where their service > endpoints are. > > > > * Furthermore, even if it the alert signal really happens to reach the > endpoint not in the target area (assuming the signal carries some info > which can be used to identify the target area), it may be more reasonable > to the underlying platform (TV stack, OS, etc.) to filter out those alerts, > instead of exposing it to web contents and relying on their ability to > gracefully handle it with their own geographical data collected via some > other way. > > > > * Besides, the description is expected to be human readable info, which > may also mention some general region info about this alert. For example, it > might look like "Hurricane XXX will be coming this Thursday and affect the > following areas: YYY county, ZZZ county, ...". So there still seems some > ways not to confuse end users. > > > > Overall, the key point is: as long as the platform fires an event of > emergency alert to the web content, we believe the alert is supposed to be > there. In the case you mentioned, it's less likely a severe storm warning > for areas in 300 miles away would be sent to your place if it's not in the > target area of the alert. (Would a regional alert really consume the > nationwide emergency channel to broadcast?) And if the alert source really > broadcasts them outside the target area, there seems more ways to stop > propagating it in the middle (service provider, platform, etc) before > eventually exposing it to the web content. So in these cases, I'm still not > convinced extra attributes for regional info need to be added to our event. > > > > Thoughts?! > > > > Sean > > > > > > On Wed, Apr 8, 2015 at 6:42 PM, Paul Higgs <paul.higgs@ericsson.com> > wrote: > > HI Sean > > > > I still don’t believe we can overload the severityLevel in this way. > > The same content is broadcast to all receivers and the transport stream > contains alert events for all regions where the broadcast could be > received. I, living in North Fulton county, Georgia USA and watching the > regional TV channel do not want severe storm warning alerts for the coastal > parts of the state (some 300 miles away) interrupting my viewing. > > > > Paul > > > > > > *From:* Sean Lin [mailto:selin@mozilla.com] > *Sent:* Wednesday, April 08, 2015 5:38 AM > > > *To:* Paul Higgs > *Cc:* TV Control API Community Group > *Subject:* Re: tvapi-ACTION-26: Contact sean and add emergency alert > requirements to the spec > > > > Hi Paul, > > > > I believe regional information is quite important and certainly a > requirement in the US (there may be different storm warning/alert levels > for neighboring counties/regions) > > > > I believe attribute severityLevel can cover this. For alerts sent to > different regions, they could be in different severity levels even when > associated with the same event. The idea of this attribute was originated > from [1], which also touches similar scenarios. > > > > PH> I don’t think it can. severityLevel could be “test”, “information”, > “watch”, “warning”, “threat”… while location would be a geographic area. > Quite often in the US you will see a “Tornado warning” issued for a few > neighboring counties and “Tornado watch” issued for those neighboring them. > > > > SL> In this case I believe we may utilize "severity level" and the new > attribute "type of emergency" we plan to add. "Type" may help identify the > nature of the emergency like Tornado, Earthquake, etc; and "severity level" > is used to distinguish the different threat levels across multiple regions. > > > > Thoughts?! > > > > Sean > > > > > > On Tue, Apr 7, 2015 at 8:36 PM, Paul Higgs <paul.higgs@ericsson.com> > wrote: > > Hi Sean > > > > I believe regional information is quite important and certainly a > requirement in the US (there may be different storm warning/alert levels > for neighboring counties/regions) > > > > I believe attribute severityLevel can cover this. For alerts sent to > different regions, they could be in different severity levels even when > associated with the same event. The idea of this attribute was originated > from [1], which also touches similar scenarios. > > > > PH> I don’t think it can. severityLevel could be “test”, “information”, > “watch”, “warning”, “threat”… while location would be a geographic area. > Quite often in the US you will see a “Tornado warning” issued for a few > neighboring counties and “Tornado watch” issued for those neighboring them. > > > > Paul > > > > *From:* Sean Lin [mailto:selin@mozilla.com] > *Sent:* Tuesday, April 07, 2015 5:56 AM > > > *To:* Paul Higgs > *Cc:* TV Control API Community Group > *Subject:* Re: tvapi-ACTION-26: Contact sean and add emergency alert > requirements to the spec > > > > Hi Paul, > > > > I am not sure what to do with the Description if it is provided in an > event with an application URL. Perhaps that behavior is implementation > specific? > > > > Yup, it could be. :) > > > > I believe regional information is quite important and certainly a > requirement in the US (there may be different storm warning/alert levels > for neighboring counties/regions) > > > > I believe attribute severityLevel can cover this. For alerts sent to > different regions, they could be in different severity levels even when > associated with the same event. The idea of this attribute was originated > from [1], which also touches similar scenarios. > > > > Thanks, > > Sean > > > > [1] > http://www.w3.org/community/websignage/wiki/Web-based_Signage_Player_Emergency_Profile#Classification_of_severity_level_for_presentation > > > > > > On Mon, Apr 6, 2015 at 7:44 PM, Paul Higgs <paul.higgs@ericsson.com> > wrote: > > Hi Sean > > > > On type of emergency – great > > > > I am not sure what to do with the Description if it is provided in an > event with an application URL. Perhaps that behavior is implementation > specific? > > > > I believe regional information is quite important and certainly a > requirement in the US (there may be different storm warning/alert levels > for neighboring counties/regions) > > > > Paul > > > > *From:* Sean Lin [mailto:selin@mozilla.com] > *Sent:* Thursday, April 02, 2015 6:02 AM > *To:* Paul Higgs > > > *Subject:* Re: tvapi-ACTION-26: Contact sean and add emergency alert > requirements to the spec > > > > Hi Paul, > > > > Thanks for the feedback. Please see the comment inline. > > > > · DOMString? severityLevel – I agree with the “localized > definition” of this, but don’t we also need some attribute that describes > the nature of the emergency, i.e. “Earthquake”, “Child Adbuction” > > Yeah, we may add this. > > · DOMString? Description – is this intended to be human readable? > What is the code that receives this event supposed to do with this > message/information. > > It's intended to be human readable, and may be directly used as the > displayable message. > > · Do we need some information on where the alert applies to, or > do we only expect it to be only delivered in-band or out-of-band to > affected regions? > > So far our emergency alert requirements haven't addressed this. Do we have > a use case which really needs this? Or we may consider to leave it for now > and open for enhancement if necessary. > > Any thoughts?! > > Sean > > > > On Mon, Mar 30, 2015 at 9:16 PM, Paul Higgs <paul.higgs@ericsson.com> > wrote: > > Hi Sean > > > > Thanks, the changes you provide are a good start to having a well defined > “emergency alert” interface, but I think some more details are required. > > > > · DOMString? severityLevel – I agree with the “localized > definition” of this, but don’t we also need some attribute that describes > the nature of the emergency, i.e. “Earthquake”, “Child Adbuction” > > · DOMString? Description – is this intended to be human readable? > What is the code that receives this event supposed to do with this > message/information. > > · Do we need some information on where the alert applies to, or > do we only expect it to be only delivered in-band or out-of-band to > affected regions? > > > > I would suggest that, if possible, people read the ATIS-0800010 “Emergency > Alert Provisioning Specification” [1] which covers North American > requirements and ATIS-0800012 “IPTV Emergency Alert Metadata Specification” > [2] which covers the implementation of those requirements. > > > > Some other reading would be > > [3] EB Docket 04-296, First Report and Order and Further Notice of > Proposed Rulemaking, Review of the Emergency Alert System, FCC 05-191, > November 10, 2005. > > [4] Code of Federal Regulations (CFR) Title 47 Part 11--Emergency Alert > System, October 1, 2005. > > [5] EB Docket 04-296, Second Report and Order and Further Notice of > Proposed Rulemaking, Review of the Emergency Alert System, FCC 07-109; > Released July 12, 2007. > > [6] Federal Information Processing Standards Publication 6-4, Counties and > Equivalent Entities of the United States, Its Possessions, and Associated > Areas, 31 August 1990, with editorial corrections January 2005. > > [7] OASIS Standard CAP-V1.1, Common Alerting Protocol (CAP), October 2005. > Copyright © OASIS Open 2003. All Rights Reserved. Based in part on prior > work contributed by the Common Alerting Protocol Working Group, copyright > 2002-2003 Art Botterell for the Common Alerting Protocol Working Group. > > [8] SCTE 18 2013, Emergency Alert Messaging for Cable. > > > > Paul > > > > [1] https://www.atis.org/docstore/product.aspx?id=22927 > > [2] https://www.atis.org/docstore/product.aspx?id=22946 > > [3] available from the Federal Communications Commission. < > http://hraunfoss.fcc.gov/edocs_public/ > > > [4] available from the Electronic Code of Federal Regulations. < > http://ecfr.gpoaccess.gov/ > > > [5] available from the Electronic Code of Federal Regulations. < > http://ecfr.gpoaccess.gov/ > > > [6] available from the Information Technology Laboratory. < > http://www.itl.nist.gov/fipspubs/fip6-4.htm > > > [7] available from the Organization for the Advancement of Structured > Information Standards (OASIS). < http://www.oasis-open.org/ > > > [8] available from the Society for Cable Telecommunications Engineers > (SCTE). < http://www.scte.org/FileDownload.aspx?A=3512 > > > > > > > *From:* Sean Lin [mailto:selin@mozilla.com] > *Sent:* Friday, March 27, 2015 6:09 AM > *To:* Paul Higgs > *Cc:* TV Control API Community Group > > > *Subject:* Re: tvapi-ACTION-26: Contact sean and add emergency alert > requirements to the spec > > > > Hi, > > > > I've reflected the adjustments to our spec [1] for now. Please feel free > to share concerns if you find something uncovered. > > > > Thanks, > > Sean > > > > [1] http://w3c.github.io/tvapi/spec/#tvemergencyalertedevent-interface > > > > > > On Wed, Mar 11, 2015 at 5:39 PM, Sean Lin <selin@mozilla.com> wrote: > > Hi Paul, > > > > Thanks for the feedback. I'm thinking to adjust the event as below. > > > > 1. What event/parameters do we signal to inform that the emergency > announcement is over? > > It appears the attribute 'severityLevel' could be nullable. So it implies > the emergency is over when the attribute is null. > > > > 2. The event only signals a channel could be switched to, however, > it could also be possible that the alert information is available at a URL > (textual “storm warning” etc). The requirements for all national bodies > should be investigated to ensure global applicability of this API. > > We may add an extra nullable attribute 'url' for now to provide this > flexibility. Yet as you mentioned, we'll need to ensure global > applicability someday. > > > > Thoughts?! > > > > Sean > > > > > > On Tue, Mar 3, 2015 at 8:46 PM, Paul Higgs <paul.higgs@ericsson.com> > wrote: > > Hi Sean > > > > This event based approach for signaling of an emergency probably the right > direction to proceed, just a couple of items to consider > > 1. What event/parameters do we signal to inform that the emergency > announcement is over? > > 2. The event only signals a channel could be switched to, however, > it could also be possible that the alert information is available at a URL > (textual “storm warning” etc). The requirements for all national bodies > should be investigated to ensure global applicability of this API. > > > > > > Paul > > > > *From:* Sean Lin [mailto:selin@mozilla.com] > *Sent:* Thursday, February 26, 2015 3:27 AM > *To:* TV Control API Community Group > *Subject:* Re: tvapi-ACTION-26: Contact sean and add emergency alert > requirements to the spec > > > > Hi all, > > > > I slightly revised the spec and added emergency alert to it. Please feel > free to share any comments on it. > > > > Thanks and best regards, > > Sean > > > > > > On Tue, Feb 17, 2015 at 10:47 PM, TV Control API Community Group Issue > Tracker <sysbot+tracker@w3.org> wrote: > > tvapi-ACTION-26: Contact sean and add emergency alert requirements to the > spec > > http://www.w3.org/community/tvapi/track/actions/26 > > Assigned to: Bin Hu > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > > > > > > -- > > Sean Lin > > Mozilla Taiwan > > selin@mozilla.com > > > -- Sean Lin Mozilla Taiwan selin@mozilla.com
Received on Friday, 10 April 2015 10:02:07 UTC