Two quakes in two days, no warning from ShakeAlertLA | Hacker News Reader

Two quakes in two days, no warning from ShakeAlertLA

latimes.com

25 pointsdanso17 comments

I'm a bit skeptical of the app. The city very briefly open sourced it. After people dug through the code and raised some privacy concerns, the city removed it from github.

The architecture of the app seems a bit weird to me. Your position is continuously reported to a central server to determine whether or not you are sent an alert, as opposed to filtering that client-side. I suppose it's easy to criticize when I'm not aware of some of the constraints they may have had to deal with, but it was frustrating that the app wouldn't even open unless I gave it fine-grained location permissions on my phone.

well maybe if you were affected by a shake and your phone sent your location to the govt then you could be helped? the only possible usecase for it though

But not an advertised one as far as I can tell now. So I don't believe they were doing anything like that.

Why not let the users pick their own threshold? People who are used to smaller quakes and not concerned can ignore these, and elect to receive alerts only for stronger quakes.

Thank you, this is exactly what should be configurable

No mention of MyShake [1], a California-based effort to use smartphone sensors to detect earthquakes. The potential is incredible but with actually popular apps - presumably like ShakeAlert - not contributing data it seems like an epic lost opportunity.

Phones have been able to detect earthquakes, changes in the weather, magnetic field changes, but we (the tech community) are not doing any of those things at scale. IMO all these things should be built-in to phones OS' as a kind of distributed multi-purpose sensor network. For now, we just have a smattering of not-very-popular apps that can potentially do neat things but are not centralized or popular enough to actually do them right.

[1] MyShake article: https://www.popsci.com/now-you-can-download-an-earthquake-se...

Being familiar with this, have you seen any info about how this would affect battery life?

The potential of using smartphones as massive sensor array for various purposes has always been a curiosity of mine, but most material I've read on the matter shows battery life to be a significant limiting factor.

I think battery life is quite simple to solve for, actually. A common method to work through this, especially for earthquakes, is to only capture sensor data when the phone is stationary and plugged in+charging. If the phone is moving or unplugged, no data needs to be gathered. Sure, this limits the quantity of data you can get in the short-term, but relaxes users long-term and keeps the app installed. Offering these settings as toggles also helps.

FWIW, I mostly work on barometers. New versions of Android lock down the sensor access so it cannot be done in the background; my solution is low-resolution data (15-minute intervals, rather than ideal 1-minute) and for user privacy, I can discard location data after summarizing pressure data into gridded anonymous areas.

I do this in All Clear Weather: https://play.google.com/store/apps/details?id=com.allclearwe... (US, Android)

Sensor code: https://github.com/JacobSheehy/AllClearSensorLibrary

I am also interested in gathering magnetic data, and presently also collect temperature, light sensor and humidity from phones that have it. I believe all this data could be useful for weather prediction, which is my current target.

But earthquake alerts, magnetic field/solar storm warnings, sound pollution (microphone), and various other distributed sensor networks should also be possible.

I have not noticed any battery degradation or over-use of resources in my dealings with sensors. As long as the developer takes care in the code and respects users' wishes, I think it's totally doable.

This may not be a good idea. Desensitizing the public by exposing them to alerts for what are essentially non-harmful events could mean they don't react in the face of a large incident.

I live in LA and neither quake was worth warning anyone. They were hundreds of miles away. Why bring it up a month later?

What is so hard about letting the app users decide at what point the app will trigger an alert? I suppose you give it a default setting and if the user decides that the alarms are too intrusive then the users can change them. That's my quick thought. My point is that there should not be one setting for all users there should be ways to customize it. My guess is that it would only take a small bit of code to let the users adjust the setting.

http://archive.is/sIrdZ

Possibly from the makers of Tornado Guard? https://xkcd.com/937/

It is strange that i cannot find this app in Google Playstore..

I am in another country but that should not stop me from seeing the app or installing it.

Yes, that can absolutely prevent you from seeing it if it hasn't been made available for your region.

You shouldn't have any trouble finding the Play Store page online: https://play.google.com/store/apps/details?id=com.app.shakea.... I believe the Play Store shows users any app they're looking for, even if it is not distributed in your region. The store will restrict your ability to install it based on the Dev Console settings by the app owners.

App developers are allowed to restrict by location on both platforms for legal reasons. The stores do not force developers to support regions that the developers do not / cannot support.

Issue is if i wish to travel to LA then i am prevented from using it...

Reply on news.ycombinator.com