Android version of popular Dark Sky weather app shuts down

The web version got an extension, but we don't know for how long.

Read the whole story

 

[Shantara]

Checking the weather app in the iOS 14 beta with Dark Sky integration, it’s apparently drizzling for the next hour, which is a surprise to me, as I’m currently sitting outside in the sun, so I’m not sure Android subscribers have really lost much.

I honestly don’t get the whole “micro weather” thing, though, so I’m admittedly not the target audience, even if it did work.

Serious meteorologists don’t generally think much of Dark Sky. Their forte is predicting exactly when rain will arrive in your area, using current and past radar data to predict future radar data. Their algorithms are more like ML image prediction and have little to no actual weather modeling.

That said, YMMV, and for some the mileage is excellent. It does an eerily good job at predicting the arrival of almost any precipitation along a weather front (think storms that sweep across a region). Similarly it does a really poor job of predicting convective weather (think rain clouds that boil up in one spot). Bodies of water and steep topography can also throw off image predictive algorithms. Whether Dark Sky works well for you is strongly dependent on how much their approach works for your conditions.

And there’s a grand tradition here. What Dark Sky does is an updated version of an early, generally useful weather prediction tech: ask someone a few dozen miles to the west of you “what’s your weather like right now?”

For me, Dark Sky's weather forecasting has been hit and miss, but where they excelled was having an enormous amount of aggregated historical weather data from multiple sources accessible via a simple, well documented and easy to implement API. This was a service which to my knowledge had nothing even remotely comparable from other companies.

 

[Jeff S]

Back to basics

That's not forecasting, that's telling me the recent past.

 

[mikiev]

use of color tags

Red is okay, blue is barely legible. At least you didn't use yellow - which is completely illegible for me.

 

[Coleman]

So what are people using for alternatives? I've been hopping between Windy and Today Weather, sort of preferring Windy but not so much that I wouldn't consider others.

I just uninstalled Today Weather, because they want a subscription to see _radar_. I don't mind subscribing to a good product, I paid for Dark Sky. I'll pay when/if I find a good replacement. But asking to pay for something so basic...nope, don't wanna support that product.

 

[GaryFisken]

In Europe, the site/app Yr is incredibly accurate. The times I've traveled to the U.S. I've used it as well and was satisfied. Behind it is Norwegian Meteorological Institute and when it comes to the weather, I trust the Norwegians...

https://www.yr.no/en

Available for Android
Available for iOS

 

[pocal]

I've tried a bunch of alternatives and not really loved any of them, however one I like is Geometric Weather, which:

- Is free
- Has no ads
- Is open source
- Allows you to edit the UI so as to to hide the stuff you don't care about

 

[numerobis]

Buying a service and shutting down support for the competitor seems anticompetitive as fuck. How do they get away with it?

Because they are integrating it directly into iOS’s weather app. I don’t see Google making sure Nest products work with HomeKit.

Google acquired Nest before Apple released HomeKit. Rather different situation.

 

[Coleman]

Are you saying they get the data from people using their app?

I hope not for the local apple users in the Netherlands, Appl eonly has a 30% marketshare here down from 44% just 5 years ago.

Yeah, that's my understanding of how Dark Sky worked so well. It used barometric (and other?) data from users (you could opt out, as I recall it asked you if you wanted to provide it...I've had it installed for so long I don't remember) to give the hyperlocal forecasts.

 

[numerobis]

Back to basics

That's not forecasting, that's telling me the recent past.

Train a webcam on a forecast stone 20 miles upwind of you. Modern technology has a solution to every problem!

 

[woodelf]

I honestly don’t get the whole “micro weather” thing, though, so I’m admittedly not the target audience, even if it did work.

DarkSky has rarely worked as long-term predictive for me. But I don’t use it for that. I use it for “okay, is it going to start raining if I go [activity]”. And for that, it’s usually pretty good an occasionally phenomenal. Especially in a time when I can’t go sit in a restaurant and have another beer or two to wait out some rain, being able to know “oh, I should call in now and go, because I’ll be walking back in the rain with a paper bag if I wait” is a massive help.

When I lived in DC, it was eerily accurate most of the time. Like, “hey, let’s go into that museum right now because it’s about to start raining, and getting inside 30 second before it starts pouring” accurate.

It’s good to know if you’ll need an umbrella when you go to work. It’s great to know if you need it right now.

Yeah, I hear that from so many people. I can't help but wonder whether my experiences are that different, or if this is selective memory where the users remember the times it was right and forget when it was wrong? I used to use WeatherBug's hourly forecast exclusively for that purpose. It was hit-or-miss. Then I went back and forth between WeatherBug and Dark Sky, randomly, not comparing their results, for a while, and felt like I was disappointed by inaccuracies from both of them about equally frequently—they were probably both about 80% accurate. Then I started using Dark Sky primarily, but the rain warnings aren't stunningly accurate: about half the time when it warns me it's about to rain, it doesn't, and about half the time it rains before Dark Sky warns me it's going to.

 

[Coleman]

I've often read (at Ars even?) that the forecasting models outside the US are much better than what NOAA/NWS uses. Does anyone know if there are any apps that use those/does any place apply said models to US weather?

 

[lcubed11]

i'm guessing all the programmers working on the android version and API have departed for greener pastures.

no programmers == no support

 

[Deleted member 250988]

Checking the weather app in the iOS 14 beta with Dark Sky integration, it’s apparently drizzling for the next hour, which is a surprise to me, as I’m currently sitting outside in the sun, so I’m not sure Android subscribers have really lost much.

I honestly don’t get the whole “micro weather” thing, though, so I’m admittedly not the target audience, even if it did work.

From what I understand hyperlocal weather requires hyperlocal data. It's good where it's got a lot of data points. Otherwise, I believe it falls back to one of the standard weather feeds. I believe it got its reputation for accuracy in bigger cities where it's got a lot of inputs.

Are you saying they get the data from people using their app?

I hope not for the local apple users in the Netherlands, Appl eonly has a 30% marketshare here down from 44% just 5 years ago.

No, typically they get the info from amateur weather stations that call into an API. WeatherUnderground has a similar system, and you can even go look at feeds from certain weather stations.

Thanks, when I read that comment I was thinking about how apps for mobile speeding cameras over here work. Users of the app that see one report it and the server pushes the report to other users to show a warning when appropriate.

Then when a user gets the warning and sees the camera gone he can report that and the warning is removed.

Works like a charm in my car since it's a single tap to report when the app is running, but I was wondering how you get people to get out their phones to report weather :/ So they don't ... that makes sense.

 

[japanesehamantashen]

This was a real hard slap in the face to Android users who paid the subscription. Still pretty mad about it.

It’s $3/year. You’re being a bit over dramatic.

Which was refunded.

You're forgetting that iOS users got to buy Dark Sky for a one-time fee of $4, while android users had to pay a subscription if $3 per year. Each android user has done a lot more to fund the app over the years due to this discriminatory pricing, and they still get cut off.

Okay, so they're refunding the final $3 payment. Anyone on Android who has had the app for more than 2 years will still have paid more than iOS users. Being made to pay more only to be cut off anyway is a slap in the face.

 

[woodelf]

Checking the weather app in the iOS 14 beta with Dark Sky integration, it’s apparently drizzling for the next hour, which is a surprise to me, as I’m currently sitting outside in the sun, so I’m not sure Android subscribers have really lost much.

I honestly don’t get the whole “micro weather” thing, though, so I’m admittedly not the target audience, even if it did work.

From what I understand hyperlocal weather requires hyperlocal data. It's good where it's got a lot of data points. Otherwise, I believe it falls back to one of the standard weather feeds. I believe it got its reputation for accuracy in bigger cities where it's got a lot of inputs.

Are you saying they get the data from people using their app?

I hope not for the local apple users in the Netherlands, Appl eonly has a 30% marketshare here down from 44% just 5 years ago.

No, typically they get the info from amateur weather stations that call into an API. WeatherUnderground has a similar system, and you can even go look at feeds from certain weather stations.

Thanks, when I read that comment I was thinking about how apps for mobile speeding cameras over here work. Users of the app that see one report it and the server pushes the report to other users to show a warning when appropriate.

Then when a user gets the warning and sees the camera gone he can report that and the warning is removed.

Works like a charm in my car since it's a single tap to report when the app is running, but I was wondering how you get people to get out their phones to report weather :/ So they don't ... that makes sense.

Dark Sky may not rely on user reporting, but it definitely encourages it and makes it easy to do so. I always figured that the lack of people in my city using it was why its forecasts weren't any more accurate than competitors, in marked contrast to the experiences of people I know who have compared Dark Sky and other weather apps in other cities (and found Dark Sky noticeably more accurate, unlike my experiences). Particularly if only, say, 1 in a hundred users actually contributes to the data, you would need a lot of users for the user data to make a difference.

 

[barich]

This was a real hard slap in the face to Android users who paid the subscription. Still pretty mad about it.

It’s $3/year. You’re being a bit over dramatic.

We got refunded. The point is that we were financially supporting them and their app (adding value to the company!) and then they turned around and threw us under the bus by selling themselves to Apple.

Would the good sir like a pearl necklace and a fainting couch?

They offered a service for which you paid. Not 'invested in' and marginally 'added value to the company'. This wasn't a crowdfunding endeavor. They simply got bought and you don't like it. I fail to see the problem here.

Hmmmm... maybe I should find a way to get an anti-trust suit going for Colgate Palmolive buying out Tom's of Maine and taking away my beloved Cinnamint toothpaste? How dare they!

When did I say anything about an antitrust suit? I'm just pissed off, as you have every right to be about your toothpaste.

 

[barich]

To buy a popular product and make it an exclusive does feel terribly anti-competitive.

It would be like Walmart purchasing the best selling brands and then making them Walmart exclusive. I don't think that would go over well.

It feels like a very fuzzy line when I wouldn't complain if Apple *developed* a popular apple-only app... but still.

Are we now claiming that consoles don't pay for exclusives all the damn time?

Remind me when a game became, say, a PS4 exclusive *after* it was released and summarily stopped working on your Xbox One.

 

[aexcorp]

This was a real hard slap in the face to Android users who paid the subscription. Still pretty mad about it.

Funny how I was assured that Google buying AR glasses firm North and shutting it down for existing users was perfectly fine.

How is this different?

North was going to go bankrupt due to unreasonable product design.
All of the glasses had to be custom fitted, raising the price significantly and requiring a visit to a store (which can't really be done in a pandemic).

Dark Sky had paying users and its existing product made money.

North was going nowhere fast, and their own actions demonstrate this. They sold their entire IP portfolio to Facebook before Google bought them. See https://patents.google.com/patent/US9600030B2 as an exemple, but feel free to look up their other patents and who they were assigned to in Jan/Feb 2020 before Google bought them. So Google basically got their staff (who chose to stay...) and the knowledge and experience they have. This is not what a company that believes in its own success and growth would ever do.

Any comparison with Dark Sky (arguably the most used weather API on mobile and possibly beyond) is ridiculous and disingenuous.

 

[BullBearMS]

To buy a popular product and make it an exclusive does feel terribly anti-competitive.

It would be like Walmart purchasing the best selling brands and then making them Walmart exclusive. I don't think that would go over well.

It feels like a very fuzzy line when I wouldn't complain if Apple *developed* a popular apple-only app... but still.

Are we now claiming that consoles don't pay for exclusives all the damn time?

Or the Epic store for PC games?

Heck, I remember the outrage/hate among musicians when Apple bought the company which made Logic Pro - in 2002 - and made it Mac-only.

https://en.wikipedia.org/wiki/Emagic

On July 1, 2002 Emagic was bought by Apple Computer. Emagic's Windows-based product offerings were discontinued on September 30, 2002.

How does Good Old Games get their app store onto the XBox?

 

[Matthew J.]

This was a real hard slap in the face to Android users who paid the subscription. Still pretty mad about it.

Yeah, what a bunch of dicks.

 

[Rabid Badger]

I loved Weather Underground until they sold out and the amazing app turned to shit and half the features disappeared or broke overnight.

I settled on Dark Sky as a replacement and then Apple screwed that up.

Now I'm using WeaWow (silly name, I know) and loving it so far.

 

[Richard Cranium]

Why is this considered tolerable behaviour?

Why isn't it?

What specifically are you objecting to?

Apple buying Dark Sky? What, specifically, is wrong with one company buying another company?

Apple deciding not to support a competitor's platform?

Pro-Tip: If you replace "Apple" and "Dark Sky" with two other companies, and it doesn't bother you, then you're bullshitting.

Or are you objecting to Apple giving you three months notice PLUS AN EXTRA MONTH before shutting down Dark Sky?

Or that they refunded any existing purchases?

Look, I get that lots of people don't like this. I just don't see how anyone can say "this is intolerable!!" when it's how all business operates.

I have to agree. Maybe it’s because I grew up as a console gamer who could only ever afford one console as a kid, but I lost count of how many 3rd party companies were bought by a console manufacturer and had their games go exclusively to that console and I was left not being able to play some of my favorite games. This is no different.

 

[OccasionallyLeftHanded]

Checking the weather app in the iOS 14 beta with Dark Sky integration, it’s apparently drizzling for the next hour, which is a surprise to me, as I’m currently sitting outside in the sun, so I’m not sure Android subscribers have really lost much.

I honestly don’t get the whole “micro weather” thing, though, so I’m admittedly not the target audience, even if it did work.

From what I understand hyperlocal weather requires hyperlocal data. It's good where it's got a lot of data points. Otherwise, I believe it falls back to one of the standard weather feeds. I believe it got its reputation for accuracy in bigger cities where it's got a lot of inputs.

Agreed. I like the UI on Dark Sky, but it seems to think we have heavy fog almost every morning here, when it's 75 degrees and sunny. Maybe the microweather is correct the next block over...

 

[Warrington Ocelot]

I've often read (at Ars even?) that the forecasting models outside the US are much better than what NOAA/NWS uses. Does anyone know if there are any apps that use those/does any place apply said models to US weather?

That would be the National Weather Service. They use every model they can access -- their own systems, the European model (ECMWF), Canadian models, and a bunch of specialized local products. Every model has strengths and weaknesses, so the NWS uses a lot of them, and combines the model outputs with their forecasters' knowledge and experience to produce a forecast.

 

[RocketFeathers]

Someone on reddit pointed out to me some time ago that most US-based apps get data from NOAA, which runs weather.gov. After you enter your location, the URL includes your location, just bookmark that in your favorite web browser and voila, don't even need an app.

I use weawow on Android for their large widget.

Edit, here is O'Hare airport ...
https://forecast.weather.gov/MapClick.p ... 5999999997

 

[Chuckstar]

Someone on reddit pointed out to me some time ago that most US-based apps get data from NOAA, which runs weather.gov. After you enter your location, the URL includes your location, just bookmark that in your favorite web browser and voila, don't even need an app.

I use weawow on Android for their large widget.

Edit, here is O'Hare airport ...
https://forecast.weather.gov/MapClick.p ... 5999999997

Dark Sky actually predicts minute-by-minute precipitation for your exact location for the next hour, not just generalized forecasts like the NOAA site.

You can really only see the complete version of the interface when there is actually precipitation forecast in the next hour. As I post this, there is rain coming in NYC, so following this link (for now) will show how the short-term, hyper-local precipitation forecast works.

Dark Sky forecast for City Hall, Manhattan:

https://darksky.net/forecast/40.7127,-74.006/us12/en

People might not strictly "need" Dark Sky's level of forecast, but the point is that the NOAA site is not really a replacement for Dark Sky's hyperlocal precipitation forecast.

 

[droidmonkey]

Shadow weather, pass it on. Dev is responsive and is making a superior product.

 

[Nick Andros]

So what are people using for alternatives? I've been hopping between Windy and Today Weather, sort of preferring Windy but not so much that I wouldn't consider others.

I liked DarkSky primarily because I run every morning and it has bern great for extremely short-term predictions about inclement weather. I still have it on my iPad, but when I'm out, I only have my Android phone.

After doing a bit of research, I put NOAA Weather Unofficial on my phone, and I'm liking it so far.

 

[Sasparilla]

Have been using a Google Weather shortcut since this started. Doesn't have the next 10 mins for precipitation but has hourly, current etc. and seems to do a nice job.

https://www.androidauthority.com/google ... pp-965754/

I am mourning the loss of Dark Sky as well. I'm glad they kept it going longer than the originally planned, but its a bummer. I was hoping Apple would come around and keep it on Android.

 

[golemB]

I, too, would like to release my app to the widest, largest, greatest, most numerous amount of users, which is why I'm releasing exclusively on the Atari 2600.

Keep in mind the comparison is all iOS users vs whatever percent of Android users subscribe. The comparison isn't total iOS potential subscribers vs Android potential subscribers. I'd bet there are many more iOS users than the number of Android users that subscribed to Dark Sky. They will definitely be reaching many more people being built into iOS.

No, that's ridiculously misleading. They'd reach more people if they served all iOS users PLUS Android subscribers.

Heck, it would give Apple more money - for the price of continuing to pay one Android frontend developer on top of the backend developers who write the APIs (or heck, just have their iOS devs use a portable framework), they'd keep raking in $3 subscriptions for Android. Far more profitable than maintaining the Windows version of Safari was. It's not like they'd have to go out of their way to keep providing an Android app that's already written. They're going out of their way to lock out Android users and API clients.

The only reason for Apple to stop serving Android users and API clients is pure abuse of monopoly power. Hurt your software customers. Eliminate their market choices to push some of them to buy your hardware.

 

[Snarky Robot]

The only reason for Apple to stop serving Android users and API clients is pure abuse of monopoly power. Hurt your software customers. Eliminate their market choices to push some of them to buy your hardware.

Yeah, who’s ever heard of a company buying another company to improve their offerings. It’s not like Google has ever bought a company, right? Never forced customers into their ecosystem, despite saying they wouldn’t? Because that would a whole nest of problems for people screaming monopoly. No two Waze about it. Folks, one company buying another to improve their products at the expense of others isn’t monopoly abuse. It’s especially ridiculous when you’re talking about the minority market share holder.

You don’t have to like it. Feel free to complain. But please stop using MONOPOLY POWER as a rallying cry.

Apple doesn’t have a smartphone monopoly. DarkSky doesn’t have a monopoly. It’s an app you like, perhaps. And it sucks that all the benefits won’t be for you. But...that’s not abuse of monopoly. As I said before, Google isn’t going to go under because of this. There will be INCREASED competition in the weather app and API markets. And some customers will lose access.

The likelihood of anyone switching from Android to iOS over a weather app whose API will be available on Android for over a year is..basically zero. And if someone likes the app enough to switch platforms...that’s what improving your products is supposed to do! That improvement coming from acquisitions is...normal.

So please, if you’re going to complain, try to be just wrought, and not overwrought.

 

[newen]

It's about time to break up big tech monopolies.

 

[mikepers]

I really liked the Darksky app. Here on Long Island I assume they had good data sources because the predictions were really accurate.

I'm not happy that it got bought out and shut down but stuff happens.

I've been looking at alternatives on the app store.

First one I'm trying is "MyRadar". It hit the nail on the head tonight with regards to predicting when it would start raining. It has a lot of options. I paid the 4 bucks to remove the adds. It has a few other payable upgrades which I think I might spring for as well.

 

[spindizzy]

I, too, would like to release my app to the widest, largest, greatest, most numerous amount of users, which is why I'm releasing exclusively on the Atari 2600.

Keep in mind the comparison is all iOS users vs whatever percent of Android users subscribe. The comparison isn't total iOS potential subscribers vs Android potential subscribers. I'd bet there are many more iOS users than the number of Android users that subscribed to Dark Sky. They will definitely be reaching many more people being built into iOS.

Sure but that assumes that non-US users want to use inaccurate weather data.

Here in Australia I'll stick with something that sources our local Bureau of Meteorology data as it isn't woefully inaccurate, unlike this.

 

[f1vefour]

The best weather app is exclusively on Android: Flowx

I have no ties to the app but have used it for quite some time and want it to succeed. Definitely check it out.

Flowx and Dark Sky are two totally different apps, one has automated local rain notifications with no user interaction required (Dark Sky) whereas the other is simply a good weather forecast app (Flowx)

 

[coolblue2000]

I would be annoyed if I were losing access to Dark Sky. For me, it's the killer app for walking the dog. It's not perfect, but does a reasonable job of giving me insight into whether I would be better off (rain-wise) taking the dog out right now or waiting 10-15 minutes. For other times I may need to go outside, that hyperlocal insight is less critical for me.

I just take a coat or umbrella like in the old days if it is at all likely to rain... It works for me and Apple can't remove that functionality from my life...

 

[warr10r_za]

Is the Windy.com app not a good replacement?

 

[PrimalxConvoy]

I use "Simple Weather and Clock Widget" in my Android devices:

- https://play.google.com/store/apps/deta ... er.weather

 

[cheetah-spottycat]

Apple does it all the time. Good products get bought, and then the versions for all competing operating systems (Windows, Android, Linux, etc) get immediately axed. Eventually the acquired company gets absorbed, and sometimes all that remains is a backend service, an API or a stray pull down menu in another existing product.

Check out where all their "innovations" were actually scooped up from:

https://en.wikipedia.org/wiki/List_of_m ... s_by_Apple

 

[vershner]

While this particular piece of anticompetitive behaviour is not applicable for an antitrust case in itself, hopefully the EU will look at it as part of their current antitrust investigations into Apple.