cbiltcliffe writes:
Crittercism has performed testing on mobile platforms, and found that while Apple's release of iOS 7 is the most stable iOS release to date, reducing the crash rate to 1.7%, this still pales compared to Android versions 4.0+, which have a crash rate of only 0.7%.
Not surprisingly, gaming apps with complex graphics and sound crashed more often (4.4%) than other apps, with eCommerce apps getting the best rating of only 0.6%. Some pre-digested coverage from gantdaily.com can save having to dig through pages of research data slides, if you're not looking for the gritty details.
Is this consistent with your experience, or does your particular usage tell a different story?
This discussion has been archived.
No new comments can be posted.
Testing Finds Android More Stable than Apple iOS.
|
Log In/Create an Account
| Top
| 1000 moderator points
| 22 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(Score: 3, Funny) by crutchy on Monday March 31 2014, @05:42AM
/me grabs popcorn
(Score: -1, Offtopic) by Anonymous Coward on Monday March 31 2014, @06:17AM
Oh you again!
</popcorn>
(Score: 0) by Anonymous Coward on Monday March 31 2014, @09:01AM
Article is wrong. Says iOS less stable then specifically says they only tested apps. Once I see a couple other studies I might consider taking this seriously.
(Score: 5, Interesting) by oodaloop on Monday March 31 2014, @05:44AM
I don't own an iOS device, so I can't compare directly. But I am running Key Lime Pie (fuck Nestle) on the Google Nexus 5, and have found it to be incredibly stable. I can't think of the last time it crashed or when I had to restart to fix something.
Many Bothans died to bring you this comment.
(Score: 5, Interesting) by mrbluze on Monday March 31 2014, @08:02AM
I had instability with Samsung's android variant on my S4, so I put on Cyanogenmod which is closer to stock Android. It's more stable but the battery life suffers due to a lack of drivers I think. But stock Android is very stable. However is it more secure? No. Is it a better phone than my Nokia S40 series tough phone? No .. it wouldn't survive being driven over by a car or being run through the wash, it doesn't have a battery life of 1 week, and my Nokia never crashed.
Do it yourself, 'cause no one else will do it yourself.
(Score: 4, Interesting) by Grishnakh on Monday March 31 2014, @09:33AM
My wife used to have an iPhone 3GS, and now we both have HTC Sensation 4G phones. The iPhone crashed now and then, but the HTC is just plain horrible. It's slow as shit, and it crashes constantly. However, the definition of "crash" is different: on iOS, when it crashed, it really crashed hard, usually needing some secret button press which I forget now to force it to reboot. The HTC phones have serious problems, but they don't appear to be with the Android OS really, but rather with HTC's shitty "Sense" UI they slapped on top. So, for instance, every single time I use Google Maps to navigate somewhere, after I reach my destination and close the app, it crashes, and I have to stare at "HTC" in the middle of the screen while it restarts the Sense UI. It's not the whole phone rebooting, it's just Sense. Also, the web browser is crap, and constantly crashes, dropping me back to the main screen. I don't know if this is because the Android browser sucks, or because HTC fucked it all up somehow.
It's really a shame, because the HTC hardware is actually really nice, but the software is shit. I've thought of trying out CyanogenMod on it, but it doesn't look like the Sensation 4G is all that well supported on it (these phones aren't that new any more). I really like the replaceable batteries and microSD cards, features some other Android phones (Nexus) don't have.
(Score: 2) by mcgrew on Monday March 31 2014, @10:15AM
Thank you for that, I'll be sure to stay away from HTC. I've had a Kyocera running Jelly Bean since last summer when I upgraded my flip phone. It's a cheap phone and getting cheaper, they're half the price I paid now.
It hasn't crashed once since I got it. Of course, the fanciest apps I run on it is YouTube and TuneIn.
Hard to compare iPnone and Android, since so many manufacturers are using Android. If you buy the wrong hardware you'll think the OS sucks. Apple is based on BSD and Android is based on Linux so I wonder, what's the stability of BSD compared to Linux?
Free Nobots! [mcgrewbooks.com]
(Score: 2) by Grishnakh on Monday March 31 2014, @10:36AM
Generally speaking, the BSD and Linux kernels are both renowned for very high stability. I guess one question might be how the two differ when dealing with out-of-memory situations.
The big question, of course, is what exactly is causing all the crashes, on either OS. I suspect the main problems are not in the kernels, but the rest of the systems. With my HTC as I said before, I've never had crashes that forced the phone to reboot (I have rebooted the phone many times, however, just because it had gotten too slow or problematic, but again this is a userland problem because the system overall isn't well architected and has too much crapware in it, and doesn't give you any way of shutting it down or removing it); the problems are all in the upper layers of software and especially the bolted-on HTC Sense UI. Samsung has something similar called TouchWiz BTW, so I don't know that those phones would be any better.
(Score: 1) by Acabatag on Monday March 31 2014, @09:31PM
Possibly if the HTC 'Sense' is just a launcher, you could install and use a more stable launcher. Look up 'launcher' in the app store. The ui on android is replaceable, like a window manager on desktop linux.
(Score: 3, Informative) by TGV on Monday March 31 2014, @08:03AM
The testing seems to have measured crashes in apps, not in the OS. The final slide defines crash rate as "Percentage of app loads that result in a crash." While it can inform you what to expect from an unknown app, it doesn't say anything about the OS per se (as the title implies).
(Score: 4, Informative) by Open4D on Monday March 31 2014, @08:37AM
Yes, I think the title should be "Android Apps May Be More Stable Than iOS Apps".
TFA (direct PDF link [documentcloud.org] in case people prefer it) does contribute to the confusion though. They use statements like "Android tablets have worse crash rates than Android phones" (page 14) and "iPhone 5 crashes least" (page 15). But I do still think they're talking about app crashes.
Anyway, could it be because Android app dev is done in a somewhat higher level language than iOS app dev, thus protecting programmers from certain types of mistakes?
(Score: 1) by TGV on Monday March 31 2014, @08:45AM
Who knows? I don't know the dirty details of Android development, but most seems to be done in Java. The (few) native apps I've seen have the UI in Java, and the core in a native DLL, so start-up would still be Java. That might be enough to explain that start-up is more stable.
(Score: 2) by umafuckitt on Monday March 31 2014, @09:43AM
I looked at the presentation and it just says "Apps on platform X crashed Y% of the time" What is that a percentage of? I don't get it.
Anecdotally, my Nexus 4 with Kit-kat does restart spontaneously from time to time (maybe once very week or two). I haven't noticed my partner's iPhone 5 doing that. The Nexus 4 also get absurdly hot when doing in-car navigation and plugged in for charging. In fact, last time I used it it clocked an internal temperature of almost 60 degrees after an hour of usage (don't know where the temp sensor is) and shut down automatically. I wonder if the Nexus 5 runs cooler?
(Score: 2, Informative) by nil on Monday March 31 2014, @10:11AM
From my experience with my Nexus 5, I'd say it runs cool enough. It does get warm when doing processor or graphics intensive things like navigation and some games, but it never gets uncomfortably hot. As far as the spontaneous restarting, I don't recall that having happened on my Nexus 5, but it has happened occasionally on other android phones I've owned. For that matter, it also happened longer ago on the old iPhone 3g that I used to own as well (granted, that is a fairly old model of the iphone).
(Score: 2) by Nerdfest on Monday March 31 2014, @12:12PM
I have a Nexus 5 and I don't recall it *ever* restarting The only time it's been rebooted is for OS upgrades or when I accidentally hold the power button in too long, which has happened a couple of times. My Nexus 4 was pretty much the same. Before that I had an original Galaxy running Cyanogen, and it did occasionally reboot. I've also had my 2012 Nexus 7 tablet reboot spontaneously a couple of times, generally when running a bunch of stuff while updating apps. It only has 8GB and generally only about 200 MB free, and I would guess that that has something to do with it. It doesn't seem to handle low space situations well.
(Score: 3, Informative) by BasilBrush on Monday March 31 2014, @10:30AM
A common cause of an app crashing on startup on iOS would be the main thread timeout of 20 seconds. If an app author does any lengthy initialisation on the main thread, particularly if it relies on a network connection, then that would cause a crash.
Does Android have a similar timeout, and if so what length is it?
(Score: 3, Informative) by mojo chan on Monday March 31 2014, @12:59PM
On Android if your main thread doesn't respond for 10 seconds it is killed off and an error displayed.
const int one = 65536; (Silvermoon, Texture.cs)
(Score: 1) by BasilBrush on Monday March 31 2014, @01:45PM
So this is ANR, is that right? If so then it displays a dialog, rather than crashing straight away. With the options to "Wait" or "OK" (to close).
And of you select OK to close from the ANR, does that count as a crash, or as a clean close of the app?
I'm just thinking that maybe this difference in behaviour could explain the measured crash on opening rates. iOS behaviour is always a crash when an app goes unresponsive for 20 secs.
(Score: 3, Informative) by Bartman12345 on Monday March 31 2014, @12:25PM
Since this article comes to us courtesy of the "What's a Windows Phone" dept, I'll chime in with an answer.
Stable, that's what.
I purchased a second hand Nokia Lumia 820, mostly out of curiosity. It has been rock solid, which is more than I can say for my last two Android phones.
The stability comes at a cost though. Microsoft has locked down their Windows Phone 8 OS as tight as a fish's arsehole, thus preventing app developers from doing anything to fill the functionality gaps, but apparently also ensuring they can't easily hose the OS.
Windows Phone 8 has great potential, but in its current form is lacking too much basic smartphone functionality to be a serious contender. The WP8.1 update could be a game changer, and is only weeks away. I have high hopes this will bring Windows Phone up to speed, but somehow I just know it will be found wanting.
(Score: 0) by Anonymous Coward on Monday March 31 2014, @06:26PM
Despite your anecdote, however, Windoze has a 3-decade-long reputation where the common perception is that M$'s OSes blue screen, freeze, and require reboots for no good reason.
MSFT would have done well to have NOT re-used their legacy trademarked name[1] on their mobile OS and instead come up with a unique name for that.[2]
Foot, meet bullet.
[1] Every time I think of being able to not only trademark a generic word but having the temerity to sue when someone else [google.com] has the audacity to use a name that is different by one letter, I realize just how broken the "intellectual property" regime is worldwide.
[2] Word, Office, .NET. WTF?? [itworld.com]
For a company involved in "creative" undertakings, M$ has some of the least creative people choosing names for things. [stackexchange.com]
Wanna see folks who had this figured out before BillG was born?
Look at Big Pharma's product names.
Now, a name that I do like because it was so descriptive was "wince".
...then again, *n?x has fsck (Filesystem Check). 8-)
-- gewg_
(Score: 2) by MrGuy on Monday March 31 2014, @01:32PM
My response to everything in the presentation is "OK, that's interesting. So what?"
Graphic heavy apps like games crash more often than transactional apps, but transactional app crashes are potentially more significant? That's interesting, but so what? Should we make different kinds of apps? Is there something we can learn from about transactional apps? Can you tell me which kinds of calls are more likely crash, so I can avoid them?
iOS 7 is more stable than iOS 6. OK, so what? Should we make apps that ignore iOS 6? Are there iOS 6 features we should deprecate? Are there ways of recovering from crashes we should implement?
Some android manufacturers' devices crash less. OK, so what? Should Android app makers manufacturer-lock their apps? Are you trying to get the manufacturers of more crash-prone devices to change something? Are there specific things some manufacturers do better/worse? Are there specific calls that are problematic for some manufacturers, so developers should consider avoiding/working around them?
There is nothing remotely actionable about any of this information. Nothing anyone can do with it. Nothing it tells us to change. Nothing that we can use.
This is a classic case of someone not realizing that the fact that you have some data doesn't mean you have something interesting to say.
(Score: 2) by Tork on Monday March 31 2014, @01:43PM
It wasn't that long ago that iOS 7 was released, the key difference being it's full 64-bit. And... yes the transition has been, relatively speaking, painful. I do want to emphasize the word 'relatively', though. Basically there were a handful of situations (like starting the camera app and switching to the video mode) that would cause the UI to reset. The frequency of this was something like once a week and I didn't lose more than 20 seconds of time to it. This is the worst I've ever seen iOS 7 and it is outright heaven compared to the Palm Treo I had years ago.
There have been 2 or 3 updates since 7 came out, each improved stability, and nearly all of those problems are gone. However, I did have a browser-related reboot yesterday... but it was the first time in something like a month something like that happened. I read somewhere that the problem was that a 64-bit app eats more memory than a 32-bit app, meaning when the device gets low on memory that when something new asks for some it runs past the end. Okay I'm explaining that badly, but I also think that sounds a bit BS'y, I mean wouldn't the first thing that'd happen when switching to 64-bit is that memory allocation would be dialed in well...?
Anyway, yeah, this has not been a banner year for iOS stability. But even at its most unstable I haven't, for example, not received a phone call or missed an alarm. I've never had to reboot the phone because it was frozen. My Galaxy Tab has behaved just as well. Frankly I much prefer this era to what I had before, I am able to take these devices for granted. I no longer have a backup alarm clock.
I do want to warn you all, though, that Samsung announced not long after iOS 7 was debuted that they're going 64-bit late in 2014. My guess this time next year you'll see a similar report only Android and iOS are flipped. I would urge caution and patience, afterall that's a major transition.
Slashdolt logic: 1600 x 1200 > 1920 x 1200