Tim Cook: Android version fragmentation is 'terrible for developers' Before CEO Tim Cook handed over the iOS 7–introduction chores to other Apple execs at the company's Worldwide Developer Conference, he took a moment to slag Android and its OS fragmentation in an appeal for the hearts and minds of developers. "This version fragmentation is terrible for developers, as many of you know," he told …
Jobs spent much more time during these events talking about how great Apple stuff was, and mostly left insults against competitors up to the users. Apple may have poked fun at their competition with stuff like the ads with the Mac guy and the PC guy, but they never said the kind of stuff Cook said, which was basically "you suck, neener neener neener!"
While I'm sure this plays well with the audience at the WWDC, it makes Cook and Apple look small when he's quoted in the media.
"it makes Cook and Apple look small when he's quoted in the media."
Exactly, just like MS, Apple should realise they're in a bad place when they have to spend a chunk of time explaining (in their clearly unbiased opinion) how bad someone else's OS is instead of how good theirs is.
"Jobs spent much more time during these events talking about how great Apple stuff was"
...and this is exactly how you should do it. Jobs may have had his faults, but Apple's fortunes did turn around when he was there and negative marketing (criticising competitors) is a defeatist way to operate and is usually a sign of weakness, poor judgement and lousy marketing.
Most products and systems have advantages or disadvantages compared to other, more so when the operating environments are different - and don't forget that while iOS and Android nominally are similar their operating environments are different: Apple have a tight reign on the hardware, OS and applications where Android is much looser and open. [This isn't an argument as to which is "right", just stating facts - both approaches have major positive and negative points].
So when this guy starts to criticise a competitor like this (negative marketing) then it's an indication of weakness in him and likely his products too. Would you rather deal with somebody who is positive about their own products or somebody who is busy being negative about a competitor's where they should be telling you about theirs?
Pointing out facts is now seen as slagging off?
If he was telling porkies I'd agree, but he is in fact telling it as it is, however unpalatable Android users find it.
Though not the users fault but Android phones generally fail to get updated because either the manufacturer or service provider fail to deliver and only give the latest versions with a new phone. Those that do upgrade, google phone excepted, get flawed updates stuffed with junk from the service provider.
Apple users are more fanatical about how much they like Apple things - shock, news at 11. I don't think that means a product is better. If I love my Android phone, but also note there are some things I'd criticise; but someone else loves their feature phone, and is ignorant of the things that other phones do better, does that mean the latter is better? Better at marketing, maybe.
The Android average will be pulled down by two factors:
* People who buy lower end cheaper phones. Although just as many (more in fact) high end Android phones are bought as iphones, the lower end will still drag down any averages.
* Niche platforms always do better for users being fanatical, because a dominant platform like Android is also used by all the people who don't give a damn about operating system, platform, smartphones or whatever, and so don't feel enthusiastic about it. Back in the early 90s when the Amiga was a popular home computer, but also a computer that the mainstream users themselves would criticise, I predicted back then the same thing would happen to PCs when they became popular in homes. And look today - I was right, people moan about PCs, but the minority of Mac users are still fanatical (and as the Amiga userbase shrunk in the 90s, the users that remained were more likely to be the ones who'd praise rather than criticise their platform).
And the other charts are misleading too. Since when do I have to code specifically for say, Android 4.0, but not 4.1? It's easy to just set a minimum requirement. Even if I limit my apps to 4.0 or above, there are still more users than all IOS users. Problems with fragmentation are more about different devices to have to test on - but with numerous iphones, ipads, ipad minis, and even rumours of a 12" ipad, that's a problem with IOS too now.
And for download revenue? Yep, Apple is expensive. I'd prefer the platform with lower costs for software, and more freely available software. What next - are Windows users going to criticse Linux because "software revenue" is lower for Linux??
I think slagging off someone else is pathetic and weak. It's IMHO tragic if you think you need that, instead of trusting your audience to draw its own conclusions.
I may be fairly happy with Apple products, but I am certainly not impressed with Tim Cook's new leadership. The time to give the guy a break because he's new at it has passed, and IMHO he's not up to snuff. He suffers as much of the lieutenant syndrome as Ballmer, only has a different style. His bullying is more the continuation of this stupid trivia patent war and now this even more stupid comparison game.
Show what is better, don't give competitors airtime. The latter only tells your audience you're worried about them. If you don't trust your own products to speak for themselves you really should go and do something else, like selling hamburgers. I fear that Apple risks sliding back to where they were before Jobs came back (slowly). That would be a stupid waste of ability and market capital.
Lose this guy.
Quite clearly anyone that believes this tripe is a cretin.
Apple play games. They have 4 different versions of ios6 in reality, all with different features,but his fancy chart pretends they are all the same.
What he also fails to mention is android API levels are very static. Targeting ICS and newer gets you 70 something percent off all devices without no real loss in API features ( most are in the backwards compatibility libraries and there is action bar Sherlock to support gingerbread)
Making a web app allows you to target more than that. Something like 100% of android devices, 100% of iOS devices, Windows phones and Blackberries. So stick your api behind a javascript library extension and shut up. The world doesn't care which OS you hide the web behind, just get out of the way and let us get some stuff done!
> Oh you are soooo right, they are struggling are they not.......
Erm, well yes, they are certainly slipping.
What people were really looking for was some genuine innovation. What they got was some tweaking of already existing products.
Jobsy would have had something brand new to show the fanbois and fangals. He was an arse, but he knew what the word innovation means.
"Ugh ugh the Galaxy S4 has a ton of stuff we have no hope of integrating like NFC (our antenna guys suck) but look people are still using Android 2.3 3 years after it was released instead of buying the new shiny all the time. Don't they know they could buy the latest iPhone every release and get no new features?"
Dear silent_count: "... it takes iPhone users twice as long to get their phone to do what they want it to ..."
Yawn... Your knee-jerk anti-Apple silliness is tiring.
Grow up, m'boy, grow up, and provide some useable data rather than foolish, unsupported trash-talking.
You bore me – along with anyone else who prefers factual analysis to simple-minded snottiness.
"I think he just meant that the average amount of time an iPhone user uses his or her device for is 50% greater than that an Android owner does."
So, it only takes ios users double the time to do something compared to Android users?
Hey! Those fanboys are getting better at using computers, next time they may even be capable of preordering online instead of queuing.
This post has been deleted by its author
Well I guess it depends why that is...
Maybe iPhone uses spend 50% more time on it because it takes longer to do anything on it. It's a nonsense statistic as presented of course.
Fact is I have many Android devices in my house, and most of them are only used for a few minutes a day. Why? Because I'm using Android to power smart remotes and so on, so I just don't *need* to use them all day. I'm sure there are plenty of other reasons and probably more widespread than my example though.
The "uses it more" statistic had no useful context to determine what that meant and so it means absolutely nothing. Like most statistics.
"Fragmentation amongst Linux distributions will be the killer blow", a critical comment which actually holds some truth because setting up software on Debian can be a completely different ballgame from doing the same on SuSE. But Linux is still going strong, ever so popular. Maybe other people can see clear signs as to why the fragmentation is extremely bad, I sure don't.
What about open source software as a whole? The fact alone that you can simply cast the magical "fork" work (and issue a quick cp command when no one is looking) was also reason for complaints. Software projects would fragment, this causing damage to the entire species.
I think that the OpenOffice project (deliberately mentioned) showed us otherwise. I think it's fair to say that the forking and fragmenting of LibreOffice may very well have saved the entire project.
If this really about a fair warning to all potential developers? Or a deliberate (and very transparent) scheme to put your own environment better into the picture?
A bit lame, don't you think?
Re Fragmentation
a) is it really THAT difficult to have software supporting 2 or 3 versions of Android?
b) even if my software only work on 1 of the most popular android version (which each have approx 30% of the android install base each), 30% of all android phones is still more than 90% of all iOS phones and that gap is getting bigger.
I would think that as a developer, the biggest draw of developing for iOS over android is that iOS users are more willing to pay for their apps
Partly. Also iOS users are less likely likely to be innate tinkerers who will crack your Java and post it free download sites. And less likely to think "This app sucks, I'll write a better one" and actually do so.
iOS has significant benefit in terms of ease of use for users. The drawback is the extremely high barriers to entry (cost a Mac to develop on, need to learn the filth that is Objective-C) for developers.
Let's say I support only the "two big ones" (most used 2.x and 4.x). Assume I want to use some of the better 4.x UI features like Fragments I have to support two software trees since 2.x needs add-on libraries for that. Google overhauled the UI and added some of the Android 3 (tablet only) tree. The also reworked the guidelines
And that assumes I do not have to support multi-user (4.2, quite rare still)
And WebApps are a so-so solution. Either need extra libraries like PhoneGap or they can not access the phone functions. Not sure if PG can work with special features like the Note series provides. And the LookAndFeel is not Android/iOS/WP8 but "something foreign"
>I have to support two software trees since 2.x needs add-on libraries for that.
I don't think the support library actually requires you have 2 code bases as the support versions of Fragment etc work just find on 4.x. You could also use ActionBarSherlock to do all of the grunt work for you. It does the right thing depending on the SDK version it's running on. If you google hard enough (actually not very hard) you might find libraries that will help you out.. who would have thunk it?
>And that assumes I do not have to support multi-user (4.2, quite rare still)
How much extra work is needed to support multiuser? 4.2 isn't as rare as you think but the multiuser support isn't as hard as you think.
>Either need extra libraries like PhoneGap or they can not access the phone functions.
Or use ActionBarSherlock to deal with the support library for you? I think your issue with actually doing your research is bigger than any issue you could have with using the support library.
Maybe you are right these days. Still it IS additional effort and / or libraries. It WAS a major problem with libraries a good year ago when I was last forced to program for it at the job (never would do that in privat)
Thankfully company dropped the Android platform for Win8/x86 where tablet pc are concerned and dropped phone support completely. Me, I dropped the last Android device (Note smartie) in the (e)bay recently. Featurephone and a spanking new TPT2 do the job a lot better
"Fragmentation" seems to have broadened into such a catch-all term, that it's become nearly useless.
Back when they talked about "fragmentation" in the Unix world, it was a problem, because inter-operability between and porting applications between different, *proprietary* Unix OSes was a real hindrance. And was born... POSIX.
Then they started talking about how terrible "Linux fragmentation" was, and how crippling it would be to the success of Linux outside of hobbyist's garages / the parents' basement. Of course, in a FOSS (ie. Free/Open-Source Software) environment, this "silo effect" just didn't have much relevance -- there was no proprietary, arbitrary, artificial obstacle to interoperability or portability. The Linux "silos" turned out to be just adjacent rooms along the same corridors.
Now would-be detractors are spouting the same sort of nonsense about "Android fragmentation".
But the correct word, or at least a much better one, would be "diversity" (or maybe "adaptability, or simply "variation"). And funnily enough, this diversity turns out to actually be an advantage -- coping with a diverse OS environment and varied hardware exposes bad code, and strongly favours solid. robust coding and development practices.
Could also mean that iOS forces it's users to spend 50% longer using it to achieve the same ends and that Android is less intrusive and a more frictionless UX. Comparing my experience as an Android user with my iOS using partner, this would certainly seem to be the more likely explanation.
But who needs objective, meaningful analysis when you have numbers to throw around, eh Mr Cook ? :)
Indeed, and don't even need to check a widget to see bus times. Android will just automatically tell me it's time to leave, and Google maps will show me which buses to take. True, iphones have this too - thanks to Google - but the point is the survey is flawed if this extremely useful feature is only counted for the 5-10 seconds I look at my phone, and deemed hundreds of times less important than an hour of staring at Facebook.
I use my phone 24/7. Just because it sits in my pocket doesn't mean it's not useful for that time.
Fragmentation happens. The actual pain caused to me (as a mobile dev) is purely based on how many different screen sizes i have to make for.
Just like with the web in the olden days, you make it work for as many people as possible.
Currently, iPad mini's are really annoying due to their different screen shape (square, not rectangular meaning things dont scale well from a long thing rectanglular iphone 5).
Also mr cook, very few developers work only on iOS "or" Android,
Personally I've found the annoyance of different device and screen aspects and resolutions more annoying on iOS than on Android. Not that the wide range of screen resolutions on Android isn't an issue, but it feels like I have better inbuilt tools to deal with one application and multiple resolutions and ratios than in iOS.
Yes, there are different versions of Android to deal with - currently two main ones unless you want to be cutting edge. But even that's not too hard as you can target the cutting edge and have fallback to the older versions as the support libraries work quite nicely (at times :-) ). It does require testing but if you develop applications properly and cleanly separate functionality from interface (Model - View - Controller) then even if you have entirely different interfaces it is not always that difficult to develop, after all, many of us develop apps that can be operated in landscape or portrait mode and this kind of model is normal to us.
I agree. It also helps I think that Android was designed from the outset to scale to different sizes, resolutions and aspect ratios. E.g., there are 4 different "sizes" that one can design UIs for, and developers have always known to test for a range of configurations (which can be done in the emulator).
IOS was sold on the idea of "it's easier, because there's only one device to develop for". Unfortunately it's been a slippery slope:
* 2010, ipad. Now two devices, different sizes and aspect ratios, but still claimed better than Android. Okay. Though this led to the myth that on IOS you can develop for two screen sizes separately, despite Android always allowing for 4 screen sizes...
* 2010, iphone finally gets a sensible resolution. Hacked to work by saying applications can just double width and height.
* Various changes to ipad resolutions, leading to increasing number of configurations that need testing.
* 2012, new iphone with different size and a new aspect ratio, leading to "black bars".
* 2012, new ipad with different size.
So now we're at 4 screen sizes, 3 aspect ratios, at least 4 resolutions. And now there are rumours of a 12" ipad on the way...
My dad has an iPad 1 and my mum has a 3gs and these cannot upgrade past iOS 4.6 or something similar. I call bullshit on the 6% stat for non iOS 6 phones. That or most iTHINGYS crap themselves after 2-3 years...
Either way... I call bullshit.
Also. Android 4.0, 4.1 and 4.2 are virtually the same kernel afaik. Should count them as one.
This was a weak WWDC from Apple.
>Android 4.0, 4.1 and 4.2 are virtually the same kernel afaik.
If you're talking about Linux kernel versions.. That doesn't actually matter too much.
>Should count them as one.
4.1 has some features that 4.0 doesn't like some video acceleration stuff in libstagefright IIRC. 4.2 has the multi user stuff but it's mostly the same stuff to the user.
I basically only see there being two groups of Android devices; 2.2 and 2.3 devices and 4.x. It's not hard to support 2.x devices if you are careful. You can actually put most of the 4.x UI into your application and have it look exactly like it would on 4.0+. Does make your APK at least ~15MB though ;).
If your app is based around something like NFC that wasn't officially in 2.x then there isn't much point supporting anything but 4.0+.
...makes it hard to develop iOS apps too. Docs (especially tutorials) are rarely in sync with the latest version which invariably doesn't run on version -2 or sometimes version -1 of the current Mac OS. This is fragmentation too, and it's well within their power to prevent, and yes it hurts developers. Apple's third party developers suffer from Stockholm syndrome.
"We paid out five billion of that in just the last year."
There was me thinking that it was the Apps that made money and Apple take a very hefty commision.
Apparently it's Apple's money which they share with the developers.
We must all bow down before the wise and generous Apple.
The Windows PC ecosystem has always been fragmented but it hasn't put off developers. Microsoft have several flavours of one version of Windows for a start (basic,home, pro, ultimate etc) and different versions being supported at the same time (XP, Vista, 7, 8 and server) and lots of different speeds of CPU, amounts memory, screen size variations. Yet it hasn't stopped Windows becoming the dominate OS for PCs.
Because the system neatly hides those differences. Yes, the new versions have added abilities - but the old API is still there and still works. Yes, servers have special skills - but no person with a three digit IQ one uses a server as a client and vice versa.
Sure the better versions have added capabilities - but they are mostly irrelevant to end user software since the system hides them well. Properly written software does not care if the HDD is encrypted, the box gets authentification from an ActiveDirectory instead of locally etc. Speach recognition is a nice to have for some jobs but the way it is implemented is transparent, not part of the end user software. No need to program against a special API
Screen size does not matter much. The vast majority of Windows units has at least 1024x768 for decades and that is the "initial size" for many programs. One of the reasons the little 1024x600 cripples aka Netbooks died. And using a bigger screen has not been a problem for Windows for at least a decade
The main difference between Win XP (the lowest NT-line platform sensible to support) and Android 2.3 is that XP already was a stable mature OS with user accounts, proper access control and a capabel graphics subsystem and UI libraries that can use all the stuff a modern box offers(1). Stuff that has not changed much (if at all) for a user program since than. Windows has done changes in other areas post XP.
Android 2.3 OTOH lacked a lot of those / the latter versions added tons of features (like "proper" tablet support). So choosing 2.3 as a baseline either forces you to use additional libraries/tools/workarounds - or to not use quite a few features.
The only stuff that needs to know the exact version and variant is admin level stuff. And the MS internal one "knows" while external software is rarely used by the average end user
(1) Pens need a add-on package aka tablet edition but even that is compatible to Vista and above
There seems to be no shortage of dorks who draw a comparison between 'Android' and 'iPhone'. Grow a brain people. One is a device and the other is an OS.
So if you're saying 'there are more installs of Android on the 100 odd devices capable of running it, versus the few models of iPhone' then you'd probably be right. But if you take the signature devices like the Galaxy S2 - S4 versus the iPhone 3GS - 5, you're not even close. Pound for pound, iOS is superior.
And for all the fandroids who insist on giving us the 'emperors new clothes' speech about buying iDevices, we pay the extra for the level of quality that comes with the product. In the case of Android devices, it's often a case of 'buy cheap, pay dear'. For development, I own both iPhone 5 and Galaxy S3 and the iPhone never crashes. Whereas an unmodified and fairly empty S3 crashes regularly and without good reason.
People are comparing 'Android' and 'iOS' as found on the 'iPhone' and 'iPad', so who's actually the dork?
To add to the anecdotal 'evidence'... I have an iPhone 4 (still on iOS 5) and a Galaxy S3 (on stock 4.1.2). The iPhone has crashed at least twice in the past two years. The S3 has never crashed in the six months I've owned it.
"For development, I own both iPhone 5 and Galaxy S3 and the iPhone never crashes. Whereas an unmodified and fairly empty S3 crashes regularly and without good reason." this is a case of 'your mileage may vary" as I've had both have froze in the middle of a call, needing a hard reset.
But "iPhone" isn't a device, unless you're really talking about the 2007 model. It's a product range - and therefore fairly compared against, e.g., the Samsung Galaxy range (which massively outsells it).
Alternatively we might compare operating systems (and yes, you can include tablets too if you like). Again, Android wins.
Or we can compare by company - Samsung wins.
Comparing by individual devices models though makes no sense. As I say, "the iphone" isn't a model, you're now talking about say, "iphone 4S". If one cake shop sells 101 cakes a day, and another sells 100 cakes, plus another 100 cakes with cherry on top, only a fool would try to spin the former as more popular based on the individual model.
But if you really want to compare individual models, okay. The S3 was the most popular device (beating any iphone, Q3 2012). And the most popular smartphone of all time is the 2009-released Symbian powered Nokia 5230. Sorry, no iphone device comes anywhere close.
"Apple's App approval process is 'terrible for developers'"
Unless things changed, as an iOS user I remember some apps taking ages to be updated while Apple reviewed them and decided if they should be allowed or not. I don't think its cheap to get started as an iOS app developer either if you include the price of buying a Mac and yearly fee.
Android development tools are free and the registration fee was only $25. I don't think the fragmentation is hurting Android or its developers that much at all.
Sensitive
I watched the keynote and from what i saw they spent much more time showing the 'new' features (many seen on Android before yesterday, but what appear well implemented) than dissing any competitor.
I wouldn’t necessarily call it dissing either, I would call it banter and then the Android and iOS 'crowd' getting too sensitive. As always happens. iOS 7 looks like a good improvement over iOS 6 and I’m looking forward to having it on my iPhone. Equally, I’m looking forward to getting the next release of Android on my nexus 10 and in addition I’m looking forward to finding a tech blog to read that has non-biased authors, which seems unlikely to happen any time soon.
Is the iOS figure that incredible? People seem to be arguing that Android phones are outselling iPhones.... but don't iPods also run iOS?
How many iPods have been sold that are now running iOS6?
Anyone who unquestioningly believes when company X produces statistics showing their own product/company/whatever is the most popular/cost effective/blah shouldn't be allowed to handle sharp things in-case they hurt themselves.
What would be more interesting is the actual figures instead of %s. I was once told in statistics you use % if the numbers are too close. The cosmetics companies understand this... what looks better... 50 people rated us better than the competition or 75% of those surveyed?
What would be more interesting to see would be having to do product reveals to an audience who prefer competitor products and not the faithful. Then you know if you've done something truly amazing.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
