They are. And as explained to davesw, all of them are based on the AOSP release, which is the vanilla base code for Android. As such, they are all compatible with apps from the Market, as long as the developer has written the app for 1.5/1.6/2.1/2.2 compatibility. So developers still only need to target at most 4 OS releases.
They are, the important info includes the share of each one, how important it is to support and how much it differs from Google's version but that info isn't presented.
As a consumer, I like Apple's model better. A device manufacturer should like Google's model better. Some consumers would say Google's model gives more hardware choice, which is fine but I haven't seen better hardware than the iPhone. Cheaper certainly but not better so having a choice of hundreds of poor quality phones instead of one isn't really a selling point for me.
On the fragmentation though, if developers aren't having issues, there isn't a problem. Not much more than the iPhone - I get a fair amount of apps that tell me I can't run them on my model of phone.
It's a little unclear, however, as to exactly where the "not a problem for developers" line crosses over into "yet still a problem for consumers."
Since Android is doing such a good volume business, devs may feel pretty OK about building apps that work on some or most phones, and letting the rest fend for themselves. At that point it's a numbers game-- are there enough handsets of a particular flavor out there to make a good business, without having to worry unduly about alienating the owners of some number of other, variant handsets?
For the individual consumer, however, that may mean that a particular handset keeps running into irritating problems with running particular apps. My guess is that "fragmentation" over the long haul doesn't necessarily mean a million, equally large segments, but rather a bell curve, with the majority of the installed users base being more or less on the same build, and becoming increasingly inoperable with the primary ecosystem as you move off the central bulge.
That does mean, however, that a not insignificant number of Android handset owners will perceive the platform as being a frustrating mess, which may have consequences for long term loyalty and adoption rates.
At that point it's a numbers game-- are there enough handsets of a particular flavor out there to make a good business, without having to worry unduly about alienating the owners of some number of other, variant handsets?
Quote:
Originally Posted by davesw
a bunch of users/developers are already complaining about Android's fragmentation
The complaints are mainly about supporting Android 1.x, which is currently under 25% of the Android market. Compare it to the new Unreal engine for the iPhone. The iPhone 2G, 3G and original iPod Touch can't run it nor do they have multi-tasking enabled - the originals are limited to iOS 3.x. This share is at least 20 million devices out of the total 80 million so similarly 25% fragmentation.
Android is far more likely to have subtle variations beyond this type of thing given that it's not just one company calling the shots over how the OS is built but it's not Google's fault. It's the people who are deploying it who choose to change the system Google provides that are to blame. It's also an inevitable consequence of hardware development.
It's not important though. For Apple to win, Google doesn't have to lose. Apple's focus is on quality not volume and it's better that Google control the volume because they are heavy Apple supporters unlike RIM, Nokia and Microsoft.
The complaints are mainly about supporting Android 1.x, which is currently under 25% of the Android market. Compare it to the new Unreal engine for the iPhone. The iPhone 2G, 3G and original iPod Touch can't run it nor do they have multi-tasking enabled - the originals are limited to iOS 3.x. This share is at least 20 million devices out of the total 80 million so similarly 25% fragmentation.
Android is far more likely to have subtle variations beyond this type of thing given that it's not just one company calling the shots over how the OS is built but it's not Google's fault. It's the people who are deploying it who choose to change the system Google provides that are to blame. It's also an inevitable consequence of hardware development.
It's not important though. For Apple to win, Google doesn't have to lose. Apple's focus is on quality not volume and it's better that Google control the volume because they are heavy Apple supporters unlike RIM, Nokia and Microsoft.
They are. And as explained to davesw, all of them are based on the AOSP release, which is the vanilla base code for Android. As such, they are all compatible with apps from the Market, as long as the developer has written the app for 1.5/1.6/2.1/2.2 compatibility. So developers still only need to target at most 4 OS releases.
Just another red herring...
Not true given I had to jump through hoops because Sense does something stupid with their soft keys. I haven't run into any blur landmines but I haven't tested any blur devices yet either.
The complaints are mainly about supporting Android 1.x, which is currently under 25% of the Android market. Compare it to the new Unreal engine for the iPhone. The iPhone 2G, 3G and original iPod Touch can't run it nor do they have multi-tasking enabled - the originals are limited to iOS 3.x. This share is at least 20 million devices out of the total 80 million so similarly 25% fragmentation.
Android is far more likely to have subtle variations beyond this type of thing given that it's not just one company calling the shots over how the OS is built but it's not Google's fault. It's the people who are deploying it who choose to change the system Google provides that are to blame. It's also an inevitable consequence of hardware development.
It's not important though. For Apple to win, Google doesn't have to lose. Apple's focus is on quality not volume and it's better that Google control the volume because they are heavy Apple supporters unlike RIM, Nokia and Microsoft.
I'm not suggesting that Android will "lose" or that Google is to blame for anything. Just that, going forward, the modifiability of Android and the struggle for the handset makers to differentiate themselves in the market may lead to even more variants, custom UIs, proprietary add-ons, and the like.
My suspicion is, at some point, "Android" loses all cache as an OS, and just becomes the default "smart phone" OS for cheap 'n plentiful handsets. Pretty much the same as the anonymous operating systems that powered untold legions of dumb phones, just updated for the smartphone era. This will be exacerbated by the fact that HTC phones look and act like HTC phones, LG phones like LG phones, etc-- just like they did when they were making feature phones. Then we're back to hardware manufacturers with extremely erratic ideas about usability and integration. Android is sort of dissolvable, in that sense, and when the "It's Android so you know it's not total crap" phase wears off Android-ness will be soft-pedaled-- if that isn't happening already.
Nothing wrong with that from Google's perspective, of course, but then you're going to have people looking to upgrade from that, just as they are now upgrading from dumb phones. Apple, for instance, is going to keep hammering away at a seamless ecosystem, something handset makers can't really address. Google can, to an extent, but I wonder how long they'll maintain real enthusiasm for advancing the platform, once it's become lowest common denominator ubiquitous. Seems to me at that point their work is done.
The 2G iPhone shipped 6 million in the first year, the 3G was > 6 million in year 2 and the original Touch was at least 6m over 2 years so at least 20 million units. Early this year, the number of iPhones and iPods was reported at 85m so the total might actually be over 100m by now bringing the fragmentation percentage down a bit to 20% but the first 2 years number might also be 25m so it's in the same ballpark.
The Google numbers have been quite widely reported and the source is probably Google themselves but the Apple numbers come from Apple.
Quote:
Originally Posted by addabox
the modifiability of Android and the struggle for the handset makers to differentiate themselves in the market may lead to even more variants, custom UIs, proprietary add-ons, and the like.
Maybe but one difference with Symbian is that Android has high profile apps and it acts against the distributors to break these apps so I think they will go with the stock Android as much as they can and simply add a skin to it.
Not true given I had to jump through hoops because Sense does something stupid with their soft keys. I haven't run into any blur landmines but I haven't tested any blur devices yet either.
I will give you that there will be some variance in regards to the OEMs. However, the list is nowhere near as massive to target as the OP wants us to believe.
Just for my knowledge, is it really that much more work?
Android 1.5/1.6 accounts for barely over 25% of the total vanilla Android OS versions. It would be interesting to see a similar chart showing current iOS versions.
Maybe but one difference with Symbian is that Android has high profile apps and it acts against the distributors to break these apps so I think they will go with the stock Android as much as they can and simply add a skin to it.
I see what you're saying, but that's assuming that not breaking apps is considered a competitive advantage. With the number of Android phones being released, at the rate that their being released, can any substantial number of consumers really keep up with which handset is doing what and how, and how well?
It looks to me like the Android market is starting to resemble the old feature phone market, where bling and specs and advertising and a few high profile features might get a handset its day in the sun until it's supplanted by the next one. With 2 year contracts creating a constant supply of restless buyers looking for the next big upgrade, I think all a phone need do is hit a few functionality marks out of the box (web, email, media and IM) and most folks will reckon its good to go. It there turn out to be some compatibility issues down the road, well, that's just how phones are, as far as the average Android customer knows, right?
But again I think that that might create a kind of gradual, aggregate sense of quality problems for the platform, without most people being really very clear about where the problems are. Its open source roots and Google backing and general air of geekiness has given the impression that Android users are largely tech savvy and discerning about stuff like this, but now that it's blowing up mass market the average Android user is increasingly going to be completely clueless about app compatibility or fragmentation or getting and running the latest OS version or any of that. They'll just use the phone and if stuff doesn't work be bummed. If it doesn't work enough times for enough people, then you start getting a reputation, no matter who is at fault. I think that's one thing Apple is keenly aware of, and they go to great lengths to avoid those kind of scenarios.
So then, as you say, one might sake "But won't handset manufacturers make sure they maintain compatibility to avoid exactly that?", but they're playing a cell phone game, not a computer platform game. In that game you release phones at a breakneck pace, each one "packed" with specs, and run ads with laser wielding robots and flying cars and rain slicked streets erupting into chaos. It's all about driving excitement for the next group of off-contract people looking to upgrade, and there's always another group of them. "App compatibility" doesn't much matter to that mindset, as long as main apps people mostly use are fine-- which is what, about 6?
I will give you that there will be some variance in regards to the OEMs. However, the list is nowhere near as massive to target as the OP wants us to believe.
Just for my knowledge, is it really that much more work?
Yes and no. Yes, in as much as published APIs break and you need to find a work around that works across all variants. Sometimes you find out because a user tells you which is not good. No, in that typically the work around isn't that much work...it's just stupid and generally an ugly hack.
It just seems that the number of ugly little hacks in Android development is moderately high.
Comments
Most of those look like third party ROMs.
They are. And as explained to davesw, all of them are based on the AOSP release, which is the vanilla base code for Android. As such, they are all compatible with apps from the Market, as long as the developer has written the app for 1.5/1.6/2.1/2.2 compatibility. So developers still only need to target at most 4 OS releases.
Just another red herring...
Most of those look like third party ROMs.
They are, the important info includes the share of each one, how important it is to support and how much it differs from Google's version but that info isn't presented.
As a consumer, I like Apple's model better. A device manufacturer should like Google's model better. Some consumers would say Google's model gives more hardware choice, which is fine but I haven't seen better hardware than the iPhone. Cheaper certainly but not better so having a choice of hundreds of poor quality phones instead of one isn't really a selling point for me.
On the fragmentation though, if developers aren't having issues, there isn't a problem. Not much more than the iPhone - I get a fair amount of apps that tell me I can't run them on my model of phone.
Since Android is doing such a good volume business, devs may feel pretty OK about building apps that work on some or most phones, and letting the rest fend for themselves. At that point it's a numbers game-- are there enough handsets of a particular flavor out there to make a good business, without having to worry unduly about alienating the owners of some number of other, variant handsets?
For the individual consumer, however, that may mean that a particular handset keeps running into irritating problems with running particular apps. My guess is that "fragmentation" over the long haul doesn't necessarily mean a million, equally large segments, but rather a bell curve, with the majority of the installed users base being more or less on the same build, and becoming increasingly inoperable with the primary ecosystem as you move off the central bulge.
That does mean, however, that a not insignificant number of Android handset owners will perceive the platform as being a frustrating mess, which may have consequences for long term loyalty and adoption rates.
http://thenextweb.com/mobile/2010/07...-stick-around/
http://apps.ycombinator.com/item?id=1397941
At that point it's a numbers game-- are there enough handsets of a particular flavor out there to make a good business, without having to worry unduly about alienating the owners of some number of other, variant handsets?
a bunch of users/developers are already complaining about Android's fragmentation
The complaints are mainly about supporting Android 1.x, which is currently under 25% of the Android market. Compare it to the new Unreal engine for the iPhone. The iPhone 2G, 3G and original iPod Touch can't run it nor do they have multi-tasking enabled - the originals are limited to iOS 3.x. This share is at least 20 million devices out of the total 80 million so similarly 25% fragmentation.
Android is far more likely to have subtle variations beyond this type of thing given that it's not just one company calling the shots over how the OS is built but it's not Google's fault. It's the people who are deploying it who choose to change the system Google provides that are to blame. It's also an inevitable consequence of hardware development.
It's not important though. For Apple to win, Google doesn't have to lose. Apple's focus is on quality not volume and it's better that Google control the volume because they are heavy Apple supporters unlike RIM, Nokia and Microsoft.
The complaints are mainly about supporting Android 1.x, which is currently under 25% of the Android market. Compare it to the new Unreal engine for the iPhone. The iPhone 2G, 3G and original iPod Touch can't run it nor do they have multi-tasking enabled - the originals are limited to iOS 3.x. This share is at least 20 million devices out of the total 80 million so similarly 25% fragmentation.
Android is far more likely to have subtle variations beyond this type of thing given that it's not just one company calling the shots over how the OS is built but it's not Google's fault. It's the people who are deploying it who choose to change the system Google provides that are to blame. It's also an inevitable consequence of hardware development.
It's not important though. For Apple to win, Google doesn't have to lose. Apple's focus is on quality not volume and it's better that Google control the volume because they are heavy Apple supporters unlike RIM, Nokia and Microsoft.
Where did you get your numbers?
They are. And as explained to davesw, all of them are based on the AOSP release, which is the vanilla base code for Android. As such, they are all compatible with apps from the Market, as long as the developer has written the app for 1.5/1.6/2.1/2.2 compatibility. So developers still only need to target at most 4 OS releases.
Just another red herring...
Not true given I had to jump through hoops because Sense does something stupid with their soft keys. I haven't run into any blur landmines but I haven't tested any blur devices yet either.
The complaints are mainly about supporting Android 1.x, which is currently under 25% of the Android market. Compare it to the new Unreal engine for the iPhone. The iPhone 2G, 3G and original iPod Touch can't run it nor do they have multi-tasking enabled - the originals are limited to iOS 3.x. This share is at least 20 million devices out of the total 80 million so similarly 25% fragmentation.
Android is far more likely to have subtle variations beyond this type of thing given that it's not just one company calling the shots over how the OS is built but it's not Google's fault. It's the people who are deploying it who choose to change the system Google provides that are to blame. It's also an inevitable consequence of hardware development.
It's not important though. For Apple to win, Google doesn't have to lose. Apple's focus is on quality not volume and it's better that Google control the volume because they are heavy Apple supporters unlike RIM, Nokia and Microsoft.
I'm not suggesting that Android will "lose" or that Google is to blame for anything. Just that, going forward, the modifiability of Android and the struggle for the handset makers to differentiate themselves in the market may lead to even more variants, custom UIs, proprietary add-ons, and the like.
My suspicion is, at some point, "Android" loses all cache as an OS, and just becomes the default "smart phone" OS for cheap 'n plentiful handsets. Pretty much the same as the anonymous operating systems that powered untold legions of dumb phones, just updated for the smartphone era. This will be exacerbated by the fact that HTC phones look and act like HTC phones, LG phones like LG phones, etc-- just like they did when they were making feature phones. Then we're back to hardware manufacturers with extremely erratic ideas about usability and integration. Android is sort of dissolvable, in that sense, and when the "It's Android so you know it's not total crap" phase wears off Android-ness will be soft-pedaled-- if that isn't happening already.
Nothing wrong with that from Google's perspective, of course, but then you're going to have people looking to upgrade from that, just as they are now upgrading from dumb phones. Apple, for instance, is going to keep hammering away at a seamless ecosystem, something handset makers can't really address. Google can, to an extent, but I wonder how long they'll maintain real enthusiasm for advancing the platform, once it's become lowest common denominator ubiquitous. Seems to me at that point their work is done.
Where did you get your numbers?
The 2G iPhone shipped 6 million in the first year, the 3G was > 6 million in year 2 and the original Touch was at least 6m over 2 years so at least 20 million units. Early this year, the number of iPhones and iPods was reported at 85m so the total might actually be over 100m by now bringing the fragmentation percentage down a bit to 20% but the first 2 years number might also be 25m so it's in the same ballpark.
The Google numbers have been quite widely reported and the source is probably Google themselves but the Apple numbers come from Apple.
the modifiability of Android and the struggle for the handset makers to differentiate themselves in the market may lead to even more variants, custom UIs, proprietary add-ons, and the like.
Maybe but one difference with Symbian is that Android has high profile apps and it acts against the distributors to break these apps so I think they will go with the stock Android as much as they can and simply add a skin to it.
Not true given I had to jump through hoops because Sense does something stupid with their soft keys. I haven't run into any blur landmines but I haven't tested any blur devices yet either.
I will give you that there will be some variance in regards to the OEMs. However, the list is nowhere near as massive to target as the OP wants us to believe.
Just for my knowledge, is it really that much more work?
Where did you get your numbers?
The Google numbers have been quite widely reported and the source is probably Google themselves but the Apple numbers come from Apple.
Straight from the Android dev blog:
http://developer.android.com/resourc...-versions.html
Android 1.5/1.6 accounts for barely over 25% of the total vanilla Android OS versions. It would be interesting to see a similar chart showing current iOS versions.
Maybe but one difference with Symbian is that Android has high profile apps and it acts against the distributors to break these apps so I think they will go with the stock Android as much as they can and simply add a skin to it.
I see what you're saying, but that's assuming that not breaking apps is considered a competitive advantage. With the number of Android phones being released, at the rate that their being released, can any substantial number of consumers really keep up with which handset is doing what and how, and how well?
It looks to me like the Android market is starting to resemble the old feature phone market, where bling and specs and advertising and a few high profile features might get a handset its day in the sun until it's supplanted by the next one. With 2 year contracts creating a constant supply of restless buyers looking for the next big upgrade, I think all a phone need do is hit a few functionality marks out of the box (web, email, media and IM) and most folks will reckon its good to go. It there turn out to be some compatibility issues down the road, well, that's just how phones are, as far as the average Android customer knows, right?
But again I think that that might create a kind of gradual, aggregate sense of quality problems for the platform, without most people being really very clear about where the problems are. Its open source roots and Google backing and general air of geekiness has given the impression that Android users are largely tech savvy and discerning about stuff like this, but now that it's blowing up mass market the average Android user is increasingly going to be completely clueless about app compatibility or fragmentation or getting and running the latest OS version or any of that. They'll just use the phone and if stuff doesn't work be bummed. If it doesn't work enough times for enough people, then you start getting a reputation, no matter who is at fault. I think that's one thing Apple is keenly aware of, and they go to great lengths to avoid those kind of scenarios.
So then, as you say, one might sake "But won't handset manufacturers make sure they maintain compatibility to avoid exactly that?", but they're playing a cell phone game, not a computer platform game. In that game you release phones at a breakneck pace, each one "packed" with specs, and run ads with laser wielding robots and flying cars and rain slicked streets erupting into chaos. It's all about driving excitement for the next group of off-contract people looking to upgrade, and there's always another group of them. "App compatibility" doesn't much matter to that mindset, as long as main apps people mostly use are fine-- which is what, about 6?
I will give you that there will be some variance in regards to the OEMs. However, the list is nowhere near as massive to target as the OP wants us to believe.
Just for my knowledge, is it really that much more work?
Yes and no. Yes, in as much as published APIs break and you need to find a work around that works across all variants. Sometimes you find out because a user tells you which is not good. No, in that typically the work around isn't that much work...it's just stupid and generally an ugly hack.
It just seems that the number of ugly little hacks in Android development is moderately high.