Android Compatibility and Fragmentation
tbray writes "Here are the details on the Android Compatibility Program — which combines the source, a formal compatibility spec, an open-source test suite, and access to the Android Market as reward for good behavior (program page). People like to rant about the subject of fragmentation, so here's TFM that they should be R'ing first."
* Bugs - devices might simply have bugs, such as a buggy Bluetooth driver or an incorrectly implemented GPS API.
* Missing components - devices might omit hardware (such as a camera) that apps expect, and attempt to "fake" or stub out the corresponding API.
* Added or altered APIs - devices might add or alter APIs that aren't part of standard Android. Done correctly this is innovation; done poorly and it's "embrace and extend".
Each of these is an example of something that can make an app not run properly on a device. They might run, but they won't run properly. These are the things that I spend my time preventing.
The only thing I might add is that devices of different resolutions can be annoying, especially if your app has static images or custom widgets.
Qxe4
I think there's only two right now - HVGA (320x480), and WVGA (480x800), though there may also be VGA sized screens as well.
The big issue is that the density increases but the screen size remains the same, so if your app isn't DPI aware things get small and hard to control. Desktop app developers tend to be fixed DPI - a larger window lets you show more information. Ditto web developers. But high-DPI displays means you don't want to show more information, but you should scale everything up. Even pictures if it makes the picture grow from literal thumbnail to larger blob.
DPI-awareness is a difficult thing and many apps still get it wrong on the desktop, if you switch your Windows desktop to high-DPI mode.
"The thing is, nobody ever defined “fragmentation”"
Let me try: You have several different versions of Android 'in the wild' on different phones, different carriers, etc. There are different stances on whether these version of Android can be updated (based on manufacturer) etc. yadda yadda
Now, looking at that situation, I would say 'fragmentation' is more along the lines of 'Is it going to remain easy for to target Android phones in general considering how many versions currently exist [/not obsolete] concurrently?'
So yes, it is mainly about compatibility. But it also means (much like any other platform) if the version leaping continues (and so many versions exist concurrently all the time) playing to the 'lowest common denominator' of supported features will be required
The hardware vs onscreen keyboard does not require two apps.
The iphone is not one platform, there are 3 different phones and there is about to be a forth.
Which you can just treat as 480x800, who will really notice the missing 54 pixels?
You will have either fragmentation, or stagnation as the device gets older. Even something totally controlled like the iPhone. As time goes on, one of three things will happen:
1) Apple will introduce new iPhones with features the old ones do not, and cannot, have such as higher rez screens, faster CPUs, etc. Software for these new phones will not run properly, if at all on the old ones. The phones will fragment along the lines of new and old.
2) Apple will refuse to introduce any features that would break compatibility with older phones. They maintain total compatibility through keeping various things at one spec. This leads to stagnation of increasing proportions as time goes on, such that new iPhones are literally years behind competing products.
3) Apple discontinues all support for older iPhones. They have the service providers remotely kill the hardware so you are required to purchase new hardware.
There's just no way around it. If you want new devices to have new capabilities, well it will lead to fragmentation. There are plenty of things you can do to help, particularly in making sure newer devices can seamlessly run older software, but you can't have all sorts of new stuff and have everything work just like it did before.
It's not 54, it's 54*480 which is 25920. Which is a non trivial amount on a device that small.
It's funny, I saw this article on my RSS and knew that the first post would be this. You don't give any reason as to why the fragmentation stories are FUD.
I am personally very hesitant to look at any android device due to the fragmentation. Hell, you have 1.6 devices being released alongside 2.1 devices. If this isn't fragmentation then what is?
As has been posted, read some comments from users about how pissed they are that their 1.6 device won't run certain apps or is lacking features that could be implemented by a 2.x release but their carrier won't deploy any updates.
While the fragmentation can't be squarely put at Google's feet, there's a shared responsibility between the hardware manufacturers, the carriers and Google to ensure that this doesn't happen. Unfortunately this hasn't happened and Android is headed squarely toward a cluster fuck.
I don't think every piece of hardware should be supported forever, but I think a year of official support for the latest version releases after the last date you sold the product is a minimum
If the contract period is 2 years, and the early termination fee is based on the contract period, they better support the phones for 2 years.
Even now lots of web developers want to treat the browser view like a physical sheet of paper, demanding iron-fisted control over placement of everything on the page. This leads to all kinds of work to maintain the control. I keep wishing more web developers would get over this. Not only would I find it a lot easier to actually make use of their work on smaller screens (my Android phone, my netbook...), I think in the end it would be a lot easier for the developers, too.
It sounds like mobile application developers often have the same problem...
Hacker Public Radio is our Friend
Android Market makes sure your app is only visible to those devices where it will run correctly, by filtering your app from devices which don't have the features you listed.
I've seen this a lot with friends who have different android phones. Friend A with HTC: "Hey try this really cool app". Friend B with Motorola: "What App? I can't find it anywhere in the market place." My understanding is that the app doesn't even show up if their phone is not compatible, it's invisible. I guess they don't want a bunch of apps that when you bring up the page says, "Sorry your phone is not compatible".
I also know from the QA side of the house that you can have the same phone hardware with 2 different OSs depending on the carrier. Usually we don't have major problems, but there can be glitches and bugs that are introduced. And then there is the speed of the upgrades. With the iPhone we looked at one major update per year. When we started working with Android last year it was 1.5 (which had been our for a while on the G1) and 1.6. Then suddenly it was 2.0, then 2.1, now 2.2 all the last 9 months or so. Apple has generally release a new OS every year with a minor .1 release shortly there after. Same with blackberry.
"The problem with socialism is eventually you run out of other people's money" - Thatcher.
The stories just say that fragmentation is killing Android, but they provide no evidence. They are making the claims, here. I have found no evidence that there's a fragmentation problem. There are apps that don't work for all devices, but iPhone OS has that issue, too, and it's only going to get worse.
Agreed, however the killing of the platform has started in minor ways, unless there's collaboration between carrier/manufacturer/Google it will end up as a platform that users don't want a bar of.
Let's face it, we're not talking about techies who can handle that some apps don't work, we're talking about users. They don't know the difference between a 1.6 & 2.1 device, they just see a phone with Android across the box. If it doesn't work they'll never buy another Android phone.
Sometimes it is difficult to step back and ask "What would a user do in this situation?". It's something that the IT industry seems to forget, that the users are what matters for a consumer device.
Most of the phones that run 1.5 right now are terribly underpowered -- OpenGL on a G1 is almost a sick joke.
If you're targeting OpenGL, you probably should cut your losses and cut them.
The Twitter application is an Android showpiece app, which is why it targets 2.1. They wanted to use animated wallpaper, quick contact bars, and so on, to highlight the best of the contemporary platform. Aside from the fact that about 50% of Android phones are running 2.1 right now, most other phones are going to see a 2.1 upgrade in the relatively short term. I suspect Google intentionally targets 2.1 to try to motivate the vendors to expedite their upgrades.
Engadget has been stoking the FUD fire for a long time. Do you own research and find out which devices are slated for upgrade. Practically the only one not slated for upgrade is the G1.
Engadget is toeing that sycophant line dutifully. I think I have seen 3 articles with them calling Android horribly hopelessly fragmented in just the past two or three months. It's ridiculous. They are sensationalizing and anger mongering.
Oh for.. look, how do you propose to make e.g. a GPS tracking app run on a device that doesn't have GPS? (yes, I'm aware all Android devices must have GPS) How is allowing apps to use specialised hardware on Android any worse than allowing apps to use the compass on an iPhone 3GS, but not on a 3G? or camera apps on an iPhone but not an iPod?
The only way to avoid fragmentation as you define it is to have one unchanging, stagnant piece of hardware that only runs one version of the OS, on one network from one vendor. Even Apple doesn't do that.
Why would anyone engrave "Elbereth"?
Indeed. There have been several occasions on which I was forced to upgrade the OS on my iPod Touch, just to use a new app. ("Forced" as in: The app would not bloody install on the software version I was running.)
And, of course: On the iPod Touch, OS upgrades sometimes cost actual money.
So. If we assume that fragmentation is a problem, then it is very plain that it is not an Android-specific problem.
Kid-proof tablet..
"It's funny, I saw this article on my RSS and knew that the first post would be this. You don't give any reason as to why the fragmentation stories are FUD."
It's actually been explained, and I have to question the motive for your posting, because it takes a certain selective ignorance to not see the given reasons. All the same, to give you the benefit of the doubt, here are the reasons they are FUD:
1) There's a choice between fragmentation and stagnation, if a set of devices does not change over time to avoid fragmentation, they rapidly become outdated.
2) Fragmentation has been fundamental to the most successful computing platform of all time- the PC. Without fragmentation there would've not been such rapid development or adoption.
3) Because of the above, fragmentation has allowed the likes of PC to be extremely versatile, and software developers have coped with it for decades without any real problems.
4) Because fragmentation has been so fundamental to the development of the PC, and dealing with it is so well understood, Android has been able to be designed with the issue of fragmentation in mind, and hence handles it in an extremely graceful manner- more so than most other platforms.
5) Because of point 4, the supposed issue of fragmentation hasn't actually really caused any problems, hence why the Android marketplace has some of the most innovative apps on the market (particularly those from Google such as Skymap, Translate, as well as the likes of Layar) which work perfectly fine even on my old HTC Magic running Android 1.6.
So in other words, despite all the fragmentation FUD being talked about, there's no actual evidence being produced of it causing any real life problems. The only whines seem to be from those who simply don't have the first clue about software development because professional developers understand concepts like abstraction layers and how they help you deal with differences in platforms.
I hope this answers it for you once and for all, but judging by your tone it sounds like your mind is made up, to you it sounds like Android has a fragmentation problem because a few blog posts told you so, even though they provided no evidence at all to back up that theory, let alone any "evidence" that can't easily and demonstrably be refuted as being FUD in itself.
This is my thoughts exactly.
You're gaining a huge install base by having Android compatible with multiple handsets. People like choosing different feature sets, not everyone wants a keyboard for example, but some wouldn't live without one. Some people will pay extra for a phone with excellent gaming performance, others won't care.
This is akin to figuring out if a hypothetical game will run on User X's computer. Ship the game for a console if you don't want to deal with it.
To be fair, for the wide variety of devices Android supports, there are very very few compatibility glitches.
- Michael T. Babcock (Yes, I blog)