Shared publicly  -

(round two) I'm sure I'm not about to say anything that hasn't already been said before but here goes.

1. When you buy a PC, you are deciding at the point of purchase what level of PC you wish to buy. If you are a gamer, you buy a gaming machine that's capable of running those games at full capacity. You can buy a less spec'd machine and still play those games though you might discover parts of game play that aren't necessarily ideal. Or you could choose to buy a netbook that can't even play those games. That's the decision you make when you research and buy your PC. We are ok with that in the world of the PC. We understand that not all PCs are made equal and that, based on our purchase decision, we end up with a machine that is (or isn't) capable of a number of things.

2. Android is proving to be no different. There are a number of highly spec'd Android devices out that are capable of running "everything" and plenty of lesser spec'd devices that are ideal for things like email and sharing. As a consumer, you make the choice. Beyond that, it's entirely possible for you to buy or get for free a top notch device even a few months after it's release. The number of new handsets in the Android world mean our options are plentiful. Beyond that, as a consumer, you don't normally have to struggle with app incompatibilities now that the Play Store allows developers to filter their apps so they only show up for those devices that they are compatible with in the first place.

3. Further, I've tested a good number of Android devices recently and one thing I've noticed on all of them is that this oft regurgitated notion of incompatible apps is much less of an issue in reality than it is in pontification. It's simply assumed that with a higher number of devices, the compatibility of apps will take a hit as developers struggle to make their app work on all three-thousand-bagillion devices (doesn't that sound like an impressive amount of work?) Yes, developers tweak their apps to better target a large number of devices, but lucky for you, many times they target those devices that best represent the landscape as a whole. In my testing, I can't even tell you of an app that I've loaded and found later to be incompatible with the device I'm using. It RARELY happens if ever!

And if you are still running a device that is more than two years old, expecting it to hold suit with the latest and greatest devices, then I'm sorry to say that right now, it's incumbent upon you to update your hardware IF it's important to you to be able to keep up. All hardware gets outdated eventually. So if it's important, that's what it takes to keep in the game. Maybe someday the improvements to hardware will slow to the point where most handsets are similarly spec'd for a reasonable cost at which time maybe holding onto devices for longer than 2 years will be more realistic, but sad to say we aren't there yet. What is important to you? Guide your purchase decision based on that.

4. The coverage of this study seems to merely mention the custom ROMs that make up 39% of the total number of devices found in this report. First it's important to point out that this doesn't mean that 39% of their user base is running custom ROMs but rather 39% of the DEVICES reported represent these one-of ROM builds. Also, people need to understand something about the ROM community. Those who ROM (including myself) take a device into their own hands when doing so. It's part of the nature of rooting and ROMing your device. When doing so, personally, I feel like any downsides that accompany MY choice to do that (ie app incompatibility) are MY problem, not yours. So developers, don't worry about that percentage of folks who choose to take their device into their own hands. You will never be able to keep up with this infinitely growing list of custom ROMs, and as far as I'm concerned, its ok if MY choice to do so results in my own inability to run your app. It's me, not you. We are outliers in the grand scheme of things, and to revolve this kind of "fragmentation" discussion around including those who ROM just seems kind of unfair to me. Who tracks the Jailbreak community in iOS coverage? You just don't see it. Cause its an outlier and really doesn't represent the majority of the user base.

5. One final #^%&^# in memory of the first draft of this post that was infinitely better written than the one you just read. I'll put my hat into the ring for G+ to gain an auto-save/drafts functionality to help idiots like me from erasing my post before I have a chance to share it.
Liam Cox's profile photoDarryl Drury's profile photoDan Shernicoff's profile photoBrian Wechtenhiser's profile photo
the whole custom ROMs being included in this chart is ridiculous. Agree 100%, Jason. You CANNOT include custom ROMs in any Android 'fragmentation' discussion
I'd like to add that over 90% of the complaints we get about our apps are on Samsung devices. That is despite us testing our apps on several handsets by Samsung. As far as I know similar incompatibility issues don't plague other manufacturers.
+Lars Fosdal but they're not distinct API levels that are different than the AOSP base they're made from. If they're the same API level then they should be included with those.
+Gina Trapani said it best on This Week in Google: the "fragmentation" issue is generally better for the consumer, giving more choices to fit the exact needs and wishes of each, and worse for the developer. And, given the choice between what is good for the consumer and what is good for the developer, we should err on the side of the consumer.
+Lars Fosdal Because they aren't SOLD with custom ROM's. You should only count what is being SOLD. Like Jason said, you don't count the hacking community in a graph or complaints about things not working.
I feel like a slow clap is in order.
Very well said...

If people dont want choice, feel free to go get an iphone... Where APPLE makes all the choices for you and tells you what YOU want...

Me? sticking to android...
As an app developer it is not that bad, I target my app for the Nexus S and the Galaxy Nexus and make sure they work in both, that covers me for pretty much all phones 2.3 and then I test on the Galaxy Nexus which covers me for 4.0. If the app is for low end phones I like to test on the IDEOS U8150 which is the most low end 2.2 phone in existence. I also create each graphic multiple times for each screen resolution (it is really easy as Google has a webpage to do it for you). It may sound like a lot but android is easier to develop for then most of you think.
Well done +Jason Howell and I hope you discuss this on Monday. As much as I love the TWIT network Android news items covered on the other programs (except TWIG because of the awesome +Gina Trapani ) are not discussed fully, especially on TNT when you are not there as the guests, 9 out of 10 times have an iphone and tend not to have a grasp on Android. So I count on All About Android for a full discussion.
As a game dev, I agree with +Nick Webster, fragmentation isn't an issue really. If I add some custom device specific code/art, it's because I already have platform independent stuff, and am just tweaking things to add polish. It's essentially the same as when I author (cross platform) games on PCs -- Actually, its easier because there's only one OS API to code to. Closed beta helps discover & work out the special kinks on different platforms, but for the most part we just test on a few devices for the different approx hardware speed & OS version. Also, targeting an older version of the API typically means that version and higher of Android will run it fine too.
Oh, I'd just like to add -- The ROM community is great. They typically have good support out of the box and are quick to fix & accept patches when things break (yes, us devs can FIX the platform in the ROM community). In my experience, It's the crazy UI mods that MFGs and carriers like to add that are usually the cause my platform incompatibility issues, but YMMV.
Very well said +Jason Howell . It's nice to hear an educated viewpoint addressing this article which, in my opinion, has just been thrown around by folks who haven't actually read it.
Where's the harsh criticism for automobile fragmentation? I can't remember which tech writer first introduced this analogy, but I think it's very applicable. I know there is no real app market equivalent for automobiles, but in terms of usability I think the comparison still holds.

Cars come in all different sizes and shapes with different engines and drivetrains. They come with manual and automatic transmissions, front wheel, rear wheel and all wheel drive. From one car to the next, things are never in the same location; wiper controls, stereo buttons, gas caps, shifters, trunk and hood releases, climate control knobs, you name it. But all of these differences don't make automobiles "fragmented" or unusable, it makes them unique. People don't complain about these differences, they seek them out. Whether you want to drive a rusty 40 year old pick up truck or a brand new cutting edge super car, a pocket sized hybrid or a super sized gas guzzling SUV, you as the consumer have options.

Android offers the same kind of options. What is the alternative? A single device from a single manufacturer? I think I'll stick with the "problem" of fragmentation, aka choice.
I think the second draft is quite good.
If we could throw out that troublesome word "budget", I would have no problem replacing my Viewsonic G Tablet with something more modern. But it still does about 80% or more of what I need to do. And I keep finding apps and hacks that will do a specific task as well as any high priced replacement. Just can't do it by over-the-air update...
I will put up with my Droid X until Verizon shows its hand on cancelling unlimited data plans. May need to change providers!
Why can't people just accept fragmentation as a result of such an open platform? Instead of apologizing and taking cheap shots at the competition, embrace it and leverage good design to make it go away from a user standpoint. So far it seems Android fans just want to make it seem like they're superior. Bullshit...make superior stuff first, then claim the prize.
What a well written post! I hope you talk about this on +All About Android
I see no major issue with fragmentation. Especially if I'm installing a custom ROM anyway.

P. S. I agree on the auto save feature. I've started writing my long posts in notepad then cut and paste them into Google +.
I would like to see how this pans out for Apple and AT&T now that they are selling the 3GS (a 3 year old phone) for free under a 2-year contract. This is, in essence, what +Jason Howell is talking about. They are going to have to deal with fragmentation even more than they do now.

On a related note, I was just talking to a coworker about her antiquated Android phone (Motorola FlipOut - and how she might switch to iOS because Android is awful. I explained to her that the phone was low end when it debuted, even worse when she bought it, so comparing it to an iPhone (i.e. - high end) is not the same thing. Sorry to say for her, but it's still on Eclair (2.1), which is another issue entirely.

+Roy Walter, there are superior Android phones out there (Nexus series, Samsung Galaxy S I/II/III, HTC One Series, Motorola Droid OG-4, etc). There are also many not-so-superior phones (Motorola FlipOut, Pantec Pocket, HTC Status "The Facebook Phone", LG Ally, etc). Basically, if you choose a smartphone for less than $100 on contract, you are buying a sub-par phone, and shouldn't complain that you don't/can't have the latest and greatest work properly. This now holds true for Apple, as I mentioned above.

Ultimately, you get what you pay for (even an iPhone) when you decide to buy a low end phone (< ~$100).

I have a VZW Samsung Galaxy Nexus running CM9 and a firesale HP TouchPad (dual booting webOS/CM9). My wife has a Samsung Stratosphere (stock 2.3.7 w/ physical keyboard) and a Samsung Galaxy Tab 2 7.0 (stock with 4.0.4). And all this because of fragmentation choice!!!
Oh, and I forgot to mention that if the WSJ is correct in saying that the next iPhone will have a > 4" screen, then what will app developers have to do to make sure there app can run correctly on all these different screen sizes.
The original galaxy s is still a great phone 2 years later. I would use it stock with GB but mine is cm9'd up and it runs all the apps and games me and my friends like to use with no problem. I don't plan on upgrading until next year.
I knew all the action was on Google+! I agree with most of your points Jason. However the PC compared to Android argument doesn't totally make sense to me.

For instance, if I buy a PC today, I'll choose something thats up to what I want spec wise. However when I want to upgrade to windows 8 later next year I don't have to wait for Dell, or HP to give me permission to upgrade. I can upgrade not questions asked, the drivers are mostly publicly available and I'll be up and running in no time.

With the Android platform you choose your phone, with nice specs etc. However you don't know if that game next year will still run on your OS, and if it does will my phone be upgraded. I can't go to the store and buy ICS and install it, and get the drivers through the individual companies, like from broadcom.

Instead I have to wait 5 months. If I even get the update. Then they'll have to write the update and make it compatible with that specific phone, get some certification done, send it to google for a check up, fix the things google complained about. Send it to the carriers, hope my carrier will approve of the update, etc etc. Sony made a great blog post about this I believe.

Luckily we are getting options like the nexus phone which has the images right on googles website and of course the ever so great cynagonmod.

But that isn't a real alternative to the average ATT smartphone consumer.

And I truly believe that if you make this whole upgrade process more transparent and faster everybody will be on one of the latest OS versions, which takes allot of problems away for developers. (besides the different aspect ratios and such). But a big problem right now is the broad spectrum of Android versions in use right now.

So yeah, I hope android because more of a PC in that sense that google says ICS is coming in July, and I can get it for my phone in July.
again +CNET with spreading FUD . +CNET go back to your little puny boring iphone..maybe one day you may get lucky and call it the cPhone
It might be nice of I could read the chart. It is all blurry on my screen
Is it just me it does the g+ app for android (on a HTC desire at least) really reduce the resolution/quality of pictures so much that pictures like this are unreadable. Zooming in just gives unreadable blur.
For 5, I highly recommend checking out Lazarus: Form Recovery ( I use it on both Firefox and Chrome and it keeps me out of trouble when text fields lose the text I typed in one way or another (inadvertent navigating away, deletion, browser crash, etc.).

Your second draft is still good. I completely agree with everything you said.
The fragmentation argument is fragmented.
well do to fragmentation of android i OS but no one with other phones will not support for their phones. I get reports that the app does not work on their phones but i can not help out not enough cash. With over 230,000 active users i does work on some phones
Hey +Jason Howell I've just found this A great way to test your app as a developer on over 170 phone platforms. So the "fragmentation factor" just diminished a level. Where there's a will, there's a way
Add a comment...