Most developers are coding for Google's Android instead of Apple's IOS | Hacker News Reader

Most developers are coding for Google's Android instead of Apple's IOS

theinquirer.net

62 pointsbakbak64 comments

That's very likely a simple function of cost: You can't do iOS programming without the (relatively expensive) mac hardware.

Also, the actual report is here:

http://www.visionmobile.com/rsc/researchreports/VisionMobile...

Other points from the (CC/BY) report:

Platform revenue potential. Not all platforms are born with equal revenue potential. Our research revealed large discrepancies across platforms in terms of the revenues applications are bringing to developers. iOS topped the chart, making 3.3x more money per app than Symbian developers followed by Java ME (2.7x) and BlackBerry (2.4x). Android (1.7x), mobile web (1.6x) were the weakest performing platforms in terms of revenue per app and only ahead of Symbian (1.0).

....

App store fragmentation is an under-hyped challenge for developers. Each of the fifty-plus app stores available has its own developer sign-up, app submission process, artwork and paperwork requirements, app certification and approval criteria, revenue model options, payment terms, taxation and settlement terms. The marginal cost of distributing an application through one more app store is significant, contrary to popular perception.

....

Platform priorities For companies going mobile, platform priorities are mixed, but the core challenge is common – market penetration and reach across the customer base. Organisations developing B2C apps (targeted at consumers) are extending their offering first Apple and then to Android, to mobile web, to BlackBerry and finally to Windows Phone 7. For B2B apps (applications paid by the corporate IT manager or CIO), HTML is already the platform of choice- not just for deployment on mobile web browsers, but also by converting HTML and JavaScript into native iPhone and Android apps using tools from companies such as Appcelerator, PhoneGap, RhoMobile and Sencha.

I think fear is involved, too. I know that's why I picked Android over iPod when I jumped in. (Disclaimer: I don't have anything ready for market yet... And may never.)

I started looking at Apple's policies and how apps would be rejected without any reason given. And that there was -no- other way to launch a product if Apple rejected it.

And then looked at Android and saw that even if Google rejected it, I could still get the app onto the device in other ways. (And then Amazon came out with their own Android market, making things even better in that regard.)

Actually, Apple will specifically tell you which rule(s) in the App Store terms and conditions that your application breaks. You can change your app and then resubmit, or appeal the decision (I had a little app rejected, appealed and it was approved about four days later)

They most certainly do not. I will not develop an app for Apple again and abandoned a "startup" based around app customization.

I developed an app with a partner, released it to App Store then took it down to upload it under a new account. Same app.

Apple rejected our app. They refused to explain by email why. They insisted on only dealing with me by phone. Over the phone they said the policies had changed in the last two weeks and my app would not be accepted. I asked which policy and they refused to say which one. I asked what I could do to get back in the app store and they said they couldn't tell me. I suggested some ideas and the Apple employee on the other end of the phone said that those would be "helpful" but they couldn't say that that would be enough to get my app listed again.

Time to cut and run. I cancelled both orders for customized versions of the app that I had taken from clients (and apologized profusely). I asked Apple for a refund on my developer account. They refused. I said that since the policies had changed then I was entitled to a refund because I bought the account under the old policies and if they'd like to change the terms and I don't agree then I'm entitled to have my money refunded. They said it was a "one time exception" but grudgingly refunded my developer account.

What did the app do?

CampaignPhone.com

Yes, but this process is infamous for being capricious, for rejecting apps for their basic purpose, for taking unexplained weeks to process, on and on. So yes, it worked for you, once.

Which is precisely my problem with it. You build an app (at cost) and then you can't sell it because the turf you're playing on has restrictions.

They need to offer:

1. Certified apps from the app store which are approved via this process. 2. Apps which you can just download and chuck on with no Apple or carrier approval.

Ironically just like the Mac App Store i.e. you don't have to go there.

Oh and like Windows for the last however many years it has been around.

I disagree. They shouldn't do that.

The price of dealing with the restrictions is justified by the exposure you get to a market willing to pay for goods. The few apps I've had that were rejected were done so justifiably, and when they weren't, we worked it out with them.

Most of this is hearsay, not fact.

If it was a free market, you'd be SOL and are defending it on that basis.

Do they? At the time I was looking into it, there were many complaints that they didn't. (To be fair, Android has had some of the same complaints, but as I noted, the Android Market isn't the only path for an Android app.)

I think they changed the policy at some point but it's been that way for at least a year. In my experience with them, they've been quite helpful.

Thanks for linking to the actual report. It is very long but worth skimming, as it covers issues and perspectives that most app developers don't even think about (for example, the pain of the Ovi 5 to 6 iterative review cycle process v. the mega downloads once it is on the Ovi store).

Like most people can't afford to buy a mac mini for development..

Have you ever tried to compile Xcode samples on a MacMini? I've had clients who do. White Macbooks (1k) are passable, but MBA is pretty much the real bottom line for any day to day development

Also, remember we're talking GLOBALLY. What's affordable to you may not be to random Indian programmer 7

2 things that make me think cost is the issue:

>Developer-market balance. Android is the one and only platform that is trilaterally adopted by developers across all three major continents active in application development - Europe, North America and Asia. On all other platforms, there is an imbalance of developer supply and market demand across the globe. iOS is lagging in developer mindshare in Asia while BlackBerry developers are almost completely lacking in Europe. The traditional sweet spot for Java developers has moved out of Europe to emerging markets, with 42% more respondents coming from Asia, Africa and South America. Flash Lite has weak supply in South East Asia where the platform can deliver best-in-class experiences on mass-market phones

And the fact you're highly likely to make little more than 700 back:

> In the gold rush to the applications economy, not everyone is making money. About a third of respondents make less than $1,000 USD per application in total, which is loss-making given that an application often takes months to develop

> Have you ever tried to compile Xcode samples on a MacMini?

Yes, I have been doing it on a four year old, Core Duo one for the whole three years since the iOS SDK became available and have had zero compilation performance issues. Keep in mind that the vast majority of iOS applications are laughably small compared to desktop apps.

I'm not disputing that a Mini is still too costly for many, but I think it is unfair to exclude it as development machine.

This barrier is quickly going away as well.

Mac OS 10.7 (Lion) is virtualizable. I'm sure someone will build a rack and offer cloud-based Lion accounts for developers that can be remoted into 24/7 so long as you have internet access.

I'm a fulltime iOS developer. Standing there waiting for it to compile was a considerable length affair (multiple minutes)

2-screen iPhone apps with no media might compile quick enough, but development of the type of apps which make money that would be a huge productivity suck.

So invest the money first, or put up with it and use the proceeds to buy something better.

"I don't have the money to spend so making money is faster and easier" is not a problem unique to iOS development.

Today's Mac mini's are more than powerful enough to write code on. But the gist of your argument seems to be that the cost of buying a mac seems to be a big barrier to entry in some markets.

While I would agree with that to some degree, the fact that there is over 400k apps in the appstore tells me that the barrier is either non existent or a lot lower than people are making it out to be.

Furthermore, there are other costs besides the computer itself when developing for iOS. So if the developer isn't making enough money to cover all of those costs, the that developer needs to try something else to make money.

Which isn't to say "not to do iOS development" but to find out why their app isn't selling and adjust. The "build it and they will come" is a fallacy and many developers fall into that trap.

> the fact that there is over 400k apps in the appstore tells me that the barrier is either non existent or a lot lower than people are making it out to be.

Barriers leading to higher numbers of hobby users trying android app dev than iOS, which is the title of the HN post.

There are more pros on iOS and Win7Phone from the survey (both numerically and percentage wise).

I agree with the gist of your statement, but a if an Air is good enough for iOS development (and there are a few blog posts out there that say it is), then a baseline Mini will be fine.

That SSD is a huge differentiation.

It's amazing how much that updates compile speeds.

Buying a mac mini to do iOS development is buying pain and a computer you're just going to have to replace.

I'm curious why you wouldn't just buy a small SSD (i.e. 40GB) and put it in the mini yourself?

That might make it work. I think that's the big differential between MBA and mini. I'd be curious to see that build some various app types.

I don't know of anyone putting that forth as the low cost solution (nor is that sort of change out easy for someone who's really terrified to break their $700 computer and not used to fiddling with that stuff).

Current Mac Mini's solder the cable to the drive and it's not swappable. (It's like the new iMac's drives -- which have apple specific firmware and thermal systems, but worse.)

> Current Mac Mini's solder the cable to the drive

ifixit.com seems to indicate otherwise: http://www.ifixit.com/Guide/Repair/Installing-Mac-mini-Model...

I stand correct. My friend went to an Apple certified tech to get her laptop HD swapped with her Mac Mini HD, because they wanted more of their music on the laptop because they traveled often, and the Apple tech couldn't do it. Issue with the Mac Mini and how it makes the HD a non-user servicable part.

Have you ever tried to compile Xcode samples on a MacMini?

I do it on a 6 year old Macbook and have no problems.

Affordability isn't really the reason, though.

Take the class of devs who, in their spare time, want to learn mobile programming. If they don't have a Mac, their choices tonight after work are to 1) sit down on their own familiar computer right now and start Android dev, or 2) go buy a Mac and a dev account and get familiar with that and start iOS dev. Paying $800+ for option 2 doesn't exactly sweeten the deal.

A modest app in the store will pay back the costs of the hardware, even though it probably won't pay for your time. Doing iOS work for hire will pay for your time, however.

So people should do what they want, but it's easy to see why Android would be the favorite of new casual mobile devs, many (or most?) of whom don't have Macs.

Good points.. it would be the same for a person who wants to get into programming but doesn't own a computer.. they still have to get the cash from somewhere to buy the equipment.

Many years ago, I was in this exact situation. I planned accordingly and saved until I could buy a cheaply made desktop and taught myself how to code.

You're sure your app will make at least $700 to recoup those costs? I like your optimism.

A half decent app with only thousands of users might get that back in maybe 6 months.

I know I did and I wrote a very niche app and did no promotion what so ever. I have not made much on my app but it has paid for a couple of gadgets so far and I am still making a (very) little bit every month.

Lots of people aren't writing that. They don't have the skills to write that on Android either. But they're still counted in that survey as "Android Developers", yielding that HN title.

Maybe not, but my cost of materials are rolled into cost of the development I do for my clients.

Lots of people aren't writing apps for third parties. As a third party dev myself, it's significantly easier to justify costs (even to people like my wife) than if I did self-published apps that were hobbling along

That's true.. I just find that owning a mac to be in the cost of doing business. That just maybe outside of the means of the average hobbyist developer.

I suspect a lot of these people are in the OSS hacker demographic (in my mind this means a college student running linux with lots of free time), in which case, why would you bother getting a Mac?

It appears lots of Android devs are exactly in the category you're talking about:

On experience of developers:

We can also quantify the signs of Apple’s allure towards experienced PC and Internet developers, since the iOS platform attracts significantly more developers with sevenplus years PC/Internet experience, compared to other platforms. This confirms that experienced software developers are moving into mobile, using iOS as an entry platform, in what we believe is driven by the sudden rise in demand for developer talent, especially in North America. Since launching in late 2010, Windows Phone 7 has done pretty well in attracting seasoned developers. We see experienced mobile developers coming to Windows Phone, with a significant bias of current Windows Phone developers having between three and six years of mobile experience - an indication that Microsoft’s strategy to tap into PC and Xbox developer segments is paying of

How much credibility does this have? BlueVia is sells a product for developers that targets Android and not iOS, while VisionMobile is a market analysis company they presumably paid to produce this study. I'd never heard of either before, and it sounds like they have a substantial financial interest in presenting Android as more interesting to developers.

I agree with everything you say but I wanted to add one thing. I'm so sick of these surveys being published without any methodology information. Methodology can very easily impact a survey. We see it all the time in political surveys where they seem to indicate something until you look deeper and find their sample was skewed towards a political persuasion.

I mean, does anyone doubt that Open vs. Proprietary can be just as political as Democrat vs Republican in tech circles?

I agree completely. At the least they could have included a link to the source.

The iPhone developers, the good ones (e.g. have been Featured in App Store) that I am fortunate to know personally, they're not even touching Android.

Reasons: Unfamiliarity and leaving the comfort zone.

iOS development is comfortable. When you do Android dev, you give up a lot of comfort for raw, unfettered access. It is a rule of thumb when you develop for Android that when you release you will get ANR (Application Not Responding) simply because of the fact that you can't test on more than a few actual devices. I get crash reports from the market site that I can't fathom how they even can happen (stuff like libraries that are statically linked not being found)

And, as an primarily Android dev now, can you imagine giving up Core Data (for all its faults, the latest iOS 5 [redacted] makes it better), Interface Builder (ADT 11 is getting closer but still not close) and even stuff like Objective-C Categories?

On fragmentation:

We were able to quantify that indeed Apple’s iOS is the platform with the least fragmentation (on average four versions per app), as has been widely noted from empirical evidence in the past. Apple manages fragmentation through two primary means: first, it has standardised the screen size and resolution for its handsets and tablets; and second, as an OEM and platform vendor, it has commercially streamlined the means by which most iPhone or iPad users are upgraded to the latest OS version.

In contrast, our research indicates that Android developers must create six versions of their apps on average, which is on par with mobile web apps. The stark difference in fragmentation across Apple and Android devices is also evident amongst the different platform versions in the installed base of devices. According to Google data released on May 2011, 25% of active Android handsets run on platform versions more than 18 months out of date. Meanwhile, according to app analytics firm Localytics, only 20% of existing Apple 3GS devices had not yet been upgraded, just two months after the introduction of iOS4. In other words, Apple devices have the youngest runtime age in the mobile industry

> In contrast, our research indicates that Android developers must create six versions of their apps on average.

That part makes no sense to me. I've never seen a single app in the Android Market that has 6 different versions. And it would probably be obvious. The Market will hide apps that aren't compatible with your device, but each app still has to have a separate, unique name.

The most versions of a single app I've been able to find is TouchDown, which has 3 versions: 1 for Android 1.x devices, 1 for Android 2.x devices, and 1 for MobileIron customers. The third version arguably would still exist even if there were only a single Android handset. http://www.appbrain.com/browse/dev/NitroDesk%2C+Inc.

Anyone have a concrete example of an Android app with 6 versions in the Market?

They're probably also counting the bastardized content pack thingy many people used instead of In-App purchase where you buy/download other items on the store for levels or whatnot.

I enjoy how he rambles on about how Java is dying, and that it isn't secure, but earlier says that 67% of mobile developers are working on Android applications.

Indeed, I think the author needs to rephrase 'Java' as 'Java ME' for a lot of that.

Misleading headline. It implies that >50% of developers are coding for Android exclusively, but that's not the case. It would be just as (in)accurate to say that "Most developers are coding for IOS."

"...67 per cent of developers code for Android..." I'm a little confused. 67% of all developers? 67% of mobile developers?

Should be of "mobile developers", upon reading the article. But nonetheless, the numbers are a bit pointless since they don't indicate overlap (gosh, you can develop both iOS and Android?). What would be interesting is to know the share of "original" development (not "just" porting - but apologies for the generalization) on platforms, or if it is done multi-platform from the start.

Yeah, I guess you could read it as "33% of mobile developers completely ignore Andriod"

Haven't read it yet, but here's a direct link to the report: http://www.visionmobile.com/rsc/researchreports/VisionMobile...

FYI for those on mobile/capped access, it comes in at 15 MB.

"Entitled Developer Economics 2011, the study shows that 67 per cent of developers code for Android, up from 59 per cent in 2010. IOS lags behind at 59 per cent, but that's also up from 50 per cent last year, most likely due to Apple's introduction of the iPad."

Last time I checked 67% + 59% != 100%

Well, I suppose developers could develop for more than one platform.

Which of course means that the "instead of" in the title is misleading. I mean, 67% vs 59% warrants a "most" and "instead of"? What a crap article.

The title didn't reflect that hence my snakiness about the numbers

I am amazed that we've got 56 comments on this thread, and nobody's brought up market share.

Android has been beating the bejeezus out of iOS on unit sales for a while now, both in the U.S. and globally. Even more importantly, Android's growing at a faster rate.

Developers will gravitate towards the platform with the most users. Period. This has happened time and time again over the past 30 years. IBM, Microsoft, Oracle, Adobe, Google...

There's a lot of nuance underneath the headline numbers, and certain communities (like HN) will be friendlier to iOS. However, in general I would expect Android to capture the bulk of developer attention going forward.

Developers will gravitate to what makes them money. If it was marketshare then more developers would be writing apps for Symbian than for iOS or Android.

Who says they haven't been? They certainly were in markets where Symbian was dominant. When I was in China (2008-2009), the developer support for Symbian was dramatically better than BlackBerry or iOS.

Now that Android has exploded in that market, I expect most of those developers to migrate away from Symbian to Android if they haven't already.

One thing to remember is that we're talking about the entire population of software developers. Certain demographics (like HN readers) will show up more heavily on certain platforms (like iOS). However, the general trends will always mirror marketshare.

Personally, I am building an app for the iPhone first because of public perception and marketing. It seems more people write and review iPhone apps versus Android apps - you can just tell when they start Android review articles with a tone of appeasement ie. "Ok, finally here's something for all you Android users.."

When someone finds your app and sees there's an iPhone version but no Android version, their normal reaction is: "That's understandable, Android is coming next, right?". If you have an Android version but no iPhone, their reaction will probably be "Hmm.. wait, that makes little sense...". It's just a side effect of iPhone being the first one to come up with app store.

The attitude towards android in general is dismissive. You need look no further than this HN thread. Imagine how different the posts would be if the numbers had broken in the other direction, in the iPhone's favor.

Well that might be all well and good if you want to be a programmer with the rest of the herd. But the most profitable app store is the Apple store. Personally I'm happy that there aren't so many Objective C programmers around.

I've been able to get by in Apple's app store with minimal knowledge of Objective C and a massive usage of C++. Thank God for *.mm files!

Reply on news.ycombinator.com