Or you may not realize that a control layout that made perfect sense for your test cases perhaps doesn't work as well in the real world when you are jumping in and out of the app, etc.
Or... Android is so fragmented you can't possibly test the hundreds of different models with different OS versions and different screen sizes, memory, cpu, etc. Switching to one particular model with one particular OS version isn't going to solve problems in any meaningful way with respect to developing software for Android.
Or... Android is so fragmented you can't possibly test the hundreds of different models with different OS versions and different screen sizes, memory, cpu, etc. Switching to one particular model with one particular OS version isn't going to solve problems in any meaningful way with respect to developing software for Android.
Apparently your knowledge of app development is quite limited. The number of different screen sizes on iOS has also increased dramatically (iPhone 4, 5, 6, 6+, iPad mini, normal, pro). App development on both Android and iOS is more and more made in a resolution independant way (comparable to responsive web design for the web). I test iOS apps for iOS v7, v8 and v9, I test Android apps for v4.2, v4.4, v5.0, v6.0.
People interested in installing new apps don't have old devices, so this basically covers the whole addressable market. There is no real difference in the number of test configurations between iOS and Android. Once an app is tested on the standard development device, an experienced developer can predict very accurately if an app will run fine on another configuration, unless one uses a very special new feature like force touch.
Not at all, I stated only facts.
The component and production price is known of the iPhone and the AWatch.
The retail price is also known.
fact - fact = fact (and sometimes a ripoff)
Not at all true. The component and production prices are NOT known. They are guessed at.
It shouldn't be about android specifically. It should be about older generation phones. Slower connection is a good target, but using current gen devices is foolish. Facebook is one of the most abysmally slow experiences on the web (along with several web-based "magazines" and "tech blogs") due to slow javascript etc. Try living with an iPhone 4 for a few months.
Apparently your knowledge of app development is quite limited.
Admittedly my comment was a bit of an exaggeration, but you are right, I only maintain one application and it is on Windows, however I do a lot of web development and I do have to test on all kinds of different devices, platforms and browsers so I probably test on more configurations than you do.
BTW have you heard of the programming environment PhoneGap? You can build simple apps using regular CSS and Javascript and it then compiles them for any of the phone platforms. It is a Mac app and it previews inside of an iPhone app. I tried it out and it certainly works. There are a lot of plug-in for it too. It is owned by Adobe.
I own a software company that develops apps for both Andrioid and iOS. In terms of development costs, Andtoid development is definitely less expensive. Test devices costs less, development machines cost less, development tools are better and more efficient (Android Studio just blows away Xcode) and licensing and validation costs less. Only Swift has the potential to become better than Java.
But this does not means that the business case is better. This largely depends on the app and the market you are in. The biggest share of my revenue is from developing apps for 3rd parties. In this context the TCO of developing Android apps is clearly better for me. Of course for the owner of the app the business case for iOS might be better if you live in a country with a high iOS marketshare.
I would dispute most of that. Firstly the cost of the Mac and iOS developer licence should be less than a few hours contracting. Maybe two. Xcode is free and I'm surprised to hear anybody think it much worse than Android Studio, though it's not perfect. And if you are developing for both iOS and Android your best option is a Mac. The savings are – you just need one machine.
As a developer who is contracted just to develop you probably don't see all the costs of testing. It isn't just the android fragmentation of the OS, which despite what you say is more relevant – if your client wanted something that only worked on iOS 8 and later you would still get 90% of customers – but the sheer number of machines you have to test for.
Of course many companies test on a small matrix of machines for Android, but as we can see from the very topic of this thread that's not really good enough – not for the low to mid end devices. If you are producing primarily for the US market you can get away with a few test devices, but not for a worldwide market.
You do know how Google counts active users as you already stated it: Unique Google Android devices that actually use the official Google Play Store are counted, and only once for each device ID. No Google services , no count. To be counted in the 1.4B unique active users Google mentions you had to have visited Google Play.
It says 1.4b Android, 1b Google Play. That suggests 400m using Android but not Google Play. Whatever the real figures are, there's no way that over 80% active share matches up with usage and revenue stats and the 1.4b that Google has put out certainly doesn't so people need to stop using the stats until they have something to back them up. You can't just make up 1.3 billion devices that may or may not exist. That's the entire population of China.
It says 1.4b Android, 1b Google Play. That suggests 400m using Android but not Google Play. Whatever the real figures are, there's no way that over 80% active share matches up with usage and revenue stats and the 1.4b that Google has put out certainly doesn't so people need to stop using the stats until they have something to back them up. You can't just make up 1.3 billion devices that may or may not exist. That's the entire population of China.
Because it doesn't immediately make sense to you then it must be made up? That's an interesting way of determining facts and fallacy I suppose.
Well by this rational, the staff shouldn't be allowed to use a flagship Android but a more typical unit that would be found outside the US/Europe. As kpluck noted, many run Kitkat not to mention less RAM (1.5GB or less) and have slower processors, lower res screens and cameras.
Naturally, Mr Cox should also switch so that he can also get the proper perspective. Somehow I doubt he uses a low-end phone.
It is simple, ~500 million iPhone users are sold by Facebook to advertisers. This income in turn is used to develop Android version for the remaining ~2 billion.
He's an idiot. I know of him. 80/20 rule - 80% of your revenue will come from 20% of your customers, the remaining 80% will be problem customers.
I would dispute most of that. Firstly the cost of the Mac and iOS developer licence should be less than a few hours contracting. Maybe two. Xcode is free and I'm surprised to hear anybody think it much worse than Android Studio, though it's not perfect. And if you are developing for both iOS and Android your best option is a Mac. The savings are – you just need one machine.
As a developer who is contracted just to develop you probably don't see all the costs of testing. It isn't just the android fragmentation of the OS, which despite what you say is more relevant – if your client wanted something that only worked on iOS 8 and later you would still get 90% of customers – but the sheer number of machines you have to test for.
Of course many companies test on a small matrix of machines for Android, but as we can see from the very topic of this thread that's not really good enough – not for the low to mid end devices. If you are producing primarily for the US market you can get away with a few test devices, but not for a worldwide market.
You can dispute what you want, but here are the fact and figures.
All my Android developers have a Dell XPS13 developer edition (Linux based), a Moto G phone and a Asus Memopad tablet, all my iOS developers have a similar equipped Macbook Air, an iPhone 5 or 6 and an iPad Mini 2. Cost difference 700 Euros. Android Studio is also free and really blows away Xcode in easy of use and number of integrated development tools. It was a great move by Google to hire the people of Jetbrains to make their IDE, I wish Apple did the same. My feeling, based on the figured I measure on the development floor, is that the productivity with Android Studio is at least 20% higher than with Xcode.
We also developing hybrid apps using cross platform Cordova, which is great for less CPU intensive apps, so no games. Cordova apps are developed on Android and then ported to iOS.
Once an app is developed it is taken over by the testing and validation team who has a battery of test devices. For iOS there are 7 test devices, covering all possible screen sizes (iPhone 4,5,6,6+, iPad mini, Air) and iOS version 7, 8 and 9 (I'll drop 7 in a few months). For Android I have 9 devices with different screen sizes testing Android 4.2, 4.4, 5.0 and 6.0. This covers 99% of all the users who actually install new apps.
The time won by testing only 7 iOS devices is more than compensated by the more efficient publishing procedure for Android and by the time consuming app validation by Apple. For an average app a full testing cycling takes 5 weeks for Apple and 4 weeks for Android.
Good point that most people in other countries especially, run older phones with older Android versions, so having employees by Android phones here will still result in having better phones than their customers. I'll take it further. The requirement will not be successful. No employee will agree to use, as their own phone, an android phone with poor functionality, as older devices must have.
Facebook is going to have to put some deep thinking into deciding what devices to be tested. Another factor is device location. No device is going to perform the same across multiple countries and regions. Their best bet is to hire locals across the world as their testers and perhaps members of their project teams. If Facebook is global in reach, so should their teams.
Comments
"Facebook forces some workers to switch from iPhone to Android..."
I'm thinking this constitutes "hostile work environment"...
Or you may not realize that a control layout that made perfect sense for your test cases perhaps doesn't work as well in the real world when you are jumping in and out of the app, etc.
Or... Android is so fragmented you can't possibly test the hundreds of different models with different OS versions and different screen sizes, memory, cpu, etc. Switching to one particular model with one particular OS version isn't going to solve problems in any meaningful way with respect to developing software for Android.
Quote:
Or... Android is so fragmented you can't possibly test the hundreds of different models with different OS versions and different screen sizes, memory, cpu, etc. Switching to one particular model with one particular OS version isn't going to solve problems in any meaningful way with respect to developing software for Android.
Apparently your knowledge of app development is quite limited. The number of different screen sizes on iOS has also increased dramatically (iPhone 4, 5, 6, 6+, iPad mini, normal, pro). App development on both Android and iOS is more and more made in a resolution independant way (comparable to responsive web design for the web). I test iOS apps for iOS v7, v8 and v9, I test Android apps for v4.2, v4.4, v5.0, v6.0.
People interested in installing new apps don't have old devices, so this basically covers the whole addressable market. There is no real difference in the number of test configurations between iOS and Android. Once an app is tested on the standard development device, an experienced developer can predict very accurately if an app will run fine on another configuration, unless one uses a very special new feature like force touch.
Not at all true. The component and production prices are NOT known. They are guessed at.
Apparently your knowledge of app development is quite limited.
Admittedly my comment was a bit of an exaggeration, but you are right, I only maintain one application and it is on Windows, however I do a lot of web development and I do have to test on all kinds of different devices, platforms and browsers so I probably test on more configurations than you do.
BTW have you heard of the programming environment PhoneGap? You can build simple apps using regular CSS and Javascript and it then compiles them for any of the phone platforms. It is a Mac app and it previews inside of an iPhone app. I tried it out and it certainly works. There are a lot of plug-in for it too. It is owned by Adobe.
This and the Bond movie thing, so much material to throw at a fandroid has been cropping up lately.
I would dispute most of that. Firstly the cost of the Mac and iOS developer licence should be less than a few hours contracting. Maybe two. Xcode is free and I'm surprised to hear anybody think it much worse than Android Studio, though it's not perfect. And if you are developing for both iOS and Android your best option is a Mac. The savings are – you just need one machine.
As a developer who is contracted just to develop you probably don't see all the costs of testing. It isn't just the android fragmentation of the OS, which despite what you say is more relevant – if your client wanted something that only worked on iOS 8 and later you would still get 90% of customers – but the sheer number of machines you have to test for.
Of course many companies test on a small matrix of machines for Android, but as we can see from the very topic of this thread that's not really good enough – not for the low to mid end devices. If you are producing primarily for the US market you can get away with a few test devices, but not for a worldwide market.
That doesn't seem to be what it's saying here:
http://www.theverge.com/2015/9/29/9409071/google-android-stats-users-downloads-sales
It says 1.4b Android, 1b Google Play. That suggests 400m using Android but not Google Play. Whatever the real figures are, there's no way that over 80% active share matches up with usage and revenue stats and the 1.4b that Google has put out certainly doesn't so people need to stop using the stats until they have something to back them up. You can't just make up 1.3 billion devices that may or may not exist. That's the entire population of China.
He's clearly wondering about the other stats that contradict these figures, Internet usage, revenue etc.
Naturally, Mr Cox should also switch so that he can also get the proper perspective. Somehow I doubt he uses a low-end phone.
The heading of this article should be: "chumming for snarks."
He's an idiot. I know of him. 80/20 rule - 80% of your revenue will come from 20% of your customers, the remaining 80% will be problem customers.
Yeh. No need for the average employee to be burdened with a crap phone for that though
Yes, making your employees less productive and mad, what a way to go. Unless they get the S6, then they're just frustrated, not mad...
I would dispute most of that. Firstly the cost of the Mac and iOS developer licence should be less than a few hours contracting. Maybe two. Xcode is free and I'm surprised to hear anybody think it much worse than Android Studio, though it's not perfect. And if you are developing for both iOS and Android your best option is a Mac. The savings are – you just need one machine.
As a developer who is contracted just to develop you probably don't see all the costs of testing. It isn't just the android fragmentation of the OS, which despite what you say is more relevant – if your client wanted something that only worked on iOS 8 and later you would still get 90% of customers – but the sheer number of machines you have to test for.
Of course many companies test on a small matrix of machines for Android, but as we can see from the very topic of this thread that's not really good enough – not for the low to mid end devices. If you are producing primarily for the US market you can get away with a few test devices, but not for a worldwide market.
You can dispute what you want, but here are the fact and figures.
All my Android developers have a Dell XPS13 developer edition (Linux based), a Moto G phone and a Asus Memopad tablet, all my iOS developers have a similar equipped Macbook Air, an iPhone 5 or 6 and an iPad Mini 2. Cost difference 700 Euros. Android Studio is also free and really blows away Xcode in easy of use and number of integrated development tools. It was a great move by Google to hire the people of Jetbrains to make their IDE, I wish Apple did the same. My feeling, based on the figured I measure on the development floor, is that the productivity with Android Studio is at least 20% higher than with Xcode.
We also developing hybrid apps using cross platform Cordova, which is great for less CPU intensive apps, so no games. Cordova apps are developed on Android and then ported to iOS.
Once an app is developed it is taken over by the testing and validation team who has a battery of test devices. For iOS there are 7 test devices, covering all possible screen sizes (iPhone 4,5,6,6+, iPad mini, Air) and iOS version 7, 8 and 9 (I'll drop 7 in a few months). For Android I have 9 devices with different screen sizes testing Android 4.2, 4.4, 5.0 and 6.0. This covers 99% of all the users who actually install new apps.
The time won by testing only 7 iOS devices is more than compensated by the more efficient publishing procedure for Android and by the time consuming app validation by Apple. For an average app a full testing cycling takes 5 weeks for Apple and 4 weeks for Android.
Jogeny Ess - First honor to comment . I'm OFW ,just NOKIA ASHA . LOL poor one!
Carlota - Nokia asha when im in phils and now samsung galaxy Ace 3 cant afford iphone its too expensive need to save bit of money.
Just a random comment - "Asha" means "hope". So technically, we who use the iPhone are hopeless!!
Facebook is going to have to put some deep thinking into deciding what devices to be tested. Another factor is device location. No device is going to perform the same across multiple countries and regions. Their best bet is to hire locals across the world as their testers and perhaps members of their project teams. If Facebook is global in reach, so should their teams.