Is it advisable to develop and publish an Android application without first testing it on a real device?

Question

I want to develop an Android application and publish it. This is a very simple messaging app.

To test it on a real device, I will first need to buy an Android phone. Usually I just set up several virtual devices with different screen resolutions and different sdk and run the application on these devices using an emulator .

So, would you recommend testing the application on a real device? Any other suggestion?

Answer 1

I do not think this is appropriate. If possible, I would test the application on at least one device, preferably more. Can you borrow an Android phone from a friend for hours? If your application is really simple, you should be able to get someone to leave the phone for a short period of time (then get them out for lunch after: P). Or you can try to hook a cheap used phone with eBay or a similar site.

The only reason I do not fully trust the emulator is that it can be very slow and unresponsive, especially for games.

Answer 2

So, would you recommend testing the application on a real device?

That's right. I think publishing an Android app without testing on hardware is considered insane.

Answer 3

Yes, you must test the application on a real device. The performance you get on the emulator will be very different from the real device, and interacting with the application using the mouse and keyboard is very different from using the touch screen.

Do not borrow too much from one another or find a cheap, used phone on ebay or craigslist.

Answer 4

I test my game on three devices: a low resolution, hi-resolution and ultra-high resolution tablet, with two different versions of Android 2.1 and 2.2.

I am glad that I spent money on a low-resolution phone because I learned a lot about the density of the screen and the need to palm the bitmap without the possibility of scaling to maintain its quality.

For some direct layouts with viewing animations, emulators work pretty well just slower. Thus, it all depends on how demanding your applications are ... but as soon as it is convenient, it would be logical to get a device.

Answer 5

One of my biggest fears is the fact that I do not have access to multiple devices. I can only check it on my current phone, so I have no idea how it will respond to an older device with a lower screen resolution or Android build, and I will not know how it will interact with the equipment, and not on my current phone (trackball, etc.)

Also, with all the different user interfaces (sense, touchwiz, etc.), I am afraid that the use of color may not reflect very well on a particular user interface (for example, if I have an application with an amazing graphic effect, color).

Answer 6

If you do not want to buy a phone, there are some services that allow you to remotely use real devices over the Internet. Perfecto Mobile has a pay-as-you-grow payment plan (I have no connection with them other than a satisfied customer).

Answer 7

Do whatever you want. You can buy 30,000 devices, but if no one downloads your application, what will happen?

The idea is to develop FAST, and then improve your application and the testing process, if users really love and show some craving for your application. If buying an expensive device makes you unhappy and less to program, then do not buy it! :)