Developing for Android

[OwlHoot]

Anyone looked at this or even earning a crust at it?

I'm wondering which of Android 2.3 or Android 3 I should start working with.

One would naturally assume Android 3 is a superset of 2.3, and will supercede/obsolete the latter; but I keep seeing references to Android 3 being targetted more at tablets rather than mobile phones. So could it be they are separate "strands" which will both remain in use?

[Solent]

I've just started Android development using the Eclipse IDE and Android Toolkit. It's very intuitive. I recommend the following Hello, Android: Introducing Google's Mobile Development Platform Pragmatic Programmers: Amazon.co.uk: Ed Burnette: Books

[xoggoth]

Just released first app yesterday as it happens. Will tidy up a few things and issue final version tomorrow.

https://market.android.com/details?i...=search_result

Many phones are being sold with 2.1 or were when I looked into it a few months back. Could not get anything above 2.1 to work on emulators when I started then although that too may have changed. In practice, most views for displaying images or text/web pages and the standard controls appear to have been stable for a while, my app works equally well on 1.6. It doesn't use any fancy stuff like GPS etc.

Compared to programming in .net I have found Android in Eclipse a total nightmare. A book might be good idea, you can spend ages trying to find decent guidance or solutions to problems on the net. There are some rough notes on all the various problems I encountered here:

http://www.xoggoth.org/pages/android.txt

[NickFitz]

As xog suggests, much of the market (i.e. end users) is still on 2.1 or less, including some on 1.x versions. As there is no overall control, and mobile service providers are pretty much able to dick around at will with the actual capabilities of the versions they deign to provide to their customers, the business of developing for Android can be a minefield. One can only be certain that a given app (beyond a certain degree of complexity) will work for a large part of the market if one tests on a few dozen devices.

[TimberWolf]

For what it's worth, the Wiki says the biggest market share is in Android 2.1 & 2.2:
"based on the number of Android devices that have accessed Android Market within a 14-day period ending on the data collection date noted below"

Usage share

Data collected during two weeks ending on April 1, 2011

Platform        API level               Distribution

Android         3.0 (Honeycomb) 	11 	0.2%
Android         2.3.3 (Gingerbread) 	10 	1.7%
Android         2.3 (Gingerbread) 	9 	0.8%
Android         2.2 (Froyo) 	        8 	63.9%
Android         2.1 (Eclair) 	        7 	27.2%
Android         1.6 (Donut) 	        4 	3.5%
Android         1.5 (Cupcake) 	        3 	2.7%

Android (operating system) - Wikipedia, the free encyclopedia

[amcdonald]

I've just started Android development using the Eclipse IDE and Android Toolkit. It's very intuitive. I recommend the following Hello, Android: Introducing Google's Mobile Development Platform Pragmatic Programmers: Amazon.co.uk: Ed Burnette: Books

I started off with that, it's a good starter book

[TimberWolf]

Has anyone made any money from Android app development?

[amcdonald]

Has anyone made any money from Android app development?

Only if you count something that strictly isn't an App, but runs on Android phones

[Freamon]

Anyone looked at this or even earning a crust at it?

I'm wondering which of Android 2.3 or Android 3 I should start working with.

One would naturally assume Android 3 is a superset of 2.3, and will supercede/obsolete the latter; but I keep seeing references to Android 3 being targetted more at tablets rather than mobile phones. So could it be they are separate "strands" which will both remain in use?

Android 3 is still closed source at the moment, so I suspect for that reason it would be a bit trickier? The open source nature of Android suggests to me it might be easier to develop for than iOS, but I have never actually tried either.

[xoggoth]

will work for a large part of the market if one tests on a few dozen devices

That what bothers me. Mine has been tested on a whole 1 real device. Given the negligible profit an app like mine will make testing on a range of devices, even using those test services you can pay by the month for, are not practical.

On the other hand, as I said, you are probably ok if you stick to fairly basic methods and services and everyone has a chance to see if it works on their phone for 15 mins before their payment is final.

One thing bothers me, if you go to android market on your phone you will find you are already signed in to Google checkout and there appears to be no way to sign out. You can only use android market and some other google apps if you phone is associated with a google account. Does that mean that if somebody steals your phone they can go shopping on your credit card at any online store that uses google checkout?