Why developing for Android sucks | Hacker News Reader

Why developing for Android sucks

webosroundup.com

14 pointsechubb17 comments

This article is citing a lot of out of date information (unsurprising, since its 9 months old). I suppose that the article should be taken with a grain of salt given the source anyway, but just a heads up on a few:

a)the Wired article that says 70% are still on 1.5/6 is from 9 months ago. Figures from November 2010 indicate that 75% of users are running 2.1/2.2: http://gizmodo.com/5679410/over-70-of-android-phones-are-now...

b) Google has since implemented some antipiracy measures on the app store.I'm less certain about its details/market penetration, but it exists in some form

c) android now has a lot more than 10 million handsets.

Another point is that most Linux distributions have way more than 60% free apps in their repositories.

I don't know if the information being out of date matters in this case, the problems of OS version fragmentation still remain.

Android antipiracy measures are extremely brittle and rely on server-side verification, this has been circumvented by decompiling app archives, patching them and rebuilding them, which is entirely possible using apktool or your decompiler of choice. Frankly im surprised that the platform isnt crawling with malware.

I wouldn't deny that fragmentation is an issue, although from my (limited) personal experience the biggest issue has been designing for all the different screen sizes/resolutions and still having your app look okay.

And yes, any antipiracy mechanism can be circumvented. That said, 90+% of android users probably dont have the knowhow to grab the apk, decompile it, and remove the server side check. And I'm not sure on this last bit, but I know at least in the US some carriers (ATT) disable installing apps that aren't from the appstore. I haven't read about the details of the antipiracy mechanism, so I can't really say more than that.

Also, its unclear to me if you're linking the fact that theyre so easily decompiled/cracked and that you're surprised it isnt crawling with malware.

Good article (apart from advocating developing for Palm??), bad title. It should be called "Why distributing your Android app sucks." No commentary on actual app development itself. My thoughts on that are that Google should develop an IDE for Android developers - Eclipse is unstable bloatware, the Android emulator sucks and supporting different screen sizes and resolutions is odious to say the least.

People often repeat that Eclipse is "unstable." I can use one Eclipse process for weeks, even months on-end with JVMs from Sun and IBM on Linux, and it never crashes or needs a restart (except if installing new plug-ins). The ADK tools don't have the polish of the core features, and I haven't used them as much as other tools, but what's causing instability?

I've had serious problems with Eclipse's autocomplete/intellisense prediction feature. It often begins listing identifiers and then just hangs indefinitely. I don't know if the fact I'm developing using the Windows build is significant - it is Java after all so it should be platform agnostic.

I seem to have bad luck with IDE's though. I develop using Visual Studio 2008 and that has some horrendous problems, though VS was never known for stability (six service packs for VS6, anyone?)

If you do not like Eclipse, don't use it (for Android application development).

There are at least two other free IDEs (IntelliJ IDEA CE and Netbeans) and you don't need one anyway. Use a text editor and you're just fine. I'd recommend building with Maven but if you really want to use ant.

> Over the Air updates. If you have a Palm phone you will be on the latest version that your carrier has available

I wonder what happens when you use an unlocked phone on a foreign telco. Will it get upgraded using Palm's servers?

Some of the points are only valid if your revenue model is a paid app. How about a free app + paid service?

If piracy is as big a problem on Android as the article claims, the free client/paid subscription model is probably the most viable way to generate revenue. However, Android is not proprietary software-friendly - It's INSANELY easy to decompile APK archives and reverse engineer content.

Interesting. How vibrant is the dev community for Palm phones?

OTA Updates are another story on Android. Different manufacturers price OS updates differently like Samsung, who price them as feature updates whereas some other Android phone manufacturers price them as cheaper maintenance upgrades. It can be a bit of a lottery as to whether your phone will ever receive an upgrade.

I'm part of the WebOS dev community and its small but growing. I've sat on the fence for a while because I wasn't sure where Palm would be a year ago. And while I found the interface to be the best in the smartphone space (sorry iOS) the hardware it was deployed on was deplorable.

They are moving from Mojo (javascript based) to Enyo (C++ based) platform so quite a few developers are also waiting for the new SDK. Most of the apps that are currently written in Mojo will continue to work but I sense that most developers are in the same boat I am. No point in learning a framework that has one foot in the grave when the next one is right around the corner.

Enyo is NOT a C++ platform. There's a C++ "plugin development kit" that's been out for a couple months now and AFAIK that doesn't change when Enyo gets introduced. I didn't really look, not particularly interested in WebOS C++ development.

Enyo is a palm-developed js framework originally created by the Ares editor team that performs (reportedly) dramatically better on the same hardware than Mojo. I believe the difference is that Enyo uses a javascript layout system a la Sproutcore/Cappuccino instead of the html+css layout in Mojo so they're basically innerHTMLing everything instead of performing a bunch of DOM manipulations/reflows. I'd expect a number of other changes but aside from that and allowing multiple-layouts (i.e. for tablets) that's all I came away with from the developer day talk on it.

I started looking into WebOS when CES was going because I was wondering at Palm's absence. It turns out that the homebrew community is surprisingly vibrant and they have (basically) an independent market app (preware) for their apps and patches. I was most impressed by the patching. The entire interface is interpreted, so changing how things work is just a matter of modifying the appropriate javascript and rebooting the device.

All in all, I was impressed enough to get a developer account and start intermittently hacking apps. Enyo isn't out but it'll have to come out soon in order to handle tablet layouts so I'm heistant to commit to a Mojo app before the big Feb 9 event. I don't know about a webos phone (google apps+NewsRob+Kindle on android do basically everything I want to do with a phone) but I'll definitely be getting a WebOS tablet for reading/development.

What about free + admob? Angry Birds is free on Android, but uses Ads to support it. Being the top free game on the platform, this means they probably do a lot of money with it.

Yes, But why is it so? Angry Birds first became very popular on iPhone and then was ported to Android. It could be that if they started on Android the outcome could have been different.

Rovio specifically said that they went Ad-supported on Android because NONE of the top apps were paid, and they didn't think they'd make much money with a paid app. There was a lot of talk about make a paid version without ads, but it never appeared.

Reply on news.ycombinator.com