From: Björn Stenberg <bjorn_at_haxx.se>
Date: Mon, 10 Oct 2011 11:03:27 +0200

Rockbox on Android is a complex topic, subject to the interpretation and expectations of each developer and user. Here is my view:

Android is a very different target from our hardware targets. On an iPod or a Sansa, we are the masters of the device. We control everything. Nothing happens without our say-so and we are usually overwhelmingly superior in capability to the original firmware. And even when we aren't, we are still the only alternative available to running the original firmware. It's Rockbox or bust.

Additionally, the people who load Rockbox on their MP3 players are typically tinkerers. People who are willing to try a "hacker firmware" on their device, are prepared for different behaviour and ready to do some tweaking to get things as they like.


On Android, things are very different. We are guests there. We must coexist with an operating system and heaps of other running apps. Users have a multitude of options and we will be merely one of dozens of music players to choose from.

The users who download Rockbox for Android are going to be very different from the people flashing Rockbox onto their iPods. They are not replacing the firmware of their phone, they are just testing another app. They will expect an app that works roughly like the other music player apps they've tried. If we deviate too much from that expectation, many users will quickly give up and uninstall. The first 15 seconds of user interaction are crucial.

I want the default experience of Rockbox for Android to be geared towards people who have never heard of us before. To the people for whom Rockbox is Just Another Music App. Eventually, they will find that we are more capable than other music apps. We can play more file formats, and we can play them better than the others.

But unless we survive the first 15 seconds of scrutiny, none of that matters. We cannot demonstrate our superior file format support in the first 15 seconds, because people don't store music they can't play on their phones. We won't even have time to show off our gapless playback, crossfade, equalizer or any other fancy stuff.

We may be technically superior, but the first impression is the user interface. We need to tailor it to feel familiar and simple, even for first-time users.

One part of that is the theme. I want to forego cabbie and use a discreet default theme that more resembles native Android, using the Android font and similar line spacing, maybe even similar icons. I want an .sbs with << || >> play controls at the bottom of the screen and a simple WPS with only album art, artist/song/album name, progress bar and play controls.

Another part is the behaviour. Our small-screen cursor-button history is very visible and we need to work to smooth off some of the rough edges.

I want the context menu to be a dialog, showing the previous screen behind it. I think that would make it a lot less confusing. I want our lists to have touch feedback (the orange blink). I want the select bar disabled completely until cursor keys are used. I want to look into how the first-time user most easily can find and play his music. Ideally, I would ideally like our vol/bass/treble/bal settings to be less list-y and more touch-focused.

So why not just go native? It is my opinion that distilling Rockbox into librockbox and coupling it to an android java GUI app is the way of the future. But I also think that future is some time off still. I think we can make a Market-worthy Rockbox app with our current GUI first, release it, and *then* do the library split and java app.

-- 
Björn

Received on 2011-10-10