[OT?] Android phones

Mon Mar 29 14:15:00 UTC 2010

Hi,

I'm an iphone user and use a quite expensive app in my ongoing effort
to learn Chinese called PlecoDict. There really is no competition in
this space for them. Michael Love the developer regularly sends out
updates on where things are going. Here is what he had to say about
why he wasn't going to develop on the Android platform.

***************************

8. Android

I've gotten quite a few emails lately asking about a version of Pleco
for the Google Android OS that's used on phones like the Nexus One and
the Motorola Droid, so I wanted to provide some clarification on that
too. While we know Android is getting to be quite popular, at the
moment we're not prepared to commit to ever developing an Android
version of Pleco, and here's why.

The big problem with Android for us is that it's an open-source
operating system; this means that the mobile phone manufacturers and
cellular carriers that use it can essentially modify it however they
see fit, without any restrictions from Google or anybody else. They
don't need to pay Google any license fees, in fact they can even
replace Google as the default search engine if they so choose.

That might seem like a fine and lovely thing, but the problem with it
from a software development perspective is that without a central
authority, there's nobody in a position to make sure that every
Android-based phone is compatible with all Android software; indeed,
compatibility issues have already started to appear on some Android
phone models. Adding new features to the operating system is a great
way for mobile phone manufacturers to differentiate their phones from
those of their competitors, so many of them have eagerly done so with
Android, even to the detriment of software compatibility.

Android's open-source nature also means that, somewhat ironically,
manufacturers and cellular carriers can impose whatever restrictions
they want on their Android-based phones without anyone stopping them.
Android is often touted for its openness, the fact that unlike iPhone,
Android phones can run applications even if they haven't been
"approved" by the manufacturer, but on some newer Android phones (e.g.
the Motorola Backflip) that's no longer the case; apps for those
phones have to be submitted to Google or the phone manufacturer before
release, and can be approved or rejected just as on iPhone.

So we're very worried that in another year or two there'll no longer
be one "Android" we can develop for - there could be half a dozen or
more Android-based OSes that we have to individually develop /
optimize our software for, and multiple application stores to deal
with every time we release an update. Which could mean that instead of
continuing to improve our software and add new features, we'd be
spending all of our time simply making sure that our software works
correctly on the latest Android phones.

For some types of software, this might be OK - games, for example,
generally rack up most of their sales in the first few months after
they're released, so a developer can release an Android game and not
really have to worry about whether it'll still work correctly on the
newest Android phones a year down the line. And server-based
applications like Twitter and Facebook are usually simple enough that
they can be kept up-to-date with relatively little work. But Pleco
sells very complicated, expensive, specialized software, and it
usually takes us at least a year to add support for a new operating
system, so we need a level of stability / consistency that we don't
think we're likely to get with Android.

Now of course this could change - Google could come up with some
clever new licensing / incentive system to keep all of the various
Android phone manufacturers in line, or the fragmentation problem
could turn out to be less severe than we fear - but we'll need to see
a lot of evidence that that's happening before we can consider
developing an Android version. So even if we do eventually decide to
do one, I think it's unlikely we'd begin working on it until sometime
in 2011, which means it probably wouldn't be out until 2012.

We are certainly aware of the need to offer *some* alternative to
iPhone; I've never quite understood why so many people hate Apple
(it's not like Microsoft is any better), but they do, and anyway it's
not very good business to be dependent on just one platform for all /
most of your sales. We're just not sure if Android is the right choice
for that alternative; along with Windows Phone 7, the new Nokia
Symbian^4 due out around the same time seems like it has a lot of
potential (and, like Windows, would be easier to develop for than
Android), and a mobile web-based version of Pleco has a great deal to
recommend it too.

Best,
Robert

On Sun, Mar 28, 2010 at 3:15 PM, Christopher Browne <cbbrowne-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org> wrote:
> On Sun, Mar 28, 2010 at 2:31 PM, S P Arif Sahari Wibowo
> <arifsaha-/E1597aS9LQAvxtiuMwx3w at public.gmane.org> wrote:
>> On Sun, 28 Mar 2010, James Knott wrote:
>>>
>>> While it would be nice if they didn't diverge, it's still open source that
>>> has it's /root in Linux.
>>
>> True, that's probably the most important thing. Furthermore divergence in
>> kernel space is less an issue anyway.
>
> I suppose.  It won't affect userspace so much; adverse effects would
> be on those trying to establish hardware compatibility between new
> sorts of hardware and the platform.
>
>> IMHO a more important issue is the
>> original Android's decission to have application stack that entirely
>> different than in Linux OS (i.e. GNU/Linux). That mean that many
>> applications have to be built specifically for Android; therefore the
>> community size of the that application is much smaller than if it can use
>> common Linux application.
>
> The size of community may be small now; if the stack gets used more
> and more for mobile phones, the population increases, and "iPad-like"
> devices would add further still to the population.
>
> What a "common Linux system" means is a mighty elusive thing; in the
> 'embedded/mobile world' there are four answers that I don't imagine
> are what you have in mind:
>
> 1.  Android would be one, for sure
>
> 2.  BusyBox, and very little more, which is what you frequently see on
> storage devices
>
> 3.  Maemo (the Nokia platform, using some bits of GNU, originally with
> Python+GTk, later biasing to Qt),
>
> 4.  Moblin+Maemo is in the process of begetting Meego.
>
> None of these represent places where I'd expect that what you'd call
> "Common Linux apps" would necessarily run trivially.
> --
> http://linuxfinances.info/info/linuxdistributions.html
> --
> The Toronto Linux Users Group.      Meetings: http://gtalug.org/
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
>

-- 
Robert F. Kennedy
Handyman Services
c. 647-367-3145
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists