"In defense of Google" thread branch: J2ME and Android
Casper Bang
wrote:
> Can you define "popular"? Possibly in terms of number of installations?
> Thanks. :-)
I can try, although I won't jump on the number bandwagon as I think
that's utterly pointless. Most people with J2ME capable phones have
never actually run such an application let alone installed a 3'rd part
one.
So what do I base this on? Well for one thing, rate of adoption and
current trends:
- Google trends suggest much higher interest than there ever were for
JM2E: http://www.google.com/trends?q=J2ME%2C+Android
- Amazon lists a total of 36 J2ME books and 16 Android books.
Considering J2ME has been out since 2000, Android only since 2008,
that amounts to 8 books a year for Android vs. 4 books a year for
J2ME.
And I have no reason to believe this trend won't continue since:
- There are over 20 Android phones coming out this year.
- Manufactures won't have to pay for a J2ME license to include
Android.
- Google provides editor, emulator, debugger - the whole shebang in an
easy to use SDK.
- No pre-processor required, as the Android spec is less liberal than
the J2ME and its idiosyncrasies.
- There are already over 5.000 Android applications, I wish I knew how
many J2ME apps there are but as usual (same thing happened with the
component marked) Sun failed to provide such a collaborative and
unifying model.
- People coming from J2ME like Android better, because of the powerful
yet easy to use API's (component library, SQL datastore etc.).
- Android is a superset of J2ME and supports generics etc. making J2ME
looking more stale than ever.
Do you feel otherwise about J2ME, its achievements and future?
/Casper
Fabrizio Giudici
On 1 Jul., 14:30, Fabrizio Giudici <fabrizio.giud...@tidalwave.it> wrote:Can you define "popular"? Possibly in terms of number of installations? Thanks. :-)I can try, although I won't jump on the number bandwagon as I think that's utterly pointless. Most people with J2ME capable phones have never actually run such an application let alone installed a 3'rd part one.
For the remainder, I don't think your further points are good :-) Trends are a thing, current popularity another. Do we have a way to measure the number of installed applications, etc?
-- Fabrizio Giudici - Java Architect, Project Manager Tidalwave s.a.s. - "We make Java work. Everywhere." weblogs.java.net/blog/fabriziogiudici - www.tidalwave.it/blog Fabrizio...@tidalwave.it - mobile: +39 348.150.6941
Jess Holle
> - Android is a superset of J2ME and supports generics etc. making J2ME
> looking more stale than ever.
>
Java-5 -- and neither Sun nor any J2ME partner (at least none I asked at
JavaOne) has any plans to even attempt to do anything about this!
If Android does nothing else it will hopefully give these folks a kick
in the pants to either (1) really update J2ME, (2) support J2SE [plus
mobility libraries] and give up on J2ME, or (3) give up and support
Android instead of J2ME.
--
Jess Holle
Casper Bang
quantifiable as potential installation numbers in the wild, then the
discussion is pointless and starts to look like you're just bull-
baiting me into coming up with a number. However I would still love to
know what then you think of the situation, you are obviously of a
different impression yet didn't really provide counter arguments nor
answer my question.
Also, I'm rather curious as to how you distinguish a trend from a
popularity, I would think it's pretty clear that at no given time in
the last 7 years was J2ME as popular a search query as Android is
today. To me at least, developer interest weighs higher than naïve
installation numbers.
/Casper
On 1 Jul., 16:14, Fabrizio Giudici <fabrizio.giud...@tidalwave.it>
wrote:
r0b
Android application runs in its own process, with its own instance of
the Dalvik virtual machine. Dalvik has been written so that a device
can run multiple VMs efficiently.'
1 http://developer.android.com/guide/basics/what-is-android.html
Casper Bang
> This is incorrect. There are over a billion installations of Java on
> cellphones. Google has made J2ME based clients before Android and
> continues to do so. If you want to hit more than 5% of the market you
> have to use Java.
Whether its a million, billion or gazillion devices don't tell the
whole popularity story but I can understand this is a desirable spin
for Sun.
Pointing out that Google also targets J2SE doesn't prove much, given
that they are also more than willing to make Obj-C clients for the
iPhone and other non-J2ME devices. Google is just like that, they go
after the marked and tries to win the hart of users regardless of
underlying technology.
With your billion I assume you refer primarily to Nokia/Symbian
devices, well if Sun wish to continue using that as a pitch, they
might be in for a surprise given what's stirring up in the Symbian
Foundation:
http://blog.symbian.org/2009/05/
/Casper
Joshua Marinacci
> that they are also more than willing to make Obj-C clients for the
> iPhone and other non-J2ME devices. Google is just like that, they go
> after the marked and tries to win the hart of users regardless of
> underlying technology.
Yes, that is exactly my point. Google is pragmatic. JavaME may not be
hot or exciting, but it has the numbers, so Google ships apps for it,
and the probably will for a long time. I don't disagree that
platforms other than JavaME are growing and more exciting, especially
as smartphones grow to be a major portion of the market, but I want to
dispel the myth that Android (or iPhone, or Palm, etc) have more
marketshare than JavaME. They don't. Smartphones collectively still
have less than 5% of the market. If you want volume today, it's JavaME.
Casper Bang
marked share regardless of how many actually USES their Symbian phone
for anything but calls and messages (I have yet to meet one). I think
it's naïve to care only for such a marketing metric, but if that's
your definition of popularity then so be it. :)
/Casper
Joshua Marinacci
or might not target JavaME. I'm telling you why. Google is practical
and targets large markets, so they build JavaME clients. A much
better question is why does Google target the iPhone which has such a
small marketshare. Probably because they see it growing into
something much bigger.
The Symbian blog you sent is accurate. MIDP 3.0 hasn't gone well and
the JavaME market is fragmenting. If only there were a new platform
that would smooth out the differences of JavaME implementations and
let developers code at a higher level. A modern 21st century platform
with modern tools, but that leveraged existing code and runtimes. Hmm.
If only. :)
Alexey Zinger
From: Joshua Marinacci <jos...@gmail.com>
To: java...@googlegroups.com
Sent: Wednesday, July 1, 2009 12:15:59 PM
Subject: [The Java Posse] Re: "In defense of Google" thread branch: J2ME and Android
Viktor Klang
So basically: It doesn't matter how many installations there are, very few people like to stab themselves in the face. Even when they're paid to do so.
Viktor Klang
Scala Loudmouth
Fabrizio Giudici
> >From what I've heared from sources in the industry whom I
> unfortunately cannot name, developing for J2ME is less pleasant than
> stabbing yourself in the face with a sharp and rusty object repeatedly.
>
> So basically: It doesn't matter how many installations there are, very
> few people like to stab themselves in the face. Even when they're paid
> to do so.
Viktor, this is a well known fact and there's no need to refer "source
in the industry that you can't name" :-) I could quote tons of person
and it would just suffice to have a look at conference presentations.
The people I'm thinking of are fans of Android for the reason we're
saying. But the same persons I know are also saying that TODAY if you
want to go with a pervasive product, you'd better to stab your face with
JME rather than stab your face multiple times with different,
proprietary technologies. The point, as Joshua said, is that TODAY
Android is only a very small fraction of the market. Since it's pushed
by Google, it can do very well, but - again - I'm not talking of trends
and predictions; it's an argument that I'm not interested of. I posted
in this discussion only to argue against Casper's assertion of JME
popularity. Casper, I can argue pretty well and in a few words: it's
just ridiculous to assert that a thing is popular when 95% of the
platforms (and thus people) have never seen it. In my part of the world
(and in my country there are more cell phones that inhabitants) Nokia is
one of the most common brands, while I've still to see a single person
owning an Android phone. Now, even though every Android owners
compulsively used Android applications and only a fraction of Nokia
users run a Java game, Java would be still the most popular. Of course
there's a specific ratio where parity could break in favour of Android.
But you, Casper, brought the idea of counting effective uses instead of
installations, and it's up to you to provide numbers. Otherwise your
statement is just as pointless than Jonathan's or James'.
MIDP is stale? But we have JavaFX for Mobile. What? Yes, it runs only on
a couple of devices: but it doesn't seem much less devices than Android.
Of course, things could diverge pretty soon, and how much is strictly
related on the power ratio Sun vs Google. But I'd better say Oracle/Sun
vs Google, and at this point there are so main unknown variables that
the discussion, now, doesn't make any sense. I don't think we have to
wait too long, by the end of the year I think we will able to guess how
things will evolve: Oracle doesn't take compromises, either they shut
down JavaFX (and then the discussion is over), or they push it
seriously, and they have the power to compete with Google.
--
Fabrizio Giudici - Java Architect, Project Manager
Tidalwave s.a.s. - "We make Java work. Everywhere."
Casper Bang
> Android is only a very small fraction of the market. Since it's pushed
> by Google, it can do very well, but - again - I'm not talking of trends
> and predictions; it's an argument that I'm not interested of. I posted
> in this discussion only to argue against Casper's assertion of JME
> popularity.
count, which is a red herring.
> Casper, I can argue pretty well and in a few words: it's
> just ridiculous to assert that a thing is popular when 95% of the
> platforms (and thus people) have never seen it. In my part of the world
> (and in my country there are more cell phones that inhabitants) Nokia is
> one of the most common brands, while I've still to see a single person
> owning an Android phone.
average Joe? I'd be surprised if it were any different than up here in
Scandinavia, where it's pretty rare. Nokia are trying to get a marked
going with the Ovi store but this is the classic case of too little
too late.
Meanwhile, second generation Android devices are starting to flood the
marked, perhaps you're a little behind where you are from (Italy?) but
the rest of Europe (and the US) seem to be following along sooner
rather than later. The general opinion among reviewers are that this
is the iPhone-for-the-people and even the top Nokia stuff (5800 to
n97) really don't compare.
> But you, Casper, brought the idea of counting effective uses instead of
> installations, and it's up to you to provide numbers. Otherwise your
> statement is just as pointless than Jonathan's or James'.
popular? By your logic it is, due to the mere count of J2ME BD-J
machines out there. However, since very few titles actually make use
of BD-J. Most people would interpret that as BD-J being unpopular!
> But we have JavaFX for Mobile. What? Yes, it runs only on
> a couple of devices: but it doesn't seem much less devices than Android.
alone and an expected growth of 900%. Meanwhile, JavaFX is barely
ready for the desktop let alone cell phones on top of J2ME, so I'd
really hate to put my money on that one. But the syntax is nice, I'm
sure we'll see it for the UI layer (Activities) on Android.
Peter Becker
app I currently use is "5ud0ku" :-)
And to support your point:
- I would not consider myself an average user
- my next phone is certainly not going to be Nokia or Symbian-based
I have used a few things on the phone over the years, from web-browsing
(including the J2ME version of Opera), via instant messaging to podcast
listening. The latter was actually a major driver in picking the phone
(cheapest option for something with WLAN and audio playback), and the
podcast app by Nokia is actually comparatively good (meaning: I wouldn't
fail a student if they would deliver that, as opposed to most of the
other things on the phone). And I still use the phone for the podcast
(not for web or IM anymore).
So I am one of those people who actually used J2ME applications. But I
would not appreciate anyone targeting my current setup as a market, it
certainly seems wasteful and I'm just waiting for my current contract to
expire (I believe I have two more months to go).
Peter
Joshua Marinacci
behind the times. It is now far from cutting edge and you wouldn't
target it for new products. I agree. Sun is aware of this and we have
been & are taking steps to address it. The partnerships and new
review system announced at JavaOne are part of it. JavaFX for Mobile
is part of it too.
Android and iPhone clearly have some good things going for them and
they are growing tremendously. I only jumped in this thread to dispel
the myth that there's no JavaME phones out there. There tons of them.
They may not be exciting but they dwarf everything else by numbers.
But that's changing, and we know it, and we are taking steps to bring
JavaME up to date.
- Josh
Fabrizio Giudici
> See that's because you equate popularity with runtime distribution
> count, which is a red herring.
>
>
Let's talk about USE, but explain how it's possible that people feel
most popular a thing that most of them have never seen. Installations
clearly plays as a cap for use.
All the remainder of your post is conjugated at the future: "is starting
to", "are starting to"... It seemed clear to me that I'm talking about
PRESENT, since your original statement is "JME _is_ not popular". I have
clearly said in my previous email that things will evolve and could
evolve in various ways, including yours.
For what concerns use, joe the average uses to play and chat. Not
everybody is an engineer or a business manager. And most of mobile phone
users are teenagers. In my understanding, the largest number of games
for mobile phones are made in Java. But let's talk about Android use.
You say you're seeing a lot: what does joe the average do with them?
Peter Becker
phone convincing -- independent of the question what the application was
programmed in. To some extent the JavaME ones might have been better --
you get the longer startup time, but at least they don't crash your
phone by slowing eating all the memory available.
The one application I found impressive was actually the J2ME version of
Opera, although that clearly suffered from being J2ME: you had to allow
network access every single time you started it. But in terms of
usability, features and performance it was way ahead of the native
browser Nokia provided.
Peter
Jess Holle
> So I think you are saying that JavaME has been useful, but fallen
> behind the times. It is now far from cutting edge and you wouldn't
> target it for new products. I agree. Sun is aware of this and we have
> been & are taking steps to address it. The partnerships and new
> review system announced at JavaOne are part of it. JavaFX for Mobile
> is part of it too.
>
to actually bring J2ME into the 21st century. They touted the new SDK
at Sun, but it still does not support Java 5 language features and
completely breaks WORA (even assuming you still to the subset of Java
supported by J2ME) because you can't write modern Java code and run it
on J2ME period thanks to this.
JavaFX is cool, but in this regard it is something of a layer of shellac
or a bandaid. WORA will hopefully be preserved for JavaFX, but things
will fall on their face as soon as you need to use a Java library.
Android in some ways holds up to the WORA promise better than J2ME.
--
Jess Holle
Casper Bang
the same as I see for the iPhone (but don't see for Nokia's). You
raise an interesting aspect though, of the few J2ME applications I
have run into in the wild, exactly as you mention have all been silly
little games for teenagers that came with the phone such as not to
appear empty. But that doesn't strike me as a particular convincing
definition of popularity.
Perhaps I'm wrong and there's a thriving multi-million $ marked with
lots of companies, developers and applications - but if there are, I
just haven't noticed in spite of having a Nokia the last 10+ years.
Perhaps we should resume this thread in 6-12 months, when it's
possible to talk installation count and usage cap. because I think
this is a dead end. :)
/Casper
On 2 Jul., 07:53, Fabrizio Giudici <fabrizio.giud...@tidalwave.it>
wrote: