How does Android market handle this problem? Android supports native development using the NDK, perhaps the apps using native code appear only on devices with a supported architecture?
My HTC phone was supported for 0 months, is the most buggy phone I've ever owned, and in its life it has received 0 (zero) updates - not to upgrade it from Android 1.6 which was already obsolete one month after the phone was released (September 2009), but not even a minor update just to fix the dozens of bugs it has.
In comparison, my $60 dumbphone from another manufacturer has seen 3 firmware updates so far.
While I believe that the EU is targeting the corporate world with these rules, it would be interesting to see an attempt to enforce this against individuals.
I don't know how it will work in the rest of EU, but for the politicians of my country it will probably mean: "if you don't delete within 24 hours your blog post where you remind people that I was convicted for corruption ten years ago, I will sue you to death".
And if people will protest because, say, they might want to know if their babysitters have a past as child molesters, the politicians will say "hey, it's not our fault, it's the EU which is asking us to pass this law". They always do so.
Or they could pick a scripting language and write the UI in that. I hear that Python runs on Windows, Linux, and Mac OS X (the three platforms LibreOffice appears to support), so maybe they could use that.
Or if they're really attached to the letters J, A, and V, they could use Mozilla's JavaScript engine.
Both of which would likely be faster and lighter-weight than trying to use Java for UI glue code.
Those who actually bothered to benchmark Java against Python can tell you that Java is objectively, measurably, several order of magnitudes faster than Python.
What's "your favourite scripting language" which doesn't break between releases? And how is it deployed so much easily than Java?
And what C++ environment are you using, which doesn't break when you change compiler, compiler release, C++ standard, endianess, ABI, operating system, architecture?
Vanilla OpenOffice.org Writer version 3.2.1, on a 2.5 GHz Pentium IV (from 2002) running Linux, takes 10.3s to cold-start, 4.8s to start the subsequent times.
You're spreading the usual anti-Java misinformation. And by the way, the vast majority of OpenOffice is written in C++.
I agree with you. I don't understand why most press supposes that commercial entities must either dominate a market segment, or "die". You can see it every time a new product is introduced and it gets called "%s-killer", where %s is the name of the best competing product according to the person who is spreading the news. I've never seen such "killings" succeed - most often, the new product takes it share of the market, both products coexist gracefully, and then they usually die approximately together, when they get "killed" by the new generation of products brought about by the respective manufacturers.
To my eyes, this obsession about "extreme" competition makes the media look childish, to say the least.
Every phone I've owned, if featuring an alarm clock, didn't require me to keep it turned on to wake me up. Even the cheapest ones. I was surprised when I discovered that my Android phone didn't alert me when it was turned off, and now I learn that the iPhone has the same limitation. It's strange because even my old Motorola, which is american-designed, does turn itself on for the alarm. I'll bet Apple will add this feature in the iPhone 7, and then every technical analyst on the Internet will proclaim that Apple has revolutionised the market of alarm clocks.
Ok, I was just pointing out that the question required a solution in "pure java" and your response provides a solution which isn't "pure java", so I saw a mismatch between the two which might be worth noticing.
tepples is probably aware of JNI, as the library he pointed to was a JNI wrapper.
In fact I think that WebKit is a nice success story for LGPL, which should be looked at by people from the "pragmatic / software is a tool / BSD is more business-friendly" camp. The supposedly "business-unfriendly" WebKit has created more jobs that many stimulus packages.
No programming language will allow you to exploit every feature from any operating system without writing code that is specific to that operating system.
Does that peripheral show up as a character device on Linux? You can open it and read it from Java.
Does it requires libusb on Windows? You can call that library from Java.
The peculiarity of Java is that it really strives to let you avoid os-specific code when it's possible, and it's currently the best platform at that. Of course its standard library can't possibly cover all possible hardware/software interations because they're not a finite set.
While you're at it, you can demand that domain in Shavian. It's a script that is native to English and it's already in Unicode; so you can stop using those foreign roman letters with inconsistent pronunciation.
Being able to speak your own language has nothing to do with not being able to understand others' languages.
I believe computers should adapt to people's customs and not the opposite; current technology gives us the opportunity for this to happen more often than not, and localised domain names are a step in that direction.
True, but I see some differences: the N900 wasn't meant for developers only, and there wasn't a non-developer, locked down variant of the N900. Also, the "app store" wasn't restrained on the N900 as it was on the ADP1.
From what I've heard from Nokia, it looks like they will provide the N9 [the successor of the N900] with the ability to switch at boot time between a "DRM" mode and a free mode; while in the free mode, you won't be able to access DRM-restricted media. How this will be accomplished exactly I don't know, and that's what effectively scares me a bit.
Yes, but in theory Android is fully open, too. Perhaps it's less open from the platform development point of view (correct me if I'm wrong, but I think you can't follow or influence its development, you only get to download Google's code drops once in a while). Technically, you could edit its sources and run the modified OS on your phone.
The problem is that manufacturers lock down their phones (even Google does that, on their directly-branded models), so their bootloader won't accept unsigned OS images. Some manufacturers go even further and lock down applications installation.
Even if MeeGo is as open as it gets, it has all the mechanisms in place to become as closed as Android is if manufacturers so wish (see MeeGo's security framework). Nokia has already shown with the N900 that they won't lock their handsets, but other manufacturers might have a different vision. I hope that the development of a large "independent" application repository will encourage them not to.
1) Just download the.tgz version: it doesn't contain the documentation.
2) Extract it anywhere you like - it's self-contained. E.g./usr/java.
3) export JAVA_HOME=/path/where/you/extracted/it
4) export PATH=$PATH:$JAVA_HOME/bin
I find this much more convenient than using (in my case) Ubuntu packages which add unwanted dependencies, scatter the installation in different places all over the filesystem and thus are less maintainable when things go wrong, are slow to track the latest released Java version, and periodically require me to run update-alternatives to avoid Sun's tools being replaced with not-so-compatible alternative implementations.
Note that OpenJDK *is* Sun's next version of Java, so when you switch to it because you have no "confidence in the quality of their software", you're not solving your problem.
I find TFA pointless. It basically says that Google won't be able to maintain a fork of Harmony (and only those bits which are actually required for Android!) without IBM's help, while at the same time stating that Google already has more people working on OpenJDK than IBM has.
Inexplicable.
Slashdot Mirror
How does Android market handle this problem? Android supports native development using the NDK, perhaps the apps using native code appear only on devices with a supported architecture?
OK, I was just answering to the "missing MBR" problem.
diskpart / select disk x / clean / create partition primary
In comparison, my $60 dumbphone from another manufacturer has seen 3 firmware updates so far.
While I believe that the EU is targeting the corporate world with these rules, it would be interesting to see an attempt to enforce this against individuals.
I don't know how it will work in the rest of EU, but for the politicians of my country it will probably mean: "if you don't delete within 24 hours your blog post where you remind people that I was convicted for corruption ten years ago, I will sue you to death".
And if people will protest because, say, they might want to know if their babysitters have a past as child molesters, the politicians will say "hey, it's not our fault, it's the EU which is asking us to pass this law". They always do so.
Or they could pick a scripting language and write the UI in that. I hear that Python runs on Windows, Linux, and Mac OS X (the three platforms LibreOffice appears to support), so maybe they could use that.
Or if they're really attached to the letters J, A, and V, they could use Mozilla's JavaScript engine.
Both of which would likely be faster and lighter-weight than trying to use Java for UI glue code.
Those who actually bothered to benchmark Java against Python can tell you that Java is objectively, measurably, several order of magnitudes faster than Python.
First google search result.
And what C++ environment are you using, which doesn't break when you change compiler, compiler release, C++ standard, endianess, ABI, operating system, architecture?
You're spreading the usual anti-Java misinformation. And by the way, the vast majority of OpenOffice is written in C++.
To my eyes, this obsession about "extreme" competition makes the media look childish, to say the least.
Every phone I've owned, if featuring an alarm clock, didn't require me to keep it turned on to wake me up. Even the cheapest ones. I was surprised when I discovered that my Android phone didn't alert me when it was turned off, and now I learn that the iPhone has the same limitation. It's strange because even my old Motorola, which is american-designed, does turn itself on for the alarm. I'll bet Apple will add this feature in the iPhone 7, and then every technical analyst on the Internet will proclaim that Apple has revolutionised the market of alarm clocks.
tepples is probably aware of JNI, as the library he pointed to was a JNI wrapper.
In fact I think that WebKit is a nice success story for LGPL, which should be looked at by people from the "pragmatic / software is a tool / BSD is more business-friendly" camp. The supposedly "business-unfriendly" WebKit has created more jobs that many stimulus packages.
WebKit. You can thank the KDE developers for choosing the LGPL when using the browser on your phone.
JNI/JNA are not pure java.
Does that peripheral show up as a character device on Linux? You can open it and read it from Java.
Does it requires libusb on Windows? You can call that library from Java.
The peculiarity of Java is that it really strives to let you avoid os-specific code when it's possible, and it's currently the best platform at that. Of course its standard library can't possibly cover all possible hardware/software interations because they're not a finite set.
I wasn't seriously proposing the usage of a synthetic script from the 60s to encode a TLD for the USA. I was joking.
While you're at it, you can demand that domain in Shavian. It's a script that is native to English and it's already in Unicode; so you can stop using those foreign roman letters with inconsistent pronunciation.
"Eradicating languages and cultures" is not "compatible with western individual freedoms".
I believe computers should adapt to people's customs and not the opposite; current technology gives us the opportunity for this to happen more often than not, and localised domain names are a step in that direction.
From what I've heard from Nokia, it looks like they will provide the N9 [the successor of the N900] with the ability to switch at boot time between a "DRM" mode and a free mode; while in the free mode, you won't be able to access DRM-restricted media. How this will be accomplished exactly I don't know, and that's what effectively scares me a bit.
The problem is that manufacturers lock down their phones (even Google does that, on their directly-branded models), so their bootloader won't accept unsigned OS images. Some manufacturers go even further and lock down applications installation.
Even if MeeGo is as open as it gets, it has all the mechanisms in place to become as closed as Android is if manufacturers so wish (see MeeGo's security framework). Nokia has already shown with the N900 that they won't lock their handsets, but other manufacturers might have a different vision. I hope that the development of a large "independent" application repository will encourage them not to.
2) Extract it anywhere you like - it's self-contained. E.g.
3) export JAVA_HOME=/path/where/you/extracted/it
4) export PATH=$PATH:$JAVA_HOME/bin
I find this much more convenient than using (in my case) Ubuntu packages which add unwanted dependencies, scatter the installation in different places all over the filesystem and thus are less maintainable when things go wrong, are slow to track the latest released Java version, and periodically require me to run update-alternatives to avoid Sun's tools being replaced with not-so-compatible alternative implementations.
Note that OpenJDK *is* Sun's next version of Java, so when you switch to it because you have no "confidence in the quality of their software", you're not solving your problem.
I find TFA pointless. It basically says that Google won't be able to maintain a fork of Harmony (and only those bits which are actually required for Android!) without IBM's help, while at the same time stating that Google already has more people working on OpenJDK than IBM has.
Inexplicable.
Many Android applications already have native shared libraries built using the NDK. How do they manage compatibility and distribution now?
I'm not defending Oracle, just stating their point of view.