I think that, with time, we'll see more truly open, Nexus One style phones. I really wish they hadn't given it up so easily: something like that would be competitive in Europe where it's much more common to pay full price for an unlocked phone.
Sometimes, I think people forget just how cool open source is in terms of being able to take stuff and hack on it.
Maybe I'm being naive, but when I hear 'open' I'm thinking that that means that if I don't like a small part of my otherwise great smartphone, I can go in and modify the code that handles that specific part, compile it, and download the modified code onto my phone so that it does what it wants. That is several orders of magnitude of effort away from today's reality of:
- find a security flaw
- design an exploit of the security flaw so that I can root my device
- extract the drivers from the binary so that I can add them back into my new image after recompiling
- modify code and compile
- reload new image onto phone
I just feel that the open source community is getting sucked in to supporting this solution that is dragging us away from what would truly be an open platform. If Rubin was being honest about an 'open' Android, it would be under GPLv3, as RMS says. It's not, and the reason it isn;t is because it is not open. Rubin is spinning this, and I do not respect that.
> openness in phones is decreasing, not increasing - each new generation of Android devices seems to contain more and more locked-down proprietry code.
"openness in phones" took a huge leap forward with Android. 5 years ago, the idea of a widely deployed, almost completely open phone system would have seemed like science fiction. Just because some companies are adding crap doesn't take anything away from the core system, which continues to be free and continues to improve.
> the open source community is getting sucked in to supporting this solution that is dragging us away from what would truly be an open platform.
Which is, pray tell? Google has poured millions of dollars into Android, and we get that under a very free license. The "open source community" isn't some magical thing: it takes real work to make stuff happen, and Google is doing it with Android, including stuff like usability and GUI work that the "open source community" hasn't been so strong at in the past. So I don't really see some magical solution just wafting down from the heavens... it's simply not on the radar at this point. The only other remote possibilities are Meego and perhaps Symbian, but I don't really see much going on there.
I agree that we're still not seeing manufacturers do quite what we'd like, but I think it'll come with time, most likely in places that are not the United States: Europe and China most likely.
And, to be clear, there's not a snowball's chance in hell that it's going to come from Apple.
Before you can approach whether or not the software is free you must first define what free is. The parent poster is relatively safe in this because he provides RMS's definition of free, which is encoded in the GPLv3. Until you provide a similar definition I don't think one can state the software is either free or unfree.
I agree that we're still not seeing manufacturers do quite what we'd like, but I think it'll come with time, most likely in places that are not the United States: Europe and China most likely.
And, to be clear, there's not a snowball's chance in hell that it's going to come from Apple.
I think this is just speculation.
http://en.wikipedia.org/wiki/Open_Source_Definition - Android qualifies, except for a few bits and pieces.
> I think this is just speculation.
I'd bet a lot of money on it. Apple has a long, long history of making beautiful, innovative, forward-thinking, and fairly locked down products, from the Mac onwards.
OK, but you admit that your definition of free isn't everyone's definition of free. To be even more precise, the OSI itself doesn't even call that definition free, they call it 'open source.' The FSF, in fact, has a contrary definition that they maintain is actually free software. Both parties refer to the OSI as 'open source' and the FSF as 'free.'
This leads in to the main problem that Google/Android have -- when they market Android as being open they don't really define what open is. It is prima facie true that Android is not open on all fronts, so the question is really what Google considers "open" to mean.
Jobs put this observation into an interesting context because it's important to note that a lot of software in the iPhone is free as well. In the manual you will find a list (quite a long one) of all the GPL and BSD licensed software included inside the iPhone. The question, then, isn't who is open and who is closed, but what the definition of 'open' is and who more closely abides by it.
OK, but you admit that your definition of free isn't
everyone's definition of free
No shit ... my definition of "free" includes me distributing the code I made however I want.If it where for me I would include a new rule in the OSI definition that excludes GPL from being called "open source", because its copyleft extends to the whole package that links to GPL pieces, and for me this is not "free".
Both parties refer to the OSI as 'open source' and the
FSF as 'free.'
The Apache license has been approved as "Free Software", which is by no means the same as "free" ... an English word that you cannot trademark. It is prima facie true that Android is not open on all
fronts
In my definition of "open" that doesn't include forcing the phone manufacturers to not build locked phones.If my voice doesn't matter (I'm a nobody) here's the voice of Linus Torvalds (you know, the guy without whom you can't speak about Linux):
[Stallman] calls it "tivoization", but that's a word he
has made up, and a term I find offensive, so I don't
choose to use it. It's offensive because Tivo never did
anything wrong, and the FSF even acknowledged that. The
fact that they do their hardware and have some DRM
issues with the content producers and thus want to
protect the integrity of that hardware.
The kernel license covers the *kernel*. It does not
cover boot loaders and hardware, and as far as I'm
concerned, people who make their own hardware can
design them any which way they want. Whether that means
"booting only a specific kernel" or "sharks with
lasers", I don't care.
And I don't care about what Jobs says, the real question is: can you build your own iOS phone? can you participate in its development (like contributing bug fixes)? Can you choose phones from multiple manufacturers and multiple carriers? Can you install your own apps on it without going through that certification shit-hole?No? Well Android is a lot more open, regardless of definition.
http://www.gnu.org/philosophy/free-sw.html
Also, the FSF (and I assume by extension RMS) has no problem with the Apache 2.0 license which is a free software license and compatible with GPLv3.
And you can see this: 1) The market is strongly biased towards ad-supported aps. Who benefits? Google. Whose data is passed to the advertisers? Mine. 2) All of Google's ad-supported services are preinstalled on the devices. 3) Android handsets must be activated before use. I've paid for the hardware, but it's useless until I give more data to Google.
Google's business model is giving away cool stuff in return for milking you for information. I can't believe people think that this is "free" or "open".
Also, it is open source, so you've got a very good base to fork from to make something of your own, should you so choose. I think we'll see more of that in the future.
Sure, Google adds some stuff on top, but I think that's a small price to pay for all the free stuff you get underneath, which is more than enough to build your own Google-free system should you choose.
It's nowhere near RMS approval at all.
What else out there is closer? That has a remote chance of being successful?
I see it as a "perfect being the enemy of the good" situation. Android is the best, most open thing out there, even if it's not 100% open, and not 100% the best thing, either.
Completely irrelevant. If I say a donkey is not a horse and you then say "Yes, but it's closer than a manatee!" the donkey doesn't suddenly become a horse. Successfulness also really has no relevance to whether or not it's free.
What's missing? Some drivers and firmware? The RMS quote is, from above, "The shortest path to making it possible to run a smartphone without non-free software is to reverse engineer those non-free drivers or firmware and write free replacements", which to me indicates that Android is fairly close. He is not saying "the shortest path is to go write something else from scratch" or "gee, it's not really free yet, so I'll buy a shiny iPhone because... what the heck, I owe it to myself and I'm tired of harping about this free software bullshit anyway".
As a kernel developer I'll just say that my opinion is that this is a monumental task that says that Android is about as far away as any other embedded proprietary Linux-based device (including the Nokia phones).
Edit: That should be not including the Nokia phones. The Nokia Maemo phones are far easier to deal with than the Android phones.
First of all, I'm talking about the percentage of the source code and the system created from it, not about animals. We're all technical here, so let's not go off on tangents.
That said, it's a compromise. I want something that is open, and I want something that is reasonably widespread, and I want something that works well, because of the positive network externalities inherent in much of the software world. I'm ok with not 100% open and not the market leader - indeed, I'm fairly biased towards open, being an Ubuntu user (even though that is not 100%).
Beyond what I want, a completely free system is useless if no one produces actual hardware for it, so I think Android's compromises are good ones in that direction. Perhaps RMS does not agree, but that's his right.
What is so 'monumental' about the drivers and firmware that are missing?
I think you are discounting the value of the GUI and user space stuff, which is given to you under a free license with Android. Recreating that stuff would be a monumental task in and of itself.
Completely irrelevant. If I say a donkey is not a
horse and you then say "Yes, but it's closer than a
manatee!" the donkey doesn't suddenly become a horse
You can use a donkey to almost the same chores as a horse, you can even mate a donkey with a horse.So this particular point is completely irrelevant ;)
As a kernel developer ... Android is about as far away
as any other embedded proprietary Linux-based device.
You've got to give some concrete facts here, as this is too vague.I'm not going to play the semantics game, I'm relying on the reader to not be dense.
You've got to give some concrete facts here, as this is too vague.
OK, to give an example of a problem of similar scope, the first thing that has to be done is root the phone. Luckily, we have a canonical equivalent to bypassing the digital signed software -- Tivo. Tivo was basically the same deal. In order to load your nonsigned code into the firmware of the device, one first needed to find a security hole and exploit it. From there one would have to recreate the necessary drivers and such. This would essentially be equivalent to the effort to reverse engineer the Broadcom wireless drivers, which is still not complete (partially because I think they recently announced their intention to release at least some source code).
An example of a device which directly contradicts davidw's statement that Android is the most free smartphone is the Nokia N900. The development work needed to customize the N900 is a fraction of that of the Android platform.
davidw didn't say that. What I said is that overall, I think Android is our best bet for something that's free, but also "matters". OpenMoko is quite free, for instance, but doesn't really matter - no one will ever use it. I honestly don't have a good idea about what's in MeeGo, how free it is, how much it's possible to tweak the installation on the phone, and so on.
> I'm not going to play the semantics game, I'm relying on the reader to not be dense.
You are the one who brought the zoo into it. The rest of us were talking about mobile phone software.
Also, in terms of their DNA, a donkey is far closer to a horse than a manatee is, if you really want to get into comparisons of animals...
However, the notion that you can simply say that Android is more open than the iPhone, universally, is simply untrue. I'm attempting to promote a semblance of formal statement here, wherein the parties actually state their axioms as opposed to making otherwise semantically meaningless statements like 'Android is more open than iPhone'.
It's pretty simple:
Where can I download the source code for the iPhone's UI?
Does iPhone have anything like Android's intents that allows me to plug my own stuff in?
Can I install my own applications on my own damn phone without paying Apple?
There may be some metric where iPhone is more open than Android, but I can't think of it.
There are plenty of nice things to be said about the iPhone: it really changed the phone game, it's beautiful, well designed, and so on. But 'open' and 'free' aren't really words I'd associate with it.
Apache license gives freedom to developers, including freedom to take freedom away from end-users.
They are very good at this, and I do not fault them for it, but I also do not confuse altruism with what is clearly a profit motive.
Free software is cooler, open source is only kinda sorta cool ;p
"Everyone thought that Android would usher in a new era of freedom for handset users. Unfortunately, it just turned out that Android is more open not for users, but for Carriers to do as they please. Compare the crapware installed on Android phones vs. the iPhone. In a weird way, Apple has done more to push freedom for the (non-hacker) end user than Android has."
I think that should have said "git" instead of "repo"
There are a dozen dependencies not listed in that tweet to build android. JDK5.0-12, flex, bison, libncurses, etc etc etc.
1) I just tried that line out on my Fedora machine - it doesn't work, doesn't know what repo is.
2) after running 'make' does my computer turn into an Android device?
3) If not, what is the hardware I have to use
4) where are the instructions for loading the newly compiled code onto my device - I see a make, but no make install... I mean I like binaries as much as the next geek, but they aren't terribly interesting unless they, you know, run.
5) I tried loading the binary on my device, but it refused - something about unauthenticated code - what should I do to use your wonderful open system?
OK, obviously I'm trolling a little here, but Andy's tweet is every bit as much a troll. I doubt very much that RMS agrees with Andy as to the openness of Android... In fact, I just went and looked, here's his thoughts on the subject:
Android's source code is free software, but in many phones the binaries of Android are not free, because the phones are set up to refuse to run modified versions if the user installs them. [This practice is called ‘tivoization’, named after the product that pioneered it.] If the software in Android were under version 3 of the GNU General Public License (GNU GPL), users would be guaranteed the freedom to install their own modified versions.
Even when Android is not tivoized, it needs non-free drivers or firmware to run. As far as I know, no smartphone is made that can be run without proprietary software. None respects its users' freedom.
The shortest path to making it possible to run a smartphone without non-free software is to reverse engineer those non-free drivers or firmware and write free replacements - RMS
So that would be a 'no'