You can't build a Truly Free Version of Windows Phone. You can't build a Truly Free Version of iPhone.
But if Android doesn't do what you want, then you can fork it, tweak it, and turn it into something that does!
I think Google is doing what's best for Android, but Android is not open source anymore.
Firefox OS might be a better choice for open source comminity to spend time on. I haven't been following latest development, how's the hardware situation on FF OS?
I actually have an older phone (a first-gen Galaxy S) on which I installed Cyanogenmod without Google's services, which is basically stock Android. In spite of opinions to the contrary, the phone is quite usable without any of Google's services on it.
I've got Firefox for web browsing, OsmAnd for maps, Apollo for playing music. All 3 are open-source apps. At some point I even installed Amazon's App Store on it, to be able to get some proprietary apps, like Facebook, installed. But then I realized I don't need all that bloatware, as the web versions work fine.
You should try out this experiment if you've got an extra Android lying around. After you find apps suitable for whatever it is you want to do, it really is OK. The level of polish is not the same as with the Google-enabled Android, but it's still much, much better than the experience you get on low-budget non-Android smartphones.
People forget that Amazon already forked Android and that without Android they wouldn't have been able to deliver Kindle Fire. That's the power of open-source, even if you choose to deny it.
"For a Linux user, you can already build such a system yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem. From Windows or Mac, this FTP account could be accessed through built-in software."
This might be a fun experiment for a developer. But for the broad user perspective even replacing launcher is a big deal.
The point is, Android is currently the go-to mobile OS for anyone who wants to scratch their itch. like Linux has been on the desktop. And there have been plenty who have taken up this offer (e.g Amazon, Chinese & Indian manufacturers, Android console makers), except they aren't end-users.
I'm actually curious as to why F/OSS applications that aren't extremely high profile or commercially backed (Firefox) seemingly tend to have UI/X as what is probably the last possible priority ever.
Yes, the renderer needs some work both in terms of prettiness and speed. It's not as polished as Google Maps, but it's definitely usable. That the map data is from OSM is a feature by itself, OSM maps are often more detailed than what you get from Google, and you can download all data, which is great for traveling. It even supports navigation.
Your argument above is in the same line as people claiming that end-users do not benefit from open-source. In some sense it is correct, as end-users do not have the technical know-how to assemble their own operating system, to make changes to source-code, etc... But practically speaking, end-users don't have to do it by themselves just like they don't have to fix their cars by themselves (e.g. I helped friends install Cyanogenmod on their phones, which is a PITA to deal with, but once done, it's done).
The power of open-source is in the ability to fork it. Not all people or companies can do it, but if there's demand for it (e.g. Google's stewardship goes awry), then a successful fork will happen. Building an operating system from scratch is hard. Having stock Android available as a base for new operating systems is an incredible gift.
In the meantime, yes, forks of Android are only for hackers wanting to play around. But if this leads to the rise of alternative marketplaces, like F-droid, then this will benefit everybody. Not sure if you notice it, but the biggest problem that Android has in terms of openness and lock-in is Google Play. Fortunately Android allows installing software from third-party sources, but unfortunately there aren't many alternative marketplaces out there.
I think it would make a lot of sense if the base Android OS (AOSP, as I understand it?) were only as functional as it needed to be for useful apps to run, and then the "Google app stack" or "Cyanogen app stack" or whoever else's, open source or closed source or anything in between, could be dropped in to provide the user-facing functionality.
The key seems to be getting to a point where important pieces of that underlying functionality aren't wrapped up in the "Google app stack".
One example is the location stack. They have moved into it play services and added a few features, but the open source one is now going to bitrot. 4.4 has now a new closed source launcher, and I'm sure the old GoogleHome.apk will go bitrot in a couple years.
If you get AOSP today is quite far from an Android device. Surely it's better than nothing but the distance is widening every month as Google keeps going down to this closed source way. In another couple of years AOSP would probably be irrelevant if you also don't subscribe for the Google apps when you build your device.
Look just what a big upgrade the photo editor in the open source gallery received:
http://www.engadget.com/2013/11/03/android-4-4-kitkat-comes-...
There's something deeply hypocritical about complaining Android is not open source on the basis that Google is not working on the components themselves anymore, giving you free updates. If you actually sincerely care about open source your concern is exactly the opposite of that: the community must be able to take the source and build upon it. Google has brought us light years, billions of dollars of investment along the road to where we have a beautiful, highly optimized, slick operating system available completely for free, even to Google's direct competitors. And your complaint is now that you're not getting fed free updates to all the exact components of the OS at the rate you want? And therefore all that was given to you is worthless, "rotten", not even worthy of the name open source?
This just sounds so entitled to me.
But notice that the problem is not "the rate of free updates", but rather the fact that Google is moving more and more development into the closed-source part of the operating system, while still trying to maintain the market line that the "operating system is free". But which operating system is free? If you install AOSP on your phone (assuming you can and manage to go through the hurdles) you end up with a phone which is completely different from stock Android. That is my complaint: the difference between AOSP and the so-called "pure Android experience" (as on a Nexus) is widening more and more.
You say Android is available for free to "competitors". But as a matter of fact, AOSP (not Android) is available to competitors, and only those that are not part of the OHA. If you are a member of the OHA (so if you are an Android vendor of any significance), your contract says that you cannot fork Android; not only that, but you cannot even produce/manufacture a forked Android device as a supplier for an unrelated company outside the OHA (which is why Amazon has problems producing Kindle Fires, since all major manufacturers are also trying to sell consumer devices based on Android).
Moreover, if you fork AOSP, not only you don't have access to the Play Store, but many (most?) .apk won't even work on your device, since they rely on the proprietary Google APIs which are not part of AOSP. And this, again, is a widening ongoing process, so more and more (rather than less and less) existing .apk found on Play Store are incompatible with AOSP.
I think the way Google cleverly designed the whole "AOSP vs Android" licensing is way more than a standard freemium model; it's a model in which nobody pays, a part is open-source, and still it's close to impossible to create a forked consumer device that uses AOSP and benefits from the existing ecosystem.
The markets in which AOSP is really making a different (in my experience) is the B2B embedded device market, which is not in direct competition with Android devices, and where AOSP is quickly moving to become the de-facto Linux embedded distro. I work in that field, and I'm grateful to Google for AOSP, as it's really making the whole market progress faster than ever.
It honestly sounds like you have never installed AOSP yourself. I do it all the time and quite often use the device for a bit without any "Google". The experience of running barebones AOSP is way, way closer to stock Android than anything you will buy from an OEM. I even sometimes do it and forget to install the Google apps because it is so similar you cannot actually tell the difference until you try to use something like Gmail or Google Now. The idea that consumers buying branded phones are somehow being deceived because they think they are getting stock Android but Google has somehow slipped proprietary code in is laughable when you look at what the OEMs do in comparison.
I would challenge you to get hold of a Nexus and flash plain vanilla stock Android (to make things easy, use CM 10.2 and don't put Gapps on there). These are built from source directly derived from the AOSP, by people who are NOT members of the OHA and have no special relationships with Google, manufacturers or anyone else. You will find you can set up your email, calendar, pretty much everything you need, with the exception of Maps is available. It may make you feel better about things.
In my view, what is happening lately is that Google is drawing a sharper boundary around what is open source Android and what are proprietary Google services. The OS is remaining open source, as a hard condition. Things that belong in the OS are staying there and staying open source. But the layers above that - things that are rightly and properly user space apps - are only delivered as part of the OS as a convenience. Making better versions of those - possibly better versions than can ever be made open source, either because they depend on proprietary / live Google services (eg: maps) or because they use licensed or patent encumbered code - is fair game.
Of course the entitlement is theres because they convinced OSS devs to write code for them for free under the "do no evil" moniker.
So now if they want to keep their toolchains, they have to continue maintaining the oss-side google service apps to regain access to the tools they help built.
I think this is called bait and switch, and it's probably the #1 pattern used by big technology firms today. It's so easy to "switch" with IT. Carrot-stick pattern worked well for Microsoft, because they owned the only stick and most the bait, but no longer. I guess without a monopoly, the only way to secure customers is to fool them.
The location stack has not moved into Google Play. The location sensor APIs are still AOSP code. Address resolution and other things have always been tied to Google services, so that geofencing and Google Maps specific APIs are in Google Play Services is not surprising nor unexpected.
> 4.4 has now a new closed source launcher, and I'm sure the old GoogleHome.apk will go bitrot in a couple years.
The previous launcher from 4.3 is still available in AOSP. The new Google launcher has high integration with Google Now, but we've already seen that the launcher space has a thriving amount of competition and an open source launcher is totally do-able.
> If you get AOSP today is quite far from an Android device. Surely it's better than nothing but the distance is widening every month as Google keeps going down to this closed source way. In another couple of years AOSP would probably be irrelevant if you also don't subscribe for the Google apps when you build your device.
This is the core point. Its not that AOSP gets less free over time, its that Google-specific code and services get more functional over time compared to AOSP components. Since Google controls code commits for the main Android project, it means people have to fork and generate alternatives like Replicant is trying to do.
http://arstechnica.com/gadgets/2013/11/google-just-pulled-a-...
This leaves it unclear what you get for a launcher with an AOSP build of KitKat, but it certainly is drifting farther from the official Android experience.
You can (and I have) fetch the new Launcher from AOSP, compile it and run it very easily as a separate application. The Google Experience Launcher is a closed source Google app based on this which integrates Google Now.
https://android.googlesource.com/platform/packages/apps/Laun...
I now know which app I'm ditching once I get a 4.4 device.
I'm a huge fan of Google Now so hopefully I can continue to use it if I decide to switch to a custom launcher.
Not correct. Launcher3 is open source.
In theory you can, in practice 99.99% of the people won't be in a reasonable position to even try it.
By the way 99.99% of all android users is quite a crowd, it could fill a few cities, so if they do check the code it can and will influence other android user, if only by adding pressure on Google. You see, adding a backdoor to android is really much harder for Google than it would have been with closed source, maybe 99 times harder?
So no, you absolutely cannot do this with what almost everyone thinks of as 'Android'.
You can do it with a different thing that also has the name 'Android', but isn't the same.
I assume you haven't noticed this, because otherwise you would be deliberately trying to spread a misleading view.
Regardless of what Mozilla does, the manufacturers are not going to sell non-differentiated phones for very long. They are going to keep that differentiation closed source, as they do now.
If one of them becomes very popular (say, Samsung), the Firefox OS you can do things with will be nothing like the Firefox OS on consumer's phones.
This is the way of the world, not just the way of Android.
I note that you are implicitly conceding that what is commonly considered to be Android is indeed not open.
No, instead it's licensed in a way that lets the rest of the world decide what they want to do with it, and the complaint is "we don't like what the world is doing with it". You must also realize if android was GPL'd, nobody would have used it, right[1] ?
" and Google itself builds its most significant end-user functionality as closed components."
I don't agree this is true, but even if i did, if you want to compete, compete. If you don't, don't. Essentially this whole complaint boils down to "i have to do my own work in order to compete seriously, i don't get to freeride anymore after x number of years".
I'll point out the mariadb folks did this, rather than simply complain.
In any case, the fact that someone may build closed source stuff on top of an open platform does not make that open platform less open. It may make that platform less useful as time goes on, but not less open.
"I note that you are implicitly conceding that what is commonly considered to be Android is indeed not open."
I haven't conceded anything, actually.
It's just not a point worth arguing with someone who holds a position that clearly conflicts with mine and is highly unlikely to change it.
[1] I wouldn't try arguing with this, this is a case where i was there, and I literally know the answers and positions of the phone makers, which was essentially 'never ever ever'.
DannyBee - have you read Moral Mazes by Robert Jackall (http://www.amazon.com/Moral-Mazes-World-Corporate-Managers/d...)? It was one of Aaron Swartz's favorite books.
It's an excellent, and fair to all sides, dissection of morality in corporate management. It explains how and why managers rehearse to explain actions their company's have taken.
In this particular comment of yours, you say lots of true things. However, you are not acting in a straightforward "hacker" way, as befits this site, telling the truth how it is.
Google can continue to behave in this way as a company, but it will alas find it harder and harder to hire good geeks.
For the record, I'm specifically not accusing DannyBee of dishonesty. The reason I mentioned Moral Mazes as a book is because it explains what happens from the point of view of the managers, and why it comes across to outsiders as dishonest.
This stuff is subtle and systemic.
Fair point about speaking for the site.
I'm not even sure where to start with your comment, because it implies so much, and has so little to offer.
Rather than address it point by point, i'll just say things:
When i'm speaking for the company I work for, you'll know it, because it'll be a press release that says Google on it. Maybe you are unable to separate your personal and professional lives, and hold opinions separate from your corporations. For me, Google is a company I work for. While I love my job, Google certainly does things i don't always agree with. When it does, i'm certainly not going to avoid saying something because i am "rehearsing to explain the actions my company has taken".
There are a lot of things Android could do better. But in the end, what I see over the years is a lot of whining that it's never enough. From where I sit, Google has taken a lot of flack for actually pushing the ball forward, because it doesn't always go as far as the open source community wants. First they wanted Google to beat up the carriers (without understanding why this is pretty much impossible). Then they complain when we don't. Heck, some people complain android is "too open", because we can't force carriers to give people what they want. I could go on here for hours. Look at the complaints of a lack of a nexus 4 and nexus 5 on certain carriers . Google releases an entire platform, on which other open platforms are now based, and people complain that they have value-added apps that provide a lot of functionality, and haven't opened them. Well great, good for them. This is supposed to be an innovative community. Create your own. Do it better. Do it open, and win. Prove that they made the wrong choice.
Let's stop and be honest for a second. Do you really think something like Firefox OS would even exist (not even just because it is/was based on Android at the lower levels, like Gonk) in the market today if Android had not pushed the ball forward?
When Android was first released, one of my good friends told me no matter what Google did, people would never see it as enough, no matter how open it was. The saddest part to me of all this is that he was right.
Android was never meant to be the end solution, it was meant to be the beginning.
All the hard work on Android is much appreciated - I'm really glad Apple didn't gain a monopoly.
But the mobile web is inevitable. Eventually there will be parity in performance between web apps and native apps. (And Android is on the wrong side of that divide, and Dalvik doesn't exactly help.) Even if Firefox OS isn't doing what it is now, someone would have tried to push for a mobile OS focused on web apps. Perhaps even Samsung.
But.. You have certainly claimed to have authority to speak about parts of Google's strategy because you were in certain meetings.
These are very different things, and you should not confuse them.
As i've said before, everything i've stated about these strategies is public knowledge, and often stated by Andy Rubin himself, just not always believed :)
All i've done is basically say "believe it".
You seem to be trying to wriggle out of this by saying that what people mean by Android today has moved past the 'open beginning' that is all that was ever meant to be open, and that's not Google's fault.
I agree that this isn't Google's fault. But let's speak plainly. Android began as open and that was all that Google intended. Things have changed since then and only a part of Android is open anymore.
Why try to claim otherwise?
This is why I said that. You have your belief. You are clearly unlikely to change it no matter what I say. Your goal in this is not to possibly change your belief, but to find a way to confirm it in what i'm saying. This is what people do in these situations.
That does not often generate useful discussion, and it hasn't really here.
Realistically, I haven't tried to wiggle out of anything. It just turns out, as I suggested, we strongly disagree over whether Android is open, whether it was meant to be, and whether it still is.
Rather than continue this unproductive debate, i'm just going to go back to hacking some code on this fine sunday :)
How much time did using this port from BSD take over shipping with a GPLv2/LGPLv2.1 version of libc and coreutils?
(Note, I'm explicitly referring to the Unix base operating system, not drivers, audio or video codecs or libraries or anything else, only Bionic and the core Unix utilities)
Your comment about 'You don't get to freeride after X years' is perfectly reasonable if google's goal is to stop contributing to the open parts of android.
I am not disputing Google's right to do whatever it takes to compete, or that there is somehow something morally wrong about that.
I am disputing the canard that 'Android is open' when really only a part of it is, and that part is not what most people think of as Android.
I do thing there is something morally wrong with perpetuating this misleading idea, even though it furthers Google's PR goals.
I know that corporations mislead us all the time with communications that are technically true but pragmatically false. I realize it may be naive to expect different behavior, but it's certainly not going to stop if it isn't called out.
This is because of what you define Android as.
Of course, the humorous part is that when all of this first started, Google tried very hard to make sure Android was defined in terms of what the Android Open Source Project was and had, but failed miserably here, so they gave up.
If as you say, Google intended to have 'Android' mean the AOSP at one point, and it has stopped trying for that definition now because it failed to make the distinction to the public, then that only strengthens the point that Google (and you) know that what people think of as 'Android' is not open, and so are knowingly exploiting the ambiguity between 'Android' and the AOSP when you perpetuate that idea.