Ex OpenMoko Lead System Architect takes on Android's lack of openness

Two days ago I centered my frustration about the lack of openness in mobile handsets with Google's new operating system, like the T-Mobile G1, in the headline "WARNING: Android devices are NOT open". But now open source wizzard Harald Welte, former Lead System Architect for OpenMoko, explains Android's shortcomings much better in his latest blog post:

To me, those things are not a big surprise. As soon as you try to get in bed with the big operators, they will require this level of control. Android is not set out to be a truly open source mobile phone platform, but it's set out to be a sandbox environment for applications.

And even with all the android code out there, I bet almost (if not all) actual devices shipping with Android and manufactured by the big handset makers will have some kind of DRM scheme for the actual code: A bootloader that verifies that you did not modify the kernel, a kernel that ensures you do not run your own native applications.

He sees Android as little more than some sandbox virtual machine environment where people can write UI apps for. Nothing that gets him excited. "I want a openness where I can touch and twist the bootloader, kernel, drivers, system-level software - and among other things, UI applications", he says. And I want that too.

To Harald most Linux handsets don't deserve their name because all the freedoms of Linux software are stripped. Linux on cell phones is "definitely not to any benefit of the user" - but only to handset maker, who can skip a pretty expensive Windows Mobile licensing fee. That brave new world makes him sick.

I guess only on Nokia Internet Tablets the Android can be as open as we whish. It's time that someone takes the source code an ports Android for them, preferably without Google spyware, as we know it from Iron, the googlefree fork version of the Chrome browser. Until now Android only runs in a virtual machine on Nokia Internet Tablets.

WARNING: Android devices are NOT open

The world is excited about Google's new operating system for mobile phones. But I am not such a big fan anymore since I talked to Rich Miner, Head of Mobile Platforms at Google. We met yesterday in Munich at the Communication World 2008 where he presented the Android. The system is not as open as I hoped. People will need to jailbreak and unlock their Android smartphones, like iPhones, if they want true freedom and make the most of them.

Of course the Android source code is free and you can develop everything based on it. But then? You don't get your new OS on that Android phone!

If you want, you can redesign the entire Android operating system and eliminate every Google function, even the kill switch for unwanted software. Our German readers at Areamobile.de are worried about their privacy, many refuse to connect to Google's servers with their location aware Android phones. But the first device on the market, T-Mobile G1, does only work with a Google account. "You must sign in to your existing GMail account before you can do ANYTHING with the phone", writes jkOnTheRun. Big brother could be watching you! Therefore it would be preferable to have the choice to get an Android phone that's free of Google. I want my own GPhone, not a Googlephone but a Goebelphone!

Rich Miner shows his T-Mobile G1 at Communication World 2008 in Munich

"That's perfectly possible", Rich Miner told me. You can do whatever you want with the source code. "But will I be able to install my own Android version on a T-Mobile G1?", I asked. "No that's not possible", he answered. "You would have to change the ROM." People cannot change the Linux kernels of the Android devices which went on sale today. They can install every additional program, but Google controls the core system on the ROM. That's against the Linux philosophy and a big difference to the other open Linux device, Openmoko. As a Linux user I am used to bake my own kernels. I remove kernel functions that I don't need to make my computer faster. Or I add new features, such as virtualization, to the kernel.

With Android devices that's not possible. Only a handful of developer devices can do that, but they are not for sale to end users. So if you want to run your own fork of the Android operating system on a cell phone, you have to get a rare developer device or become a handset producer like Motorola or HTC. That sucks! Also: The marvellous G1 is locked to T-Mobile's network in the US and doesn't work with German SIM cards.

My take: It won't take long until we see a flourishing jailbreak and unlock scene, as we already know it from the iPhone. The Android system is not really open before I can bake my own kernel for the device and use it on every network. I asked Rich Miner what Google thinks about jailbreaking the G1 and he just returned: "Why would you want to do that if you can install every software?" Unfortunately I didn't find a good answer in this moment. But I should have said something like: Because it's human to reshape devices for unintended use. It's part of our DNA since the first monkey realized that fruits are not only food, but can make a good booze if you let them mature a little longer.

Mike Jennings presents the Android SDK at OSiM World in Berlin

Maybe Google doesn't even know what's coming their way, sometimes they are surprised at what people do with Android. That's what I learnt from my interview with Mike Jennings, Google's Android Developer Advocate, on September 17th at OSiM World in Berlin. He was astonished when I told him that I was running Android on my Nokia Internet Tablet since Juli. "That's not possible because the source code is still not free", he said. But yet there was an idiot-proof installer available on the internet.

OpenMoko urges Android to release everything as source code

While I was in Barcelona at the Mobile World Congress 2008, I received a message from Sean Moss-Pultz, CEO of OpenMoko, producer of the world's most open mobile phones. Their Freerunner, Neo 1973 and Dash Express devices use the open source operating system Linux and people can install every software they want on it with just an apt-get. Where other companies have a Linux kernel with a locked proprietary stack on top of it, the OpenMoko phones are open from top to bottom. You can use your own tools, compile your own kernel. Everything barring a few small drivers is open source under GNU General Public License (GPL).

In Barcelona I talked to a representative of the LiMo foundation who doesn't want to be quoted with his name. He revealed that LiMo Linux is in fact a closed shop. The only aim of the LiMo companies is to produce cheap handsets with a versatile operating system that doesn't cost them a dime. Most end users won't even notice that it's Linux because they are not allowed to install any software for „security reasons“. More open is Android of Google's Open Handset Alliance, he said, but the most open system is OpenMoko.

Read what their CEO Sean Moss-Pultz thinks about the Android and other actual developments! He has answered my interview questions by email.


What's the actual status of Openmoko?

We limited our production of handsets for developers. Our goal was to sell a small number to an enthusiastic crowd. When we sold out in 3 days, we realized that we need to build another batch of phones. From our standpoint the developers are engineers in our company. We don't have more than 10,000, currently. As the project became more noticeable we got inquiries from many different directions. So, for the next release of FreeRunner we will plan for more at the start and get ready for ramp up of consumer oriented products. The coolest thing about a Neo is what it doesn't do. It doesn't lock you out. It's a GNU/Linux computer. It does what you want.

What is the most difficult thing about building a GSM phone from scratch with Linux?

OpenMoko started as a project inside FIC (Taiwan) and has recently been turned into it's own company. So, we've had to build a open software stack, build a team, build a product, plan a future, and build a company, while everyone gets to watch.

Most of the challenges, I would say, are philosophical in nature. Not technical.

In an closed company you go through various stages of hardware development. These stages are hidden from the general public. Prototype hardware is built and passed out to a few select internal developers. Later, more hardware is built. In a pilot run. This is distributed internally to more engineers in the company. Finally, it's presented to the public like it just came out of the oven.

OpenMoko is inside out. Our prototype are shared with developers around the world. Why? because our engineers are outside the company as well as inside.

In some ways we are like a reality TV show. Showing how one builds a gadget of the future. It's like a cross of Survivor, Dirty jobs, and ice road truckers.

What can your handset do, the Neo1973? Give us some specs and tell what is so great about it, compared to normal mobile phones!

The hardware specs for the phone are on openmoko.org, but in short its a GSM phone, with GPS, WiFi, Bluetooth, 3D graphics acceleration and accelerometers. But this is not a spec war and this is not a beauty contest. Neo branded phones unlock the hardware and unlock the software. The software on the phone, the applications you use are totally open. If you program you can change them. If you don't program you will download popular programs that others have built.

Which people do already have such a handset, the Neo1973? Can people buy it somewhere? Is it planned to be sold to the general public some day?

Our Neo 1973 has been onsale since July of 2007. We're totally sold out now and won't make anymore. Our next produce Neo FreeRunner is coming soon.

I think we primarily appeal to technologists now and will stay this way for the some time to come. But this is not at all limiting. What I find most appealing about OpenMoko is that we just have to provide the right framework for change to happen. Our community is the one that pushes the revolution. The more people that develop for this platform, the larger the target market becomes.

We've found a strong and influential niche. In the age where mass market TV advertising is dead, this is crucial. Marketing a product initially to the masses is impossible, I would argue. People have too many choices and not enough time.

OpenMoko stands out because we are different. We speak directly to the needs of an extremely creative group. Our goal is to provide them tools and inspiration so they can realize the have power to revolutionize the world. OpenMoko is a company from their community. We will amplify their voice.

How did the OpenMoko project start? Whose idea was it and who pushed it forward?

About a year and a half ago I was a product manager at First International Computer (FIC) -- a large Taiwanese OEM, charged with Defining the product roadmap for our division. I was quite limited, unfortunately, by having to create only Windows Mobile devices.

I quickly realized that it didn't need to be like this. I could step outside the box any time I liked.

The phone that I really wanted to create was the modern equivalent of the programmable calculator. A device that is simple to use, but almost infinitely extensible by the end user.

The main idea was that the driving force behind the Internet is Free and Open Software (FOSS). It's the superconductive medium that powers the Internet. Without FOSS the Internet would be trapped in 1995.

In a nutshell, OpenMoko is about spreading this technology to mobile devices. Letting people everywhere "Free your phone." It's the birth of a new Internet. Not merely a revision, not simply 2.0, but rather a connected, interconnected experience wherever you wander.
Without FOSS the mobile phone still lives in 1973.

Our first open mobile device is called the "Neo 1973". It's internet connected, location aware, and completely FOSS. In a way, I see the mobile industry as a matrix. The industry is hindered by proprietary systems and Neo is trying to tear down the walls unplug humans from the matrix and give them back their power.

I took this idea and pitched it to our senior management team with the help of Timothy Chen -- a very smart businessman. Without him, this whole project would only be another unfulfilled entry in my sketchbook.

How is Openmoko organized today? How do you include all those Linux developers worldwide? Is there a boss or a structure?

We are a fully independent company now:

http://biz.yahoo.com/bw/080102/20080102005213.html?.v=1

Lots of people trying to make great open mobile devices.

What is your opinion about Android?

We support FOSS endeavors. In our philosophy, a software platform needs to be free from the iron to the eyeballs. That means FOSS code from the lowest levels that talk to hardware to the highest levels that present images to users. We hope Android moves in this direction. We encourage them to.

Is Openmoko involved with Android?

Yes. We lobby them to join the FOSS movement and release everything as source code, all the way down to the drivers like we do.

Has Openmoko been contacted by Android?

Yes.

Is Openmoko source code involved in Android?

Our source code is freely available GPL.

Maybe Android source code is involved in Openmoko?

If they posted GPL code and our community found that it was good and useful it would get used. That's what FOSS is all about. This is how FOSS gets stronger for everyone.

How does the launch of Android affect the Openmoko project?

It's rather humbling. We never expected a company like Google to endorse our concept of freeing the phone. It's also exciting because we realize that with their support of developers many new open applications will come to the FOSS platform.

Why did Harald Welte leave? He was Openmoko's "Senior Software Archtitect System Level" and we were very proud that he was from our city, Berlin.

Berlin should be proud. Harald is a great programmer and was key to getting The first phone, Neo 1973, shipped. He did a huge amount of work and is still a part of every Neo that ships.

What does this change?
(http://gnumonks.org/~laforge/weblog/2007/11/16/#20071116-leaving_openmoko)

We have our goals which Harald helped create, and we are meeting those goals and going beyond them. We are prepared because of his diligence.

How do you see mobile communications in ten years? Everything seems to "open" now. Verizon and AT&T open their network, Google bids for an open spectrum at 700 Mhz, Mobile Wimax promises more open mobile communications...

You could say there were two theories. In one theory the user pays for bandwidth (time on line) and the device (phone, set top box etc) is free. In the reverse world, bandwidth is free (like free WiFi) and people buy great devices.

In one world you pay premiums for bandwith, in the other you pay premiums for devices.

Opening the network...levels the playing field and gives people more choices. This is what Openmoko is all about.

To make money from mobile VoIP, companies have to accept certain realities

Jon Arnold has updated his very interesting portal website IP Convergence TV. This time I also wrote a guest opinion, because to make money from mobile VoIP companies have to accept certain realities: "WiFi isn't everywhere and callback costs double".

I love the mobile use of VoIP but I still find it quite uncomfortable. That's what I point out. Especially annoying is how Skype, Fring, Truphone and other SIP based VoIP services get blocked by German 3G providers. Sorry, Dean Bubley from Disruptive Analysys! The reality looks much darker for VoIPo3G than you predict for the future. (But thanks for your regular Google ads "3G mobile Voice over IP. Analyst report: is it a threat to carriers? Or a future opportunity?". I better put a direct link to your website.)

Mobile VoIP over Wifi works only at home or in the office where I don't need it. So in my guest opinion I advocate intelligent cell phone software which automatically completes calls as callback, callthrough, Vo3G or VoWifi while the user doesn't even notice. I have already installed an example software on a Nokia E61.

Maybe if more and more people use these options, Dean's dream will come true. If everyone uses only mobile callthrough, triggered by intelligent software on the handset, the mobile network operators cannot charge any other items than the tariff's included minutes for local calls. Their voice legacy cell phone networks would become dumb pipes into the internet, the way we already see it with the 3Skypephone or iSkoot, Ringfree, Mobivox, Jajah Direct, Sipbroker, Tpad, Rebtel, Mobiletalk, etc. If mobile operators wanted to charge for international calls at all, they would have to embrace VoIPo3G and could at least charge for data, the way Dean predicts it.

But until this comes true, the mobile VoIP companies should attack the incumbents with better callthrough options, to take more and more cell phone calls out of the traditional networks and into IP. Read the full text for further explanations!

Next monday is the real Google Phone day

I am really looking forward to next Monday when the software development kit (SDK) for Android, the new operating system from Google for mobile handsets, comes out. Hopefully that brings answers.

When they launched the Android last Monday there was very much buzz but few to be seen. Valleywag pretends to have screenshots of the first Googlephone app, but that's not very much of an information.

On tuesday I interviewed Florian Seiche, Vice President Europe of HTC, the most important smart phone producer in Google's Open Handset Alliance. But whenever I digged deeper he said that we have to "wait till monday when the SDK comes out".

At least he could tell me that Android has nothing to do with Openmoko or other Linux versions for smartphones. No Openmoko developers worked for the Android, no code sharing or whatever. Before I had the suspicion that Google's new mobile OS was in fact powered by the community solution Openmoko.

But that's not the case. Even Sean Moss-Pultz, initiator of Openmoko and responsible hardware product manager at First International Computer (FIC), doesn't have much knowledge about the Android, he told me in an email. He doesn't know the code yet and is waiting for something to compare.

So let's wait for Monday, November 12, 2007! That`s the real Google Phone day.

I think Google's Mobile Phone Platform Android will be great

So the Google Phone is out and the first reactions are not too good. At least at GigaOM there is more criticism than kudos. Nobody wants to hype the new product, nearly everyone is nagging. Since Google's shares are worth more than 700 dollars it's not cool anymore to be a Google fanboy.

I got an invitation to the same press call like Om, but unfortunately it started when my workday ended. May other journalist cover the story. Also it seems that the press call was not very much of a pleasure. "They completely dodged my question about how does it reconcile with other mobile linux efforts which are backed by none other than partners like Motorola", writes Om Malik.

Personally I like very much what he tells about Android, Google's new mobile phone plattform:

What is Android? A fully integrated mobile “software stack” that consists of an operating system, middleware, user-friendly interface and applications. It will be made available under one of the most progressive, developer-friendly open-source licenses, which gives mobile operators and device manufacturers significant freedom and flexibility to design products. Next week, the Alliance will release an early access software development kit to provide developers with the tools necessary to create innovative and compelling applications for the platform.

Does anybody know if this has something to do with OpenMOKO, the other open Linux cell phone platform? Maybe Android is just the same?

How open is Android compared to OpenMOKO?

The latter let's you manipulate everything to the very core of the mobile phone functions. Yet now there are thousands of great free Linux programs running on the OpenMOKO devices. I would love to see this kind of openness backed by heavy weights like Google and the other mentioned companies.

I hope that Android is as open as the Open Handset Alliance's website says:

Android was built from the ground-up to enable developers to create compelling mobile applications that take full advantage of all a handset has to offer. It is built to be truly open. For example, an application could call upon any of the phone's core functionality such as making calls, sending text messages, or using the camera, allowing developers to create richer and more cohesive experiences for users. Android is built on the open Linux Kernel. Furthermore, it utilizes a custom virtual machine that has been designed to optimize memory and hardware resources in a mobile environment. Android will be open source; it can be liberally extended to incorporate new cutting edge technologies as they emerge. The platform will continue to evolve as the developer community works together to build innovative mobile applications.

All applications are created equal

Android does not differentiate between the phone's core applications and third-party applications. They can all be built to have equal access to a phone's capabilities providing users with a broad spectrum of applications and services. With devices built on the Android Platform, users will be able to fully tailor the phone to their interests. They can swap out the phone's homescreen, the style of the dialer, or any of the applications. They can even instruct their phones to use their favorite photo viewing application to handle the viewing of all photos.

I think GigaOM's reader rohit understands it right:

i think this is a much bigger potential play at replacing the whole mobile phone software stack and aimed at making it truly an information appliance. think of it as an IP-services led “phone” design, not a telco-call based device.

It's a Linux for phones! You can do everything with it, if it's really open. I already wonder how it cooperates with Google's Ubiquisys femtocells. It annoys very much that my cell phone is not as open and flexible as my PC. Give me a command line to my cell phone and I will be happy!

Or, as commentator David Jacobs puts it:

Being an open system, hackers will have a field day with this and it could get some serious traction among the geek community who are so frustrated with the iPhone limitations.

Here you can get more quotes from Android's developers:

"Even A teenager in the basement and a senior designer in a big company - they have the same chance", says the film. That would be great because I don’t want just a Google Phone. I want many different of them for different purposes. That’s why I think the OS approach is great. The iPhone just isn’t enough anymore. It’s so 2007.

I got the offer to do interviews to John Wang, Chief Marketing Officer of Google's hardware producer HTC, and Florian Seiche, Vice President Europe of HTC, tomorrow. Let's see if that will answer my open questions.

Interesting interview with OpenMoko

PC World has a very interesting text and interview with the makers of OpenMoko, the open source "anti iPhone". Most interesting: After GPRS, built-in GPS, Wi-Fi, and Bluetooth they will add 3G next year.

We plan on going 3G next year. This first device uses existing Gerber designs from our feature phones to lower the costs. Remember that this is an unlocked, nonsubsidized phone. So $300 is actually incredibly cheap. We thought this was super-important, to help with early adoption.

There is a nice list of yet available applications, but there is more to come since the entire open source world is invited to develop and to port their applications.