About a third of the Android code base is open and nothing more,” says Williams. “And what is open is a collection of middleware. Everything else is closed or proprietary.”
Can anyone with more Android experience comment on this? What is the 2/3 that has been kept proprietary?
Symbian isn't open sourcing any of those things either, AFAICT. Symbian will still have many closed source components similar to Android. Symbian also has components that are not open source, but whose Android equivalents are--e.g. its fonts and the font rendering code. In fact, the Symbian fonts have a very restrictive license--no redistribution allowed, usage for R&D purposes only.
Further, during the open-sourcing process, Symbian created new ("open source special") implementations of some components; that is, the open-source implementations of some components are not the same as the implementations that real devices have been using. It will be a long time, if ever, before these open-source versions catch up to the production-quality versions.
The Android license (BSD) is a lot easier/safer to comply with than Symbian's license (EPL). And, if you actually want to read the source code, Android's code is a lot easier to jump into than Symbian's is. (I'm not saying Android's code is better or that Symbian's code is bad; I'm just saying that there are many layers of indirection and unusual, Symbian-specific idioms in Symbian's code that makes it difficult to read.)
Anyway, I don't understand why Symbian is so eager to start the "who's more open source" debate. On many points, Android clearly wins on open-sourciness. I'm sure Symbian wins on other open-sourciness points. Open-sourcing was a smart thing for Symbian to do, whether Android exists or not, and I think Symbian should emphasize the positive aspects of this change.
That said, it would be very cool if somebody created a table that detailed which parts of Android and Symbian are open source, component-by-component, side-by-side.
This seems completely backwards to me. The kernel of course is open, with the possible exception of some specific hardware drivers for specific phones (Wifi and perhaps GPS). The virtual machine (dalvik) and development environment are both open. The closed parts are at the apps level: Gmail, Gmaps, Marketplace, and so on.
Nonsense. It might be 'too little, too late' to take on the iPhone, but it's timed perfectly for the real market for mobile phones. It's easy to forget that there are an estimated four billion GSM subscribers worldwide, most of whom neither know nor care about Android or the iPhone.
The global mobile OS market is currently dominated by Nucleus, OSE and Series 40 - these three platforms have a combined market share of 90%.
For Series 40 users, the obvious upgrade path is Series 60 - of course Nokia will lose many of these customers, but they will keep many more.
That leaves us with the myriad manufacturers using Nucleus and OSE. These manufacturers, many of them completely unknown in the west, are producing featurephones for customers who are highly price sensitive and are staggeringly fickle and fashion-led.
When looking for an upgrade path to a full smartphone OS, the main concern of these manufacturers will be cost - of implementation and of Bill of Materials. Symbian has clear advantages over other smartphone platforms in this respect, being royalty-free and having significantly lower hardware requirements than Android. An open code base will open the door to these Chinese manufacturers and their dizzying diversity of hardware and software designs.
I reckon that the Symbian Foundation has played an absolute blinder here - it may have lost North America, but it still has a dominant position in EMEA and has set itself up perfectly to launch an aggressive assault on BRIC.
What you say sounds reasonable on the face of it, but there's one piece missing. Do these manufacturers believe that Symbian can keep up with the pace of new hardware designs and software fashions or that it even survives?
Now, I'm anything but an expert on Symbian, but my general impression was that Symbian has been falling behind and Nokia has been suffering financially partly because of that.
If I was a Chinese handset manufacturer I would ask myself who I can trust more to stick to its platform and subsidise further development, Google or Nokia (or even Microsoft if they manage to pull off a comeback).
Nokia has started to make high-end Linux based phones as well. Why would they do that if they believe in Symbian? And why would Nokia, a handset manufacturer itself, subsidise an OS that other handset manufacturers are successfully using?
There are too many unanswered questions around Symbian I think.
Nokia has started to make high-end Linux based phones as well. Why would they do that if they believe in Symbian?
The explanation is right there in your first sentence -- "high-end". Symbian can work with 1/10th of the CPU and 1/20th of the RAM compared to Nokia's Maemo Linux platform.
It will be a long time before all phones can be equipped with ARM Cortex CPUs and 256 MB RAM. In the meantime, there will be literally billions of phones sold with much lesser specs. It makes a lot of sense for Symbian to target that market.
(It's curious that Nokia gets so much flack for using multiple operating systems, when other vendors don't have a single OS that would span their whole range either... The existence of Mac OS X doesn't imply that iPhone OS is too limited. Windows Mobile isn't suffering because Windows 7 is so much better. Google has both Android and Chrome OS, etc. Different devices, different users, different operating systems.)
Symbian phones are fairly capable (just look at what many of the high end nokias have been able to do for years); just generally not as "polished" as an iPhone/Android device.
> There are too many unanswered questions around Symbian I think.
Maybe we'll get some answers now that the source is out there. I think that part of the problem is that Nokia is just confused, and spread a bit thin, in terms of handling everything from 25 euro phones to the high end Linux ones.
I agree, it reminds me of Netscape's decision to spin off their code base into open-source Mozilla. Arguably the web benefited greatly from this decision, but at the time it produced a four year delay in a new version at a time Microsoft was rapidly releasing new versions of IE:
http://www.jwz.org/gruntle/nomo.html
To gain market share, maybe; but there are an insane number of devices running Symbian. This might allow us to do some interesting tinkering.
Speaking of Symbian - I remember an entire floor at Fujitsu's building in Musashi Kosugi in Kawasaki was completely locked down; no in or out without special clearance. That entire section was devoted to coding in Symbian, and they had some proprietary information about the OS from Nokia about the system.
Massive install base, yet no hype. There's more hype for Nokia's Linux phone than for any random Symbian smartphone. I coded for Symbian many years ago -- it's a dead end. The OS literally has no future. Open sourcing it is just part of the death spiral.
At least this might give some of the dedicated developers, whomever are left, a chance to actually fix some of the more annoying bugs.
I have the N900. It replaced my E71. The Skype integration is amazing. :-)
I agree that Symbian is a dead end. I have two books on Symbian development, and Symbian's bizarre form of C++ and their relatively complex code signing system drove me away from the platform.
With the N900 as a vanguard, I'm excited to see just how far Nokia can take its open, Linux-based smartphones.
Pretty low unless someone is willing to put a lot of money into it. The problem with Symbian is gazillion bizarre corner cases and exceptional situations -- this makes it hard to develop for, and also makes making a proper emulator very work-intensive.
The article says that Symbian powers most of Nokia's phones, but that's not correct. AFAIK, Symbian is on about 20% of Nokia's offerings, with the rest running their closed dumbphone operating systems (Series 30/40).
Uh, I meant; Nokia currently develops Qt; Qt as it currently stands is developed by Nokia. Qt was developed by developed by Trolltech. Nokia didn't originally develop Qt. I didn't say "was developed by"... Buffalo buffalo Buffalo buffalo buffalo Buffalo buffalo ... English ... has a lot tense strangeness...
to dear down-voters: what is the reason to develop on really crappy platform for outdated devices with 128x128 screens which people throw away and buy new once a year?
About a third of the Android code base is open and nothing more,” says Williams. “And what is open is a collection of middleware. Everything else is closed or proprietary.”
Can anyone with more Android experience comment on this? What is the 2/3 that has been kept proprietary?