i can see linux overtaking symbian maybe but windows very unlikely. the main reason is that symbian and linux will always run on cheaper hardware. windows is resource addicted. windows phones will always be more expensive and i can't see them achieving critical mass. even with contracts the equivalent symbian will be a lot cheaper and people vote with their wallets.

linux will run on the same hardware as symbian but getting a common codebase will be difficult if not impossible as none of the companies making phones are going to share their code with their competitors. linux strenght comes from the open source development model.

moylan - Yes I agree. Linux is more of a threat. I do feel it is more theoretical at this point in some ways. I also think its going to struggle to keep up technological and feature wise wih Symbian in the high end devcies. No question it will be major competition in mid tier. The lowering of Symbian License fees is, I feel, more a defence against Linux than anything else.

Very well written article, Rafe. Congratulations.

I am even more pessimistic about Linux than you. It makes for a decent server OS, because servers are managed by experienced people. It also works well as an alternative to traditional embedded systems, where a single vendor creates a mostly closed system (like TiVo). But the reactive nature of OSS development, combined with lack of strong leadership resulting in constant fragmentation, is a killer for Linux's hopes as a broad platform, whether on the desktop or mobile devices.

MS's software has technical problems, but we've seen in the past that MS will work on that until it's "good enough" for the average person. However, the phone industry is very different from the computer industry, so I'm not sure MS will work well here, either.

So long as Symbian doesn't create another break like the v9 one (and this is the right time to do it), they should be quite successful.

"now is the right time to do it" - just interested in you expanding a little on this?

You mean now is right because the competition's not established yet?

Steve

IMHO, Windows is a high-end threat because it appeals to MIS departments to buy "Windows Everything" and they will feel that it is convenient to integrate Windows mobiles with all their existing Windows infrastructure. When it comes to managing the software installed on phones in the same sense as managing the software on employees' PCs and laptops, the MIS departments will like whichever platform has the best and most integrated tools.

Linux seems to be the lower-end threat because it doesn't have the advantage of "being Windows" so it will have to work on price alone.

Symbian could end up being between a rock and a hard place except for the fact that it is owned by Nokia. This guarantees it a large market share (for now) and plenty of investment.

I don't think that Linux investment is very large by comparison so it probably won't be able to develop and grow to support new phone designs quite as quickly.

Both Linux and Windows have the advantage of a huge variety and high qualty set of development tools which have been refined over a long period. Symbian's resources don't seem to be great enough to keep up with this yet. I am thinking in terms of modern compilers, memory debugging tools and incredibly useful tools like strace and ethereal.

There is also a very rich set of libraries that is available for Windows and Linux and the effort of maintaing ports of those libraries is likely to be much less than the effort of maintaining a port to Symbian.

Symbian tends to force one to re-design applications to fit it's model and that is going to be very bad for enterprises - it simply doesn't allow one to amortize the costs of developing software. Some might argue that phones don't run the same software as other systems but such an attitude is lacking in any vision and there are already examples that contradict it: eg. Opera, Helixplayer, KHTML and Zlib. An enterprise that knows it can recast it's internal systems to work on a .NET phone without completely rewriting them will be happier to invest in Windows than Symbian (ok - hurray for Appforge etc . .). Java was supposed to supply this for Symbian but it seems to be mainly useful for games at the moment. Not supporting a full-native API interface made Java a toy (IMHO again).

From this soup of thoughts I have a few conclusions:

1) Symbian is poorly sold in the high-end market - has no recognition. This will make it's acceptance in the enterprise market unlikely. This market isn't the main prize because it is of limited size but it is rich and a way to influence people and gain recognition.

2) Symbian will sell more than Windows anyhow and that creates an opportunity - anyone who can come up with a "killer" application will start to see a market that is big enough to be worth addressing. If Windows or RIM acquires a killer feature then Symbian's numerical advantage will allow time for that feature to be added to it.

3) The "flood" of software will not "happen" until the cost of GPRS and/or 3G downloads comes down very substantially (at least in the UK). The fact that bandwidth is expensive or of variable cost inhibits many innovations and operators need some sort of competition that will force them to reduce their thirst for ever higher ARPU (Average Revenue Per User) and alter the market in favor of the consumer. If the operators don't become more competitive, then the software market for handsets won't have any chance to develop at all.

4) Features are more useful when they are available everywhere and this means that handset manufacturers really need to make their phones more similar and less differentiated - a very unlikely thing for them to do unless the balance of power can be shifted away from the makers of equipment to the operating system provider and software vendors. MS has this power arrangement and it's a potent weapon. I think that the use of reference designs will probably help a lot. Linux is at entirely the opposite situation - the manufacturers have all the power.

5) It will be necessary to add a simple update feature for the software and firmware on a handset and this will have to have a good user interface and be very reliable and cost very little. This is supposed to be on it's way (OMA DM) but we'll see. This is needed so that older handsets can be brought up to some common standard with newer ones. Application software shouldn't have to be written to take into acount all the bugs and eccentricities of existing models in order for it to work on a wide variety of handsets.

Cheers,

Tim

Tim - very interesting post.

I agree with you on development tools. I think Nokia and Symbian are waking up to this and the new Carbide tool set does seem quite impressive. However there is room for a lot more here.

The big problem I have with Linux is that its not really a platform. That's a big advantage for Symbian and Microsoft's OS'.

The Nokia investment is both a blessing and a curse. It may put some companies off, but on the other hand is guarantees future stability. Interestingly Sony Ericsson look like they are strongly committing to Symbian.

The firmware over the air issue - spot on here I feel - a big complaint I hear from ddevelopers in having to code round annoyances in individual firmware (never mind individual phones). I know this is in S60 3rd Edition (possibly in FP1). We'll have to see how it goes. This is cearly a place where Symbian / s60 / uiq has the potential to get ahead.

Data costs are restructing OTA service growth though I think this is a matter of time.

I've had an email exchange with TDG and it's only fair to point out a few things.

The reports says 26.8 million Symbian devices shipped in 2005. I still think this is wrong (it's out by about 7 million), but it does mean the number of Windows Mobile and Linux devices are 8.45 million and 11.8 million respectively (which are perhaps more realistic numbers). The Windows Mobile numbers include all PPCPE (pdaphones), Windows Mobile 5 and Windows Smartphone devices, which I think inflates in market share for smatphones.

The report conclusion is based on two main points. Firstly that China will have a big role to play (and will primairly adopt Linux). I agree China is going to be important, but I don't think we can be certain what will happen.

Secondly the report argues that Symbian will decline because it wont be able to keep up with market requirements in a way that will make a more compelling business model. I do not agree with this argument (technically I think Symbian is ahead - look at the low level features such as the new kernel, and the report under estimates) . However it is an argument that is possible to make. It is true that the winner will not necessaiirly be the best technically but the on that makes the most sense from the business model standpoint. It is quite possible to argue that Symbian with its reduced license fees is the best.

TDG deserve credit for taking the time to contact me (and answer some of the points I raised).

At the end of the day all such market research is forward looking, and no one can know the future. Mobile Opportunity has an interesting take on this. See also this report that sees Symbian with 60% market sahre in 2009.

What I'd like to see is a market research company do market research on the accuracy of market research companies! I suspect they wont come out any better than sticking a pin in a bit of paper.

Anyway. Yes, China do like Linux and Open Source but Linux is still a very long way behind Symbian and Microsoft in terms of being a smartphone OS. Symbian is more open than Microsoft for example, you could devlop your own UI if you want to. Also including Symbian in your device will not be very expensive in the future.

The enterprise is a very new market. I you ask IT managers today which mobile solution they would ge for they will probably say Microsoft because thats the safe thing to say. However I doubt if many of them have seriously researched it! Symbian with its latest security model and more open nature e.g. supporting multiple push email solutions starts to make it a more attractive to business and possibly safer in terms of corporate security.

I'm not even sure if the enterprise market is ever really going to be a big factor. In a short while most of your employees will have phones capable of doing most of the corporate needs i.e. email and web access, etc. Why pay out for more devices? Why not just use whatever device you employee has but just make sure that when the connect to your companies servers they can only do it in a safe and secure way.