In case I had forgotten how little mindshare Symbian has in many parts of North America, the recent Robert X. Cringely piece “Why Windows Mobile will die” contained yet another stark reminder.
As usual with Cringely, the piece mixes potential insight with a lot of conjecture and then some fancy. Most of the article discusses Windows Mobile, iPhone, and Android. But it squeezes in a dismissive paragraph about Symbian:
…donning flameproof clothing: Symbian is simply too old. The OS is getting slower and slower with each release. The GUIs are getting uglier and are not user-friendly. The development environment is particularly bad, which wouldn’t hurt if there weren’t others that are so much better. Symbian C++, for example, is not a standard C++. There is little momentum in the Symbian developer community, maybe because coding for Symbian is a pain. Yes, there are way more Symbian phones in circulation, but those phones will be gone 18 months from now, probably replaced by phones with a different OS. Lately, Symbian’s success has been primarily based on the high quality of Nokia hardware, on the loyalty of NTT DoCoMo, and now on the lure of being recently made open source and therefore free. But if open source developers don’t flock now to Symbian (they aren’t as far as I can see — at least not yet) then the OS is doomed.
And if that weren’t a sufficiently strong reminder of Symbian’s lack of mindshare, I found scant encouragement in the 65 comments posted (so far) to Cringely’s piece.
Allow me a few moments to respond to individual points in this paragraph, before I return to the bigger picture.
“Symbian is simply too old” – but it has been undergoing a constant internal renewal, with parts of the architecture and code being refactored and replaced with each new point release. Just a few examples: we introduced a new kernel in v8.1b, a new security architecture in v9.0, new database (SQL) architecture in v9.3, new Bluetooth in v9.4, substantially revised graphics architecture and networking architecture in v9.5, and so forth.
“The OS is getting slower and slower with each release” – on the contrary, many parts of the operating system are humming much quicker in the newer releases, as a result of a specific and pervasive focus on performance across the whole system. Deliverables include speed ups due to smart incorporation of demand paging, file system caching, data scalability improvements, and wider adoption of separation of activity into three planes (data plane, control plane, and management plane).
“The GUIs are getting uglier and are not user-friendly” – but the UI system is increasingly flexible, which allows customers to experiment with many different solutions (whilst retaining API compatibility). New developments such as the S60 Fifth Edition touch interface, and the recently announced support for Qt on Symbian OS, take things further in the user-friendly direction.
“The development environment is particularly bad” – but documentation and tools for Symbian OS have markedly improved over the last two years.
“Symbian C++, for example, is not a standard C++” – but watch out for our forthcoming annoucements about EUserHL that go a long way to address this particular gripe.
“There is little momentum in the Symbian developer community” – but that’s not the impression given by the media reports from people who attended the Symbian Smartphone Show last week.
“Yes, there are way more Symbian phones in circulation, but those phones will be gone 18 months from now, probably replaced by phones with a different OS” – but I beg to differ, based on my knowledge of development projects underway at phone manufacturers across the world. For just one example, consider the recent remarks from Li Jilin, Huawei Communications Vice President (note: Huawei has previously not been a user of Symbian OS):
“Huawei is excited by the plans for the Symbian Foundation. We look forward to participating in the work of the Symbian Foundation and using the foundation’s platform to deliver a portfolio of devices for mobile network operators around the world. We believe that the Symbian Foundation ecosystem will enable innovation which will benefit users and drive increased customer satisfaction.”
“If open source developers don’t flock now to Symbian (they aren’t as far as I can see — at least not yet) then the OS is doomed” – but this is far too impatient. It’s too early to make this judgement. You can’t expect the open source developers to flock to us before more plans are published for the roadmap to put our source code into open source.
As for the bigger picture: despite the above individual points of fact, I don’t expect significant changes in mindset (except among the far-sighted) until there are more Symbian devices in the hands of North Americans.
It was the amazing array of devices at the partner showcase stands at the Smartphone Show last week that caused the biggest buzz of all – bigger than the announcements from the keynote hall next door. Thankfully, AT&T have publicly mentioned their “plan to introduce more Symbian phones“. North American users shouldn’t have too long to wait. And there are encouraging signs of independently-minded North American writers actually (shock horror) liking the latest Symbian phones. For example, the renowned software essayist Joel Spolsky called the Nokia E71 “the best phone I’ve ever had – I’m loving it“.
In the meantime, the Symbian Foundation has a big mountain to climb, in public perception. But it’s a mountain well worth climbing!
dw2 
>>Symbian C++, for example, is not a standard C++. There is little momentum in the Symbian developer community, maybe because coding for Symbian is a pain<<
Absolute rubbish, maybe author does not know how to write good code.
However there is a point in >>little momentum in the Symbian developer community<<
The biggest pain is security architecture not the coding. There is a saying "too much of everything is bad" and symbian security is to much to do a fast development. No one has time and money to indulge in long procedures.
Comment by Vipin — 25 October 2008 @ 5:01 pm
>>Symbian C++, for example, is not a standard C++. There is little momentum in the Symbian developer community, maybe because coding for Symbian is a pain<<
Absolute rubbish, maybe author does not know how to write good code.
However there is a point in >>little momentum in the Symbian developer community<<
The biggest pain is security architecture not the coding. There is a saying "too much of everything is bad" and symbian security is to much to do a fast development. No one has time and money to indulge in long procedures.
Comment by Vipin — 25 October 2008 @ 5:01 pm
About the developers momentum and the open source developers’ flocking:
In the Smartphone Show, the OpenC session was full, unlike the previous sessions. Which shows that there are open source developers interested in Symbian, at least in Europe.
Without open source projects, there is no future for a modern OS, so Symbian should have a more aggresive policy, helping even more open source developers. The main problem right now as I understand it is that the Professional edition of Carbide is not free. That turns away many developers.
Comment by nop — 25 October 2008 @ 6:24 pm
About the developers momentum and the open source developers’ flocking:
In the Smartphone Show, the OpenC session was full, unlike the previous sessions. Which shows that there are open source developers interested in Symbian, at least in Europe.
Without open source projects, there is no future for a modern OS, so Symbian should have a more aggresive policy, helping even more open source developers. The main problem right now as I understand it is that the Professional edition of Carbide is not free. That turns away many developers.
Comment by nop — 25 October 2008 @ 6:24 pm
As a previous owner of several Psion devices, and currently owning an S60 phone I unfortunately agree with Cringely when he says that “lately, Symbian’s success has been primarily based on the high quality of Nokia hardware”. The standard S60 applications are very limited and third party offerings poor for the most part as is evidenced that you are on record as still carrying around a Psion 5Mx (s/w 10-15 years old and the legacy squandered).
Comment by aj365 — 25 October 2008 @ 11:32 pm
As a previous owner of several Psion devices, and currently owning an S60 phone I unfortunately agree with Cringely when he says that “lately, Symbian’s success has been primarily based on the high quality of Nokia hardware”. The standard S60 applications are very limited and third party offerings poor for the most part as is evidenced that you are on record as still carrying around a Psion 5Mx (s/w 10-15 years old and the legacy squandered).
Comment by aj365 — 25 October 2008 @ 11:32 pm
Symbian is entering a new level in the open source family and there is much excitement. Huawei interest is the beginning and this represents an explosive market.
What counts, at the end, is the symbiotic relation of the Symbian family developers and of course end user experience.
Comment by H3 — 26 October 2008 @ 7:12 pm
Symbian is entering a new level in the open source family and there is much excitement. Huawei interest is the beginning and this represents an explosive market.
What counts, at the end, is the symbiotic relation of the Symbian family developers and of course end user experience.
Comment by H3 — 26 October 2008 @ 7:12 pm
Hi Vipin,
>The biggest pain is security architecture not the coding. There is a saying "too much of everything is bad" and symbian security is too much to do a fast development.
The security system is something we’re reviewing. But we are planning to improve the coding too – via more standard APIs, better tools, and so on.
Hi nop,
>In the Smartphone Show, the OpenC session was full, unlike the previous sessions. Which shows that there are open source developers interested in Symbian, at least in Europe.
Good point!
>Without open source projects, there is no future for a modern OS, so Symbian should have a more aggresive policy, helping even more open source developers.
Agreed. Watch this space…
>The main problem right now as I understand it is that the Professional edition of Carbide is not free. That turns away many developers.
Just as we are removing cost from accessing the source code, we plan to remove cost from using key development tools.
Hi aj365,
>As a previous owner of several Psion devices, and currently owning an S60 phone I unfortunately agree with Cringely when he says that "lately, Symbian's success has been primarily based on the high quality of Nokia hardware".
I agree that the Nokia hardware (including their handling of telephone calls) is first class. But I think some of the software is compelling too…
>The standard S60 applications are very limited and third party offerings poor for the most part
Examples of great apps for S60, in my opinion, include BlackBerry Connect, Google Maps, and the Python environment. I also like Opera Mini.
>…as is evidenced that you are on record as still carrying around a Psion 5Mx (s/w 10-15 years old and the legacy squandered).
Now that’s another story – too long to fit in this reply. But see here for some hints.
Hi h3,
>Symbian is entering a new level in the open source family and there is much excitement. Huawei interest is the beginning and this represents an explosive market.
>What counts, at the end, is the symbiotic relation of the Symbian family developers and of course end user experience.
Well said!
// dw2-0
Comment by David Wood — 26 October 2008 @ 8:05 pm
Hi Vipin,
>The biggest pain is security architecture not the coding. There is a saying "too much of everything is bad" and symbian security is too much to do a fast development.
The security system is something we’re reviewing. But we are planning to improve the coding too – via more standard APIs, better tools, and so on.
Hi nop,
>In the Smartphone Show, the OpenC session was full, unlike the previous sessions. Which shows that there are open source developers interested in Symbian, at least in Europe.
Good point!
>Without open source projects, there is no future for a modern OS, so Symbian should have a more aggresive policy, helping even more open source developers.
Agreed. Watch this space…
>The main problem right now as I understand it is that the Professional edition of Carbide is not free. That turns away many developers.
Just as we are removing cost from accessing the source code, we plan to remove cost from using key development tools.
Hi aj365,
>As a previous owner of several Psion devices, and currently owning an S60 phone I unfortunately agree with Cringely when he says that "lately, Symbian's success has been primarily based on the high quality of Nokia hardware".
I agree that the Nokia hardware (including their handling of telephone calls) is first class. But I think some of the software is compelling too…
>The standard S60 applications are very limited and third party offerings poor for the most part
Examples of great apps for S60, in my opinion, include BlackBerry Connect, Google Maps, and the Python environment. I also like Opera Mini.
>…as is evidenced that you are on record as still carrying around a Psion 5Mx (s/w 10-15 years old and the legacy squandered).
Now that’s another story – too long to fit in this reply. But see here for some hints.
Hi h3,
>Symbian is entering a new level in the open source family and there is much excitement. Huawei interest is the beginning and this represents an explosive market.
>What counts, at the end, is the symbiotic relation of the Symbian family developers and of course end user experience.
Well said!
// dw2-0
Comment by David Wood — 26 October 2008 @ 8:05 pm
Thanks for the reply. I understand why you use a 5mx (“… device supplements my memory, keeps track of my appointments, gathers my thoughts and ideas, marshalls my to-do items, and much, much more”), I don’t understand why this capability has been lost from current EPOC(Symbian) devices. I would be interested in reading the “longer story”.
Comment by aj365 — 26 October 2008 @ 11:50 pm
Thanks for the reply. I understand why you use a 5mx (“… device supplements my memory, keeps track of my appointments, gathers my thoughts and ideas, marshalls my to-do items, and much, much more”), I don’t understand why this capability has been lost from current EPOC(Symbian) devices. I would be interested in reading the “longer story”.
Comment by aj365 — 26 October 2008 @ 11:50 pm
Nokia's hardware and software expertise are actually less relevant to Symbian's continued presence than Nokia's outright scale and distribution.
Apple, HTC & other Android licencees probably couldn't physically scale to match Nokia's manufacturing and sales presence in 5 years, even if the demand was there. The sole exceptions is probably Samsung, which had 6 Symbian devices at last week's show.
As well as Huawei, I'm also wondering if the open-source nature of the Symbian Foundation convinces Qualcomm to produce a hardware platform to support it, now that it doesn't have to deal with Nokia directly.
Dean
Comment by Dean Bubley — 27 October 2008 @ 10:35 am
Nokia's hardware and software expertise are actually less relevant to Symbian's continued presence than Nokia's outright scale and distribution.
Apple, HTC & other Android licencees probably couldn't physically scale to match Nokia's manufacturing and sales presence in 5 years, even if the demand was there. The sole exceptions is probably Samsung, which had 6 Symbian devices at last week's show.
As well as Huawei, I'm also wondering if the open-source nature of the Symbian Foundation convinces Qualcomm to produce a hardware platform to support it, now that it doesn't have to deal with Nokia directly.
Dean
Comment by Dean Bubley — 27 October 2008 @ 10:35 am
However much benchmarks may prove bits of the OS run faster now than they used to, the N96 is a shockingly slow ‘flagship’ device for Nokia – bringing up a contact list when writing a text takes 2+ seconds even after it has ‘warmed up’ which simply isn’t acceptable for core functionality on a device of that price. Starting some functions can leave the screen frozen until after the backlight times out, over 5 seconds. This slowness is compounded by the fact that the UI just freezes between screens (and doesn’t show context well) – the iPhone’s flashy swipes and fades are actually critical to covering up understandable lag in the OS and giving the user the impression that something is happening responsively. I remember being taught that in UI lectures 10 years ago, it should be built into any modern UI layer.
The E71 proves that the latest S60 can run fast, so why do Nokia consistently underpower the hardware? The N96 is proof that end users can very justifiably claim that S60 is slow.
Comment by raddedas — 27 October 2008 @ 3:25 pm
However much benchmarks may prove bits of the OS run faster now than they used to, the N96 is a shockingly slow ‘flagship’ device for Nokia – bringing up a contact list when writing a text takes 2+ seconds even after it has ‘warmed up’ which simply isn’t acceptable for core functionality on a device of that price. Starting some functions can leave the screen frozen until after the backlight times out, over 5 seconds. This slowness is compounded by the fact that the UI just freezes between screens (and doesn’t show context well) – the iPhone’s flashy swipes and fades are actually critical to covering up understandable lag in the OS and giving the user the impression that something is happening responsively. I remember being taught that in UI lectures 10 years ago, it should be built into any modern UI layer.
The E71 proves that the latest S60 can run fast, so why do Nokia consistently underpower the hardware? The N96 is proof that end users can very justifiably claim that S60 is slow.
Comment by raddedas — 27 October 2008 @ 3:25 pm
Hmmm. Anyone who has ever had to code for WinMo *AND* Symbian would have to say that Symbian is much more consistent. The big issue (as always) is 2-stage construction, TRAP/LEAVE and the cleanup stack and descriptors. These things are just foreign to most C++ developers, so they are deemed bad.
I’d actually prefer to see this stuff go away myself – or at least see it handled by the compiler. Descriptors could be fixed by just allowing them to be cast easier. Maybe even just wrap descriptors with a CString class. Also, wouldn’t it be sweet to have Carbide.c++ flag API calls that require specific security capability?
But the jewel of the Symbian platform has got to be its client/server architecture. That the most brilliant design decision which never gets the attention it deserves.
Comment by Darin — 31 October 2008 @ 1:43 am
Hmmm. Anyone who has ever had to code for WinMo *AND* Symbian would have to say that Symbian is much more consistent. The big issue (as always) is 2-stage construction, TRAP/LEAVE and the cleanup stack and descriptors. These things are just foreign to most C++ developers, so they are deemed bad.
I’d actually prefer to see this stuff go away myself – or at least see it handled by the compiler. Descriptors could be fixed by just allowing them to be cast easier. Maybe even just wrap descriptors with a CString class. Also, wouldn’t it be sweet to have Carbide.c++ flag API calls that require specific security capability?
But the jewel of the Symbian platform has got to be its client/server architecture. That the most brilliant design decision which never gets the attention it deserves.
Comment by Darin — 31 October 2008 @ 1:43 am