I caught up with Mr. Carney last week and quizzed him on mobile development issues, as seen from the Symbian perspective.
NN: What initially drew you to small systems development?
BC: My early career was embedded systems programming --flight simulators and some pretty advanced stuff using ADSL to do video on demand on an C++ embedded OS in 1993. However, that is not why I am working for Symbian. I have been passionately interested in the internet since 1987.
NN: Interest in the internet brought you to Symbian? Explain a little about the relationship, as you see it.
B.C.: Looking back, in 1990, few computers were networked. Then within 10 years every computer was connected! By 2000, it was apparent to me that the internet would lose the wires and wireless, mobile networking would replicate what had happened in the fixed line internet.
I predicted to myself (and anyone who would listen) that by 2010 the internet would be wireless (i.e. the mobile internet would be real) so I took a career diversion. During 2007 we have seen that many of the big technology announcements are relating to smartphone technology. Comparing what happened in 1998-2000 on the fixed internet, I truly believe 2008-2010 is potentially going to be a massive period of innovation in mobile.
NN: Could you give a brief history of Symbian OS? For example, where are its technical roots? What brought the partners/consortium together? How did Symbian recruit such a broad audience?
B.C.: Symbian is neither a partnership nor a consortium. It is a separate company, born in 1998, whose sole mission is to create and market an open standard OS for mobile devices. The idea that we are a consortium probably springs from the fact we were founded with investment from major mobile device manufacturers Nokia, Motorola and Ericsson. Today’s Symbian OS evolved from an earlier mobile OS, EPOC, created by a company called Psion.
The Symbian founders saw a need for an open OS, independent of specific device vendors. Openness is a cornerstone of how we do business, and we work closely with our customers to evolve our standard in ways that support the advancing capabilities of their devices. Because of this, we have broad knowledge about emerging handset design, and this informs our direction for the Symbian OS. Everyone’s requirements get included. This is all possible because our customers don’t fear us! We don’t compete in their part of the value chain. We make a mobile device operating system. Period.
NN: What, from your point of view, are the strongest technical features of Symbian OS? Of the Symbian C++ language?
B.C: Symbian OS is designed as an object oriented OS from the ground up. This gives us unparalleled extensibility, and device manufacturers love this. It allows them to use one OS and extend, segment and differentiate their devices without significant increase in manufacturing costs. Having an open standard OS also "future proofs" them to a certain extent, because they can evolve existing device designs gracefully.
The Symbian C++ language is based on what we know to be required in the mobile space: "five nines" reliability, good battery life, performance, and minimized Bill of Material cost. These considerations are industry drivers, and we are very sensitive to all of them. Squeezing the max out of a device is what the mobile handset business is about. The Symbian C++ language and SDKs are devised to support great device variety-- differing form factors, camera types, mobility options--but with great efficiency. We expose more APIs than any competing solution, a really huge amount of functionality.
All of this exposed functionality, however, is protected by an overarching platform security, which is absolutely essential on a handset. You don’t want your phone to reboot when you need to call 911, after all. In short, we make sure that the right functions are exposed to the right people developing the right stuff, which is the elegance of our approach.
NN: What are your favorite tools and development environments and why?
Nokia’s Carbide is our current favorite. It’s an Eclipse descendant and so is stable, familiar to a lot of our developers and in its base version, free.
Nokia Tools Download Link
NN: Could you characterize a typical Symbian developer in terms of locale, experience and application niche?
B.C: In terms of the number of apps we see coming through the Symbian Signed process, a large share come from the US and Germany. We also see a good deal of activity from China, but when you normalize for population, it’s German and US development that stand out as generating the highest volume of software. We know there are as many as 2.5 million developers writing applications that can run on Symbian OS based smartphones. When we segment our developer population, we divide apps among Mobile Services, Games and Utilities. We see more or less level interest across these categories.
Of course, today we are talking about development using Carbide and Symbian C++, but it’s important from our perspective to reiterate that more "user friendly" development tools like Adobe’s Flash generate mobile apps and content that run on the Symbian OS platform. When we think about developers, we think of the ones using tools above us in the stack, as well as those using Symbian C++ to take advantage of the power and advanced features of the Symbian SDKs.
NN:Talk a bit about the signing process. Why is it necessary? Why give operators so much power over the channel? Who benefits?
B.C: ( Big Sigh...) I recognize that the signing process generates a lot of controversy. Let me say this: The point of signing is to keep devices open, not to close them. The Symbian Signed process gives applications an unimpeachable seal of authenticity. It is a way for carriers and consumers to know it is safe to buy and use mobile apps and content. This benefits developers as much as it does carriers and consumers. Users simply won’t tolerate phone apps that make their devices unavailable. Symbian Signed status is accepted or required by more than 250 carriers worldwide, which is about the best possible way to open doors to markets. It’s a "Test once, run anywhere" strategy. Everyone benefits.
NN: Is this an enterprise solution developer’s place to be, or is it a consumer driven space?
B.C: At this point, consumers are driving the mobile device market. But whether at work or not, we are all consumers. We tend to choose the best experiences, when offered a choice. This has implications for enterprise mobile applications. For many kinds of business interactions, mobile devices are simply going to be more convenient and productive.
At bottom, the question for enterprise mobile apps may be this: "At what point don’t you need a laptop anymore?" High end mobile devices literally offer everything in terms of performance and functionality as premium notebook computers, except they don’t have full size screens and keyboards. One Nokia model has 8 GB of storage and uses demand paging for memory management. This is really sophisticated stuff--not so long ago, mainframe computing technology. It’s just remarkably fast. Also, light, easy to carry, and multi functional.
NN: Do you see significant growth ahead for the Symbian developer community? Why or why not?
B.C: Well, sure. It’s a pretty big opportunity, and this isn’t going to be lost on the software entrepreneur. But remember, at Symbian, when we think of our developer community, we think of development environments above us in the stack as well as our native app developers. Symbian phones currently host Adobe FlashLite content and the S60 browser, which means that many iPhone widgets just run without changes. There is a port of .NET for Symbian and an open source port of Python. There is ongoing effort to port GNU type tools and we have a new POSIX compliant API set. It’s a great time to be a Symbian developer.
QUOTE: "B.C: ( Big Sigh...) I recognize that the signing process generates a lot of controversy. Let me say this: The point of signing is to keep devices open, not to close them. The Symbian Signed process gives applications an unimpeachable seal of authenticity. It is a way for carriers and consumers to know it is safe to buy and use mobile apps and content. This benefits developers as much as it does carriers and consumers. Users simply won’t tolerate phone apps that make their devices unavailable. Symbian Signed status is accepted or required by more than 250 carriers worldwide, which is about the best possible way to open doors to markets."
I am highly sceptical about this point of view. Saying that "Symbian Signed status is accepted or required by more than 250 carriers" shows who really wants signing: the carriers. Not the developers. Every single Symbian developer I met was frustrated by signing process. And how telling is it that "users won’t tolerate phone apps that make their devices unavailable"? Users won't tolerate apps that make their PC unavailable either, but that doesn't mean they demand their PC apps to be signed. The PC is coming along pretty decent without.
In combination with the cumbersome Symbian development process (to say the least), I feel that the signing requirement is the final nail in Symbian's coffin. With alternatives (Windows Mobile, iPhone/browser-based and who knows Android) getting better and better Symbian's chances are diminishing. I'd love to be proven wrong, but I currently believe that the signing process hoovers above Symbian like the sword above Damocles.
I find it fascinating how the current blogosphere assumption is that smartphone/mobile platforms are a miniature version of PCs; i.e. PC paradigm = Mobile/Smartphone paradigm.
If I could pass on the mind shift that I have realized in the last 5 years, smarpthones are not computer/developer led, it is a CONSUMER device. Just as consumers buy a gaming console ('cause they plug it in, buy the games and it just works) they expect much more from their mobile device and the apps they install on it.
To give you some idea of what global/consumer/mass market means; in Nancy's earlier post people many people commented that in actual fact Symbian only has single digit percentage of the entire mobile phone market, but turning that into numbers, 20+ MILLION phones shipped with Symbian OS in Q3 (is a small percentage?)...but....20M in a quarter (and 165+ million cumulative) is more shipments than all the other Smartphone platforms combined, ultimately Symbian has solved problems that most of others platforms haven't even reached yet!
When you think global mobile mass market it is beyond the scale of almost any other technology. It is big and then some!
Hence Openness vs Security vs Reliability vs Trust in the mobile mass market is a non trivial problem. Needs a bit more than markitectures and fast talkers; we've had some of the best brains in the industry working on it for 10 years.
The other urban myth is that Symbian do Symbian Signed because carriers wanted it? Symbian do Symbian Signed, because if Symbian didn't; carriers would have done it anyway. The current implementation is win/win/win; Carriers don't have to do it, developers only have to do it once (not on 250+ carriers for every global release = fragmentation) and consumers get a better and more trusted applications experience. No conspiracy theory there, just common sense?
We rolled out some major changes to Symbian Signed last month to make it simpler, cheaper and faster for developers. If you haven't already - Check it out www.symbiansigned.com
As an outsider to mobile development, I would be interested in hearing more about the process of developing applications and bringing them to market. Although it might not be like the desktop computer market, the biggest issue I've heard about is that the carriers control everything. In that case, it's probably better that Symbian does the signing to at least centralize the process. But the bottom line is that if you can't get a carrier to provide your application, you're screwed.
I would think that a better alternative is some kind of test suite to verify that your application isn't going to screw up a phone. If your app passes the suite, then users can buy it and install it, without the control of, and potential blockage by for any number of nefarious reasons, the carriers.
In the current world this sounds ridiculous, and "it just can't work that way, this isn't the PC." But that's the same attitude that Microsoft had about open source software and competing operating systems, and those things have continually nibbled and grown larger, and there's no indication that they are going away. The cell phone market is a much larger target and if the carriers don't play fair with developers, then at some point someone will begin coming up with a better, more open system (note that the reason AutoCad became so big was precisely because it allowed third party addons, without those addons being controlled or taxed by AutoCad. But yeah, that's PC thinking, I suppose).
I don't know yet, but I wonder if Android might be such a system. Maybe I can get Dick Wall (Android Evangelist) to comment about it. Or maybe Nancy will interview him.
On the whole, I’d say that I’m in your camp on the signing issue, but for different reasons than yourself. Clearly, regardless of Symbian’s motivation, required application signing allows carriers to constrict the channel in ways and to a degree that very nearly constitute a restraint of trade. I am somewhat amazed that in the litigious U.S. they get away with this. However, what’s significant to me about the situation is that there are 250 distinct carriers, each operating their own fiefdom. It seems a bit feudal, and suggests to me that despite amazingly rapid and broad adoption of cell phone technology by consumers, we are actually still in the prehistory of connected lifestyle infrastructure
I fully acknowledge that the following is an excursion to the lunatic fringe, and wish to make clear at the outset that these ideas are simply being tossed out in the interest of exploratory conversation. So.
1)I see a parallel with the early days of other sorts of utility infrastructure. Take as an example, electrical transmission. Around the turn of the last century in North America, there were dozens if not hundreds of small, local electric utilities. Some used DC; some used AC; grounding and fusing were often poorly applied, occasionally with disastrous and tragic results. But people were hungry for this new technology, which is something you can readily appreciate if you’ve ever lived in a place where you had to use a match to turn on the lights. No one remembers this now that we have regional electrical transmission inter ties and plugs work pretty much the same everywhere ( though you need the converter gizmo if you travel between Europe and North America. ) My hypothesis: we won’t always have 250 carriers and they certainly won’t continue to be able to impound their customers by means of technology.
2) I don’t think that either the iPhone or the Android is going to make that much difference in terms of developer access to phones, because the obstacle isn’t really one of platform type; It’s cracking the carrier’s lock on the market. I think ( deep breath ) that what will be the key for us is momentum toward overhaul of the internet protocol. If it delivers a method of assigning identity, priority and valuation to discreet packet streams, digital communications will come busting out of carrier jail.
For now I see Symbian as the place to be simply because of market share, and signing as the price of admittance at an early stage in the game.
From a global perspective, I don't understand all the concern about carriers and anti-competitive behaviour, my conclusion is that there are North American local circumstances causing a distorted world view of mobile development.
On Symbian OS, you can pretty much go to any country in the world and your phone/network coverage will work as will Symbian Signed application on that phone.
To the extent that you can take your phone that you bought in SPAIN, with your AUSTRALIAN carrier's SIM card to CHINA and install an application developed in CHILE that you downloaded from www.handango.com in the US....This *is* the reality today on Symbian OS handsets.