Original URL: http://www.theregister.co.uk/2004/07/21/colly_myers_interview/

Symbian founder on mobile past, present and future

Why Skype's a chimera and why the iPod is great

By Andrew Orlowski

Posted in Mobile, 21st July 2004 17:28 GMT

Exclusive Colly Myers has had a pretty low profile since leaving Symbian two years ago.

For the former Managing Director of Psion, the Symbian OS is very much his baby. Before Psion, he was a mainframe programmer, and always wanted to see a mainframe-class operating system on a handheld. So he instigated the project, ten years ago this November, and led the kernel team. In 1998 Psion nudged Nokia, Ericsson and Motorola into investing in the OS to create a standalone company.

Now he's back with the phone answers service AQA (text 63336 and for a quid, you'll get a researched answer back by SMS in a few minutes). So we were pretty curious about how he took stock of the current business, what lies ahead and what he's learned. Some of the answers may surprise you.

Software is a service

The most surprising revelation came when we asked how many Symbian applications are running on his phone (a Nokia 6600).

"Not one," he says. He simply doesn't see a mass market for software. Instead, he thinks, most people will want "a lot of things you can get on the Internet on your phone translated as a service, piece by piece".

"For raw OS software there isn't a market - it will become a Java market; one where you can download and run applications everywhere." The data downloads market for ringtones is worth billions, but that's a service, he points out.

We rifled through our back interviews and pulled one out from February 2000. Even then Myers when asked if the killer app for phone really might be something as simple as talking to people, "I really don't know [and] I don't think anyone knows".

At the time he forecast 100 million wireless information devices by 2003-4, in line with analysts and as we all know that's about 80 million short of predictions. That said, he's bullish about the prospects for Symbian and for the mobile industry as a whole, once it realizes how much transactions are worth, and developers realize that software is a service.

Myers characterized the much-hyped VoIP operators like Skype as a "chimera" and predicted the winners and losers.

But the other surprising answer is that he doesn't have much faith in the success of feature-packed phones that do a little bit of everything, not very well.

"I used to think you could convert a lot of things [to an all-in-one smartphone] but I'm older and wiser, I think," he told us. Instead he sees a bright future for best of breed devices such as the iPod.

"You end up with a 'spork' - a combination of a spoon and a fork. It's no good as a spoon and no good as a fork."

(This interview was conducted before the pre-emption process and new financing arrangement for Symbian were announced. One observation of Myers proved to be prescient. "Nokia haven't proven themselves to be bad," he said. "So you might as well go along with it until they start behaving badly." That seems to be the prevailing mood, as the refinancing boosted the head count and saw increased commitment from the other partners.)

Now read on...

So where's the mass market for software developers?

There were always two sides to this: Palm with their talk of a 'Palm economy' which is a lot of claptrap, and the PC side and the consumer side on the other. For raw OS software there isn't a consumer market. It will become a Java market: one where you can download and run applications everywhere.

The data download market is for songs and ring tones and bits of software that are part of a service. But it's service orientated, not application oriented.

There are no signature apps that aren't already in the phone. Everything is built into your phone already. I always thought that was a trivial argument. But the real market for Symbian developers is as providers of software to the phones; the telephony apps, T9, and so on.

I'm really surprised you don't have any Symbian apps on your Symbian phone. What advice would you have to, say, Lee Epting, who's got the job of evangelizing Nokia developers?

My point is that there is not a mass consumer market for C++ applications, with the emphasis on consumer and C++. My theory is that any really successful C++ application will become a signature application and will end up being built into the phone. Opera is a perfect example.

So there will be a large market for C++ applications but the market will be to ODMs [Original Device Manufacturers] and handset manufacturers, and possibly, in time, network operators. There will be a large consumer market for downloadable applications but it will be for smaller and lightweight applications and they will be in Java. They must be cost effective to download and they will have increasingly shorter shelf lives. It will be Java because a large target market must exist, and for mobile phones this will only be for Java MIDP phones. Lee still needs to do whatever she's doing because the market needs the powerful C++ signature applications, but she can spend less time trying to target the applications at a consumer market.

The best thing she can do is to try and promote a cross network operator platform that allows developers of Java programs to offer their software across all networks so as to increase the size of the market. This platform must allow the phone easy access to the apps available for download and must bill the customer. The platform must allow any developer to make their program available on the platform, set its price and receive an agreed share of the revenue billed by the network operator. This will create the right whole product offering to allow the market for downloadable software to let the market take off.

So is there a market for this abstract thing, mobile 'data' ?

The appetite for data cannot be underestimated - people are just looking at the wrong type. It's just not going to be download or video data - it's going to be transaction data. They're going to need every bit of data they can get their hands on - voice traffic and transaction traffic. It's hard to see any other means by which you can run transactions.

OFDM and other new technologies will just be subsumed into the network, just like ADSL was subsumed. In two or three years data will be economic and by 2009 it will be in the midrange.

It's down that alignment of the value chain we always talked about. Everything has to be in place.

When you've got real end-to-end computing, you need to know the IP address of each client, which is why IP6, Mobile IP are so important. You've got to have secure transaction handling too. 3G does everything here - you can back-add it all to 2.5G, but it's all part of the system in 3GPP.

We also need the final bit from the phone to the merchant. We already have ticketing or tills that take Bluetooth, but the industry will need to campaign to get people happy putting it in their phones; and the mechanics of this is happening.

Until then people will avoid it. You already don't carry money or a ticket with you. I have a wireless card for my car and I don't carry keys: if only I could get rid of my wallet! Then my passport. This has a very high consumer appeal, but you've got to make it easy to access.

Everyone can count - it's easier than texting.

Won't there by some resistance to putting all this in a phone when your phone can be nicked so easily?

But it's not your whole bank account you're carrying around. You just put in £10 a day. Risk is not a factor - you'll download money into the phone over the phone. It'll be like a float. People will be carrying less money with them than they do now. I think there'll be no problem overcoming that prejudice.

Not even something like WorldMate, which gives me the weather and the exchanges rates?

That's a service.

Issuebits was set up to publish Java software, but there's no mass market for it. It may be 2007 or 2008 before Java software is sold on mobile phones, and you're never going to make a big living out of it.

Symbian is doing its job. The products that are being built are reliable and well positioned for next ranges of phones. It's way beyond the engineering of most PCs and routers: everything you've ever seen is crammed into one device, and it's got to be simple. Symbian has persuaded DoCoMo to buy in, it's got Nokia Series 60 and UIQ which are successful, and the licensees are very happy.

We moved away from doing user interfaces at Symbian, because it wasn't a product company. It's a technology company. It provides a service and really anchors the telephony handset market. It ensures that there are standards that are built up.

How important is timing? In 1999 or 2000, we seemed to think that Symbian had a huge competitive advantage by being several months' lead over the rivals…

The systems integration problem is so huge that dealing with it takes a surprisingly long time, even if you have all the experience in the world. We're coming up to November 2004, and that's the 10 year anniversary of when I began to write the [Symbian OS] code. When I started, I said it wouldn't get into a stable system on a cost basis until 2001, and that included the phone segment.

We were a couple of years too early, even then. Even now it's showing up for mid-range phones and is now starting to deliver sensible quantities: Nokia shipped 10 million Symbian phones, and that's a lot. Our good friends at Palm used to get excited when they shipped two million.

There's so much more to a phone that the technology.

What really happened in that period was the unbelievable uncertainty in the value chain. With location based services, things are coming to an end; WAP is showing a bit of life; Java is coming to phones, but it's still nothing to write home about.

WAP, the paradigm of browsing on a phone wasn't one that was going to go anywhere, and of course they oversold the system dramatically. You could have a 640 by 480 screen and have the same content available even if you messed about with it but it's still not the right metaphor for a phone. You want a lot of things you can get on the Internet on your phone translated as a service, piece by piece.

You know, I used to think you could convert a lot of things to work well on the phone, but I'm older and wiser, I think. For example, making a phone a browser and an mp3 player. Each of those things are a lesser thing and you end up with what we call a 'spork' - you end up with a spoon and a fork. It's no good as a spoon and no good as a fork, but it's got both things.

The reality is that trying to push everything into everything just doesn't make sense. We'll see an unfolding of more things like the iPod - focused at a particular consumer solution. Everything doesn't go into there. Where you can break out groups of functions - the phone and the camera may work for some segments but not others; some might never want it, or might never use it. As we get more and more digital, all this complexity has to be tamed in a way that the consumer can access it.

People are trying to solve the computer problem, not the human problem, which is aligning the whole value chain. Everyone is fighting each other left, right and center - when they should be focusing on delivering a service for a customer. No customers ever expressed the wish that they wanted WAP.

So what we've done at IssueBits is focus on a service, clearly the customers want to answer questions on their mobile phones, so we do that really well. We make sure it's easily accessible and easily viewable when you get it back. That's the way to bring in the next level of services.

But we're seeing non-Symbian smartphones with features that we never thought we'd see, like the Sony Ericsson 700s - removable media, gigapixel cameras and Bluetooth. Isn't that good enough for most people?

They are! 2G or 2.5G has carried value through. In terms of the radio, it's all integrated into that OS. And there's hell of a value to staying on that platform. There's a lot of engineering in 2G and 2.5G, and a lot more is required to leave it.

But as you get into 3G, that problem reverses, because there is so much complexity that you can't manage. For example, you have to have fallback to GSM, and do CDMA. There, the cost value falls apart and it's much better to get a Symbian OS phone. DoCoMo weren't coping with a proprietary OS for 3G; they've licensed Symbian and now they are very happy.

Isn't there cost pressure, and Symbian phones are more expensive to make?

Not really. 3G adds the costs, and Symbian doesn't add much more. Nokia has Symbian in the middle range and it will push down. The 6600 will go where the 2200 is today.

Won't a Java phone won't be significantly cheaper?

Perhaps once they've got a proper, clean cut version. But they've evolved it over time without being able to claw back features. AWT was replaced by SWING and they're both still in there. So it's a very big system.

Standards tend to be bad when they're arrived at in a higgledy-piggledy way, like Microsoft and Java. Symbian was arrived at in a very clean way. Then there's all the other rubbish you have to put on top of them.

Java's value, which I'm a great fan of, is being able to run objects anywhere. It's a distributed world. But Java is too slow for running the low-level stuff and it has a problem where it's not in control of its memory. It can only clean up memory. You don't have a control process, so it has to do garbage collection in a real time environment. If you run out of memory, you're dead. That's a fundamental problem which Symbian doesn't have. [Symbian announced a real-time capable version of its OS last February - ao]

But people talk a lot of twaddle about it because they don't know what they're talking about.

You simply don't see Linux cutting it?

There's as much missing as there is in it. Motorola is trying to do that - but there's a huge amount to provide, it's hard to catch up from where Symbian is, it's well ahead of MS with all the communication stacks; you have to put all that in there.

The proposition remains consistent in that a standard for the next generation of mobile phones is a good thing;

But the demand for 4GB iPods has surprised everyone - critics said it was rotten value and it wouldn't sell. And two or three years out, phones will have 4GB of storage…

But if you think of user getting access to a service, what's the real possibility that a phone manufacturer is actually going to give you the whole package? There's the iTunes software, there's synchronization that works, there's the easy to access menu, the stereo headphones, the iTunes store and shop, the distribution of the songs - all that's got to be good. Maybe it will take ten years to get together, but it's a long process.

We as technologists remember our favorite things, not the whole product. For us to be the early adopters we'll put up with it because we like these. So what if there's a bit of sync missing. But that mass market won't do this. The more we add, the more we have to put it together ourselves.

[At Psion] we learned a lesson from Lotus in the early days. With Sinclair we produced four easy-to-use software packages and integrated them tightly. They were the four second best products in their field: word processing, spreadsheet, graphics, and a database. Lotus produced the best spreadsheet, so-so other applications and no word processor at all - and they won. The very best will always beat second best.

Symbian OS is a very, very good OS for a phone. It remains well ahead of the Microsoft offerings. All the comms side that we haven't seen will come into play yet. It's already in the phone - such as MobileIP and IPv6 - and that's real comms value for the Nokias.

Let's talk about VoIP - over here in California that's thought to be very revolutionary. I've always thought these arguments forget about incumbent advantage

Yes. VoIP is going to go over broadband, the incumbents sell broadband, and they have a way of connecting the local loop to the Internet, so the BTs are going to be very happy.

Some companies won't be here, some grow and some disappear. It's also true that gorillas in one sector that don't make that transition to next sector. They can skip a generation like IBM. So some people we know today will be gone.

IPv6 will be next thing we move to.

But one interesting question is will we get all the services and benefits we're promised. I learned from Ericsson that they invented 260 services for PBXs and four only ever caught on. Voicemail and Caller ID were a couple.

So what innovation and what services do you think we are going to see?

Ask yourself, what are people going to with all their pictures in the future? What are they going to do? Is writing to CD-ROM really safe? Sorry - it's gone in a few years. Are people going to do a 3-stage offering, or make one of their copies in an alternative geographical location? Nobody does that.

With digital you can do things better; for a really simple straight forward things.

No one has designed architecture for the home. We've got Wi-Fi and broadband and Bluetooth but there's no way to put it all together.

So who, then? We've seen that even with the best intentions Wintel can't do a good job. It has to come from the consumer electronics people;

So it has to go back to being vertically integrated; you have to tackle the product offering yourself. You start doing something vertically because you can't work with everybody. So somebody has to break through, starting with a niche.

Like Apple?

That's one way. Whoever does this has got to do the hardware, and the software, and the systems infrastructure, and not many people can do that; and they must have a brand that the consumer respects. On the one hand they have to be known for style going into the home, and on the other be able to manage infrastructure. And they've got to be big.

So they need to establish a beach-head, and some companies wouldn't even bother to try to cross this chasm. And it needs a really big organization to be able to deliver. So I don't even know if they know they should be doing this.

But isn't there a real advantage to an open network?

Yes but the voice problem doesn't go away.

It's like a magic trick. Skype is not offering a whole product in a mass market. It's in a small market and it's a chimera. Skype couldn't roll out their service to compete with anything, globally. OK, they might be able to, but it would be an awful and probably still couldn't get it to work everywhere you go. That's even true for 3G, now!

What would you do differently, if you had your time as CEO again?

We wouldn't have spent time on user interfaces. We'd have left that much earlier. [In 2001, Symbian left the business of designing UIs to its licensees, with the exception of UIQ, which remains part of the company]. Everyone was keen to share and we tried hard for two years, but it was never going to happen. Everything about those companies [phone OEMs] is based in their own UIs. So that was two years wasted.

In hindsight we came to the right view; but we never learnt that lesson. There were other things people were keen for us to get into early, for example WAP. We could never have NOT done it, but I had a pretty good feeling it wasn't going to be worth it. But I wasn't the customers.®

Related stories

Former Symbian, Psion boss answers all your questions
Symbian raises £50m in rights issue
Java and DRM key to mobile ambitions
Mobile operators push next gen services