Now mobile phones are struggling to evolve from the 1980s position. Windows made Microsoft a lot of money, so there is no shortage of people – Microsoft included – who want to be as big as Windows on a device as small as a phone.
Cheap phones still have their manufacturers own operating systems: Motorola uses P2K, Nokia Series 40 and Sony Ericsson OSe. The immediate challenge is from smartphones. Here we look at the leading contenders: where they came from, what it means to the manufacturers who choose them and what it means to developers.
Symbian started life as EPOC, the Psion operating system for pocket computers. Psion realised that they needed a system which could do more than the basic functions they had in their devices of the late 1990 and so Charles Davies led a team to develop something fast and capable of being extended. It was a small team, with Charles doing a lot of the work himself. Software that comes from a close knit group with a single vision is always better than that designed by committee. Psion has tried to hide this, but EPOC stood for Electronic Piece Of Cheese, which is probably no less meaningful than GPRS, GSM or HSUPA.
In a masterstroke Psion did a deal with the three major phone manufacturers (Motorola, Nokia and Ericsson) to create Symbian, an industry standard and defence against Microsoft. It was announced at the Old Brewery in London on 24th June 1998.
This started life as Windows CE, codenamed Pegasus. Having conquered every desktop, Microsoft was making a bid for the kids bedrooms with the Xbox and their pockets with Windows Mobile. Microsoft took very different approaches with the two. With Xbox they hired games industry people and did it the industry way, after all they had learnt through the failures of the Dreamcast and a strange home video device sold through the Radio Shack. With Windows Mobile they treated the phone as a little computer.
Codenamed ‘Stinger’ and nicknamed ‘Stinker’, the first ever review appeared in What Mobile in September 2001. The manufacturers Microsoft approached were not the big guns, but PC clone makers like Compal, Asus, Mitac and HTC. Windows was chosen by UK start-up Sendo. They were going to teach the phone industry how to do smartphones. How hard could it be?
One of the companies that was willing to be taught was Orange, which took the HTC Windows Mobile phone codenamed Canary and launched it as the SPV100. One of the people charged with making the project work was Rich Miner. He joined Orange when they’d bought his voice recognition voicemail company, the legendary Wildfire.
Rich worked for Orange Ventures based in Boston, taking cool new technologies and getting them to market. He hated Windows Mobile, because when he wanted a minor change he found Microsoft to be slow and intractable. Something simple might take two years. Thinking that there had to be a better way he left Orange and set up a company called Android with Andy Rubin. Small and focused (where have we heard that before) they built their pocket computer operating system on Linux. Android was bought by Google in July 2005.
Apple was rumoured to be working on a mobile phone for a long time before it happened. It held off because it didn’t want to have to acquiesce to the demands of the networks. Apple invented the term PDA with the Newton, but the device was too ambitious and it’s cursive (joined up) handwriting recognition didn’t work very well.
The only licensed device to ship with the Newton OS was the Motorola Marco pager. The first Apple phone venture was a Motorola too, the Motorola E398 which allowed 100 tracks to be loaded from iTunes. The limit and the poor software implementation in Java meant it was never going to fly. The iPhone OS, like that used in Mac OS X, is based on Unix (which gave us Linux), and originally developed at Carnegie Mellon University. Just before launch, programmers were taken off OS X to work on the iPhone to ensure it shipped on time.
The first Symbian phone was the Philips Illium Accent. This was a two-part device where the Symbian bit detached from a standard phone. Unfortunately delays in the smart bit meant that the phone bit was too naff to sell when it was all ready – it never saw a shop.
The first Symbian phone you could actually buy was the Ericsson R380. Touchscreen with a flip, it might have been Symbian inside but it was a closed operating system; you couldn’t add your own applications.
The first open Symbian handset was the Nokia 7650. This was also the first phone with Series 60 and Nokia’s first camera phone.
Windows Mobile had similar false starts. A Compal prototype was shown at lots of trade shows but never materialised. Sendo was the lead manufacturer, with a team composed of people who had worked at Motorola and Philips. The phone got to within days of being launched – posters and point of sale material had been sent out to the shops – when a legal wrangle caused them all to be recalled. Under the terms of the settlement between Sendo and Microsoft all phones were crushed.
What Mobile reviewed the Orange SPV100 against the Sendo Z100 and the Sendo won, but it was the SPV100, built by HTC, that ended up in the shops. Indeed HTC seems to be the only manufacturer which has really got to grips with Windows Mobile. Other attempts, such as the innovative Motorola MPx which had a dual hinge so that it could be used like a phone or an organiser, failed when Motorola believed Microsoft’s claims of what was necessary for a minimum hardware specification. This phone only shipped in limited quantities in Asia.
The iPhone is too new to have much of a history. It was launched in 2007 to mass euphoria with people queuing in the streets to get one at the stonking price of $399. When the queues withered much faster than Apple expected, the price was cut by $100 and all those who’d queued were placated with a voucher to spend on more Apple products. It was criticised for being 2G, but when the 3G version appeared what really made it special was the App store. A great way to buy really cool software for the phone.
And the Android phone is even newer. The first phone is of course the HTC built G1 for T-Mobile. There is also the ‘Kogan Agora Pro Mobile Phone Featuring Android™’, although this has now been ‘delayed indefinitely’. Rumour has it that Google was unhappy with the implementation of Android on the Kogan’s small screen. There is definitely something more to it than Kogan just wanting to maintain future compatibility. Small manufacturers don’t pull a product from launch weeks before shipment unless there is a bigger story behind it. Certainly not when they have taken deposits.
Expect a rash of new phones from small manufacturers and the likes of Motorola, Sony Ericsson and others.
Windows Mobile is a traditional closed operating system. Any phone manufacturer can go to Microsoft and licence the software. They get help (a lot of very good help these days) on how to make it work and build a phone around the platform. It’s not great at some things, such as adding new languages or things the networks want (eg. Vodafone Live). These must be added by Microsoft and can take a long time. The whole look and feel used to be defined by Microsoft, even down to which manufacturer’s screen was used, but recently this has changed allowing for innovations such as the cool front end on the HTC Diamond and the Sony Ericsson Panels.
Symbian is now a Foundation. The development is owned by Nokia and everyone who used to work for Symbian and hasn’t been fired now works for Nokia. Members join the foundation if they want access to the source code of the operating system and want to put it into a phone. They have to join, but this only costs €5,000. The user interface is officially in three separate parts: MOAP which is only used in Japan, UIQ which was used by Motorola and Sony Ericsson and Nokia’s Series 60 which had been licenced to LG, Siemens and Sendo. All the user interfaces have been contributed to the Foundation. Motorola has given up on Symbian and Sony Ericsson on UIQ, so Series 60 is now the only real option for a manufacturer. People who’ve worked with it say that the Symbian performance issues are down to Series 60 and not to Symbian.
Android is completely open source. Any would-be handset manufacturer can download it from the web and put it into a phone – they don’t even need to ask Google first. Indeed this is what Kogan has done. Manufacturers are free to make whatever changes they like to the software. Google would like to think that manufacturers would keep the standard core applications which include the phonebook and browser, but there is no compulsion to do so. Officially Google does not offer help to manufacturers (although they are doing so in the early days). Ultimately the need to support anything new becomes the problem of the person who has downloaded it.
iPhone is about as different as you could imagine. The software source code is not available to manufacturers or developers. Most manufacturers work with companies who supply parts of the software – things like the browser – and those ‘second parties’ do get a development kit with heavy secrecy agreements. Apple does everything itself.
You might think it’s odd that America is so far behind in mobile phone usage, seeing as so many of the mobile operating systems come from the US. This is because it’s the most advanced in web usage and this is the core of the new Palm Pre operating system. Like Apple, Palm is keeping the WebOS for itself and does not seem to be actively recruiting partner manufacturers. While the mobile world has been treading carefully in the world of the internet with browsers improving slowly, the internet world has slowly been discovering mobile and Palm has started somewhere in the middle.
The Pre is a mobile surfing device which brings the browser to the front of the operating system. There is a QWERTY keypad, but the 12 number keys are a minor function called ‘up through software’. Voice is less important than browsing. All the advanced technologies like Java and Flash are supported (which the iPhone doesn’t do)making it a great platform for web developers. It is based on a framework called Mojo, providing the building blocks for developing applications, and because they are supplied the graphics and boxes feel familiar over different applications. Web developers will flock to the Pre because it’s an environment they feel comfortable with. This will ensure a lot of good applications, but they will be new to many of the things mobile developers take for granted such as intermittent connections, battery life, using the device one-handed and limited storage.
The Pre uses the CDMA technology on the Sprint network in the US and it will no-doubt rival the iPhone. A 3G version is expected this summer.
The term Open Source is widely used but I suspect that many of the people who use it don’t know what it means.
Computers are billions of switches that can only be on or off, represented by the 1s and 0s we hear so much about. Those switches can be put together to form ‘gates’ that can add, subtract, increment, multiply and divide numbers. The numbers can either be instructions for the computer to do something (where feeding it through a gate makes it do that something) or a number which can have something done to it. In the 1950s all computer programs were written as numbers, called machine code. Programming was made easier by giving the numbers a mnemonic, typically a three letter abbreviation which described what it did. So instead of typing ‘96’ for ‘return from subroutine’ you typed ‘RTS’. The mnemonics still lined up with what the gates in the chip did. To turn the programs from the text into the numbers the computer could use, the text was run through a program called an assembler, and the code is also known as assembler.
A major change came when high-level languages were developed. These used more English-like commands to represent combinations of mnemonics. These languages, like C, Pascal and Cobol are generally used to write programs today. Only very specialist jobs are done in assembler. The high level language is run through a program called a compiler. The compiler translates the program into assembler and then the assembler into machine code.
The code written in the original high level language is known as ‘source code’ and the end result, the machine code, is known as ‘object code’. When you buy a program like Microsoft Word or Photoshop you get the object code – this cannot reasonably be altered to change what it does. If a programmer wanted to swap the ‘home’ and ‘insert’ menus around it would be a huge amount of work. With the source code it would be reasonably easy. Until recently it’s been normal for phone manufacturers to only get the object code from the operating system suppliers. Indeed this was the subject of a major falling out between Sendo and Microsoft. If Sendo wanted to change Windows Mobile – say to include Bluetooth – they had to ask Microsoft to do it for them.
Open Source is different. Not only do Google give the source code to phone manufacturers, they publish it on the web. Anyone who wants to build a mobile phone can download it, use it and make any changes they like. If they want the phonebook sorted by phone number instead of name they can. And it isn’t limited to manufacturers. Anyone can make changes to the Google Android operating system and then try and sell those changes, either to people with Android phones or to the manufacturers. If the changes are good they will be put back into the public domain so that other programmers can improve on it further. The idea is that in a Darwinian evolution the source code will get better and better.
There is a halfway house between the two extremes of closed and open source. Operating Systems like LiMo and Symbian require you to join a club – paying an annual fee – to get access to the source code.
Symbian is in a state of change. It aims to be more like Android – anyone can download and alter any part of the operating system, but in the past it was only available to licensees. There is a developer registration fee of €1500 but in time it is expected that this will only be necessary if a developer wants their code to go back into the mix.
No-one does developer support better than Microsoft. It has the best conferences and tools. Indeed at the birth of Symbian it did a lot of soul-searching about using Microsoft C compilers to develop for Symbian and decided that it would use the rival’s products as they were the best. The technical documentation is excellent. There is an easy path for developers who have worked on a PC to move their application to Windows mobile.
This means there is a wealth of software out there, although a lot of it is not very well optimised for handheld or single-handed use. What developers don’t get is much flexibility. There is no access to source code. If for instance you wanted to add a new field to the address book you could not do this.
It’s a little hard to find out about iPhone development because all developers are required to sign a Non Disclosure Agreement that prohibits them from talking about what they are doing.
One developer who has found a way around a problem can’t share that with another. The development tools are reputed to be excellent, and creating applications that look and feel like iPhone applications is comparatively easy. There is no access to source code and some quite tight restrictions on what the device is allowed to access within the phone. The only way to sell an application to an iPhone user (at least one who hasn’t unlocked the phone) is through the Apple Apps Store on iTunes. Here Apple has control and won’t let you sell anything which might compete with programs Apple has or plans to release. You can’t, for instance, sell a SatNav program or anything that might compete with iTunes.
The Google mobile operating system is as open as it’s possible to be. All the source code is available and can be modified. Whatever you want to change you can. Android has been ported to the Asus eee computer. The mechanism for programming an Android phone is what Google refers to as ‘native java’ which is the Dalvik language. This liberated approach gives the maximum possible flexibility, but with that comes fragmentation. It’s likely that different manufacturers, networks and programmers will all do radically different things.
If a manufacturer changes something fundamental it’s quite possible that a program which works on all other Android phones won’t work on the changed one. For developers this means a nightmare of testing, and because Google isn’t responsible for what other people do to their code the developer can’t look to anyone for help.
All the manufacturers realise that a healthy supply of good programs is important. It might even lead to the program that changes the way people lead their lives and which establishes their product as The One To Have, the elusive ‘killer app’.While many (mainly the networks) call for a reduction in the number of operating systems, there seems to be an endless number of people trying to be The Next Windows – the mobile OS that will dominate.
Symbian is ten years old and despite having the best possible start in life hasn’t managed it. It’s doubtful that a winner will emerge anytime soon.
Follow Catherine Keynes’ blog on the politics of the mobile industry at www.catkeynes.com
