The Symbian OS (Symbian Books), страница 10

(A Communicatorwas the first Symbian phone to offer Wi-Fi connectivity, for example.)Figure 2.4 The Series 60 user interface, as used on the Nokia 765018David Wood believes that Nokia’s work on ‘square’ had been in progress for at leasttwo years before the formation of Symbian.34THE HISTORY AND PREHISTORY OF SYMBIAN OSFigure 2.5 The UIQ user interface, as used on the Sony Ericsson P900Figure 2.6 The Series 80 user interface, as used on the Nokia 9500While a number of licensees worked on Quartz devices and others(besides Nokia) expressed interest in Crystal-based devices, the DFRDstrategy eventually stalled. The reality is that, although they aimed tobe generic, the designs could not escape the pull of licensees.

In theend, they were more licensee-specific than generic, reflecting particularlicensee’s views about what a phone should be. Nokia drove Crystal;Ericsson and then Sony Ericsson drove Quartz; and Sapphire seemedto split and split again, until there was a one-to-one mapping betweenDFRDs and licensees.Martin Budden’s view then and now is that the problem was essentiallynot resolvable. It was not possible to agree on a Symbian-based userinterface, in other words one evolved from the original Eikon GUI ofEPOC and the Series 5, which was suitable for the different device visionsof Nokia, Ericsson and Motorola.And this was the problem.

Each phone vendor had different, deeplyheld views about what makes a phone. Symbian was trying to create aDEVICE FAMILIES35software platform that would satisfy them all. Motorola wanted a penbased user interface. Nokia wanted a keyboard-based user interface.Nokia did not place much value in the power of the pen and to datehave only ever released one pen-based range of Symbian OS phones,the Series 90 user interface on the 7700 and 7710.

Quartz, coming froma design unit which had started out working on Microsoft Windows CEdevices, came up with a pen-based tablet format, such as the CompaqiPaq or the classic Palm devices.Not everyone in the company was convinced by the DFRD strategy. To some it seemed more like an attempt to paint over the deeperproblem, that conflicting licensee user interface requirements were irreconcilable.Martin Budden:In my view, we just could not resolve the issue. We couldn’t come to anagreement on what the Symbian-based user interface would be that wassuitable for all licensees, which I think was ultimately why Nokia went off todo their own.The designs were not so much generic as licensee-specific, reflectingeach licensee’s views about what a phone should be.Martin Budden:The DFRD idea was to have families of user interfaces, so there would beone family for devices like the Nokia 9210 Communicator and the Series 5;there would be a family that was based on the Ericsson R380 which was aphone; and at about this time, Quartz started up and that was for anotherEricsson phone with quarter-VGA tablet form-factor.

There was a basic conflictthere between Nokia and the Series 5 user interfaces, because what Nokiawanted went further than the customizations the user interface could easilyaccommodate and then the other conflict that started to manifest itself was auser interface for a phone form-factor smaller than even the Ericsson R380.After his time on the Ericsson R380 project, Budden had movedacross to work on Quartz as technical lead and spent the best part ofa year commuting between London and Ronneby.

He witnessed thedifficulties of implementing Quartz as a product-ready, concrete instanceof a DFRD at first hand. Partly the problem was one of resourcing,with the company’s main focus dedicated to the underlying featuresof the operating system, many of which were driven by the ambitiousrequirements of the Nokia 9210 project. The fact that Quartz was being36THE HISTORY AND PREHISTORY OF SYMBIAN OSdeveloped at a remote site did not help. Nor was it necessarily easy forthe Ronneby engineering team to adapt itself to the very different styleof Symbian, of which it was newly a part, compared with its previousparent, Ericsson. For its part, Symbian probably found integration of anew remote site just as painful. Beneath it all, were the basic engineeringproblems.Martin Budden:The Quartz team had very great difficulty getting anything done which wouldsupport their work and that led to a lot of fragmentation and reimplementationand it also highlighted that there was a lot of code that was not easily separable.So, for example, the messaging code had user interfaces in the engine layer,which meant that to change the user interface we had to redo a lot of theengine code as well.

So there were a lot of things that made it difficult toseparate out the bits. And this is when the idea of resolving all these conflictsby defining DFRDs emerged. I thought at the time that it was never going tofly.However, the strategy served its purpose as, between 2000 and 2002, itenabled the important focus to become that of developing the underlyingoperating system. It also underwrote the splitting of generic from specificfunctionality in Eikon, the original EPOC GUI, as a necessary engineeringstep to enable the creation of the DFRD variant implementations.When it abandoned the DFRD strategy in 2002, Symbian made atactical retreat out of the user interface space altogether.

The Pearl DFRDwhich was being actively developed in collaboration with Nokia, wastaken over entirely by Nokia to become Series 60. Quartz, which by thistime was the basis for projects with both Sony Ericsson and Motorola,was spun out into a new Symbian subsidiary, UIQ Technology AB, basedat the Ronneby development site.

UIQ became the name for the userinterface.19 Japanese licensees working under the FOMA umbrella, asDoCoMo’s new 3G network was branded, went their own way andcollaborated on a common user interface known as MOAP.20Symbian’s strategy since 2002 has been based on the concept of a‘headless’ delivery to its customers with a custom user interface integratedas part of the product creation lifecycle, either by the phone vendor (inthe case of the FOMA licensees and Nokia) or by a user interface vendor19In late 2006, Sony Ericsson announced that agreement had been reached for it toacquire UIQ Technology AB from Symbian.20FOMA is DoCoMo’s 3G network, which launched in 2001.

Mobile ApplicationPlatform (MOAP) was originally an internal designation which has now begun to appear inpublic forums.OPERATING SYSTEM INFLUENCES37(UIQ licenses its own user interface pre-integrated with Symbian OSto customers such as Sony Ericsson and Motorola; similarly, Nokia’sindependent Series 60 business unit licenses S60 preintegrated withSymbian OS to customers such as Samsung and LG).Symbian OS is GUI-centric in the sense that the user model is exposedonly through the GUI, while being designed into the operating system ata deep level; nonetheless, Symbian does not provide its own shippableGUI. This model has its challenges (as any model does) but, with thesales record as it stands, it can be considered proven.

It is driven bythe recognition that the mobile phone market is quite different fromthe desktop computing market, for example, in which users (as well asvendors) seem reconciled to the greater than 90% domination of a singleuser interface, Microsoft Windows. There are alternatives, in the form ofMacintosh and Unix/Linux, but these are not mass-market alternatives.The attempts to create mass-market alternatives (BeOS, for example, inthe late 1990s) have been spectacularly unsuccessful.Personal mobile devices, from phones to music players to cameras, arevery different.

Consumer devices, from TVs and DVD players to Hi-Fi andradio to car dashboard controls, are very different too.21 What users wantfrom them and how they wish to interact with them, is quite differentfrom what they want from the beige box underneath the desk.2.8 Operating System InfluencesWhile the user interface defines a system from the perspective of end-usersand translates the design philosophy of the system into tangible behavioraccessible to users, the real character of an operating system is defined ata deeper level, by the fundamental choices its designers make – typicallyin the form of trade-offs between aspirations and limitations, whether ofperformance against price, features against time-to-market or innovationagainst familiarity.

As a company of engineers rather than computerscientists, it seems that Psion absorbed multiple influences, but thenproceeded from immediate practicalities, almost as if no theory existed.Its first true operating system was written for the 8-bit Organiser andthen re-written for the next generation of 16-bit Organisers. Since thesewere x86-based, DOS was a possible alternative candidate. The companyhowever decided to follow its own path.21Christian Lindholm provides a fascinating public insight into the issues driving userinterface design for phones as we have known them, and for the multi-function mobileterminals they are becoming. These devices, as he says, are evolving from impersonal objectsto intimate possessions ‘containing one’s most important data and thoughts’ [Lindholm 2003,p.

154].38THE HISTORY AND PREHISTORY OF SYMBIAN OSCharles Davies:We considered using DOS and rejected it, which was controversial at the timebecause there was a view that DOS was the only kind of operating system forPCs downwards and of course it was also what the HP200 used, which wasour big competitor at that time.But DOS is a strictly synchronous system, single-user and single-taskingand it also provides an interface which insists on dragging the user downto the machine level, which was not the vision at all.

In fact Psion did laterrelease a DOS-based version (MC600) of the MC400 laptop, precursorto the Series 3, just as it had released a DOS-based version of the HChandheld. As David Wood remembers, after that first-hand experience ofMS-DOS, it became known in the company as ‘MS-dogs’.In contrast, the exposure to VMS was critical because it showed thatthere was another way.Charles Davies:From that experience of cross-compiling on VAX for the ZX80 we got to knowVMS pretty well.