Why does Symbian fail?

On the way home, I joked with my colleagues that if Nokia or Symbian asks me to be the technical director of Symbian OS, I will tell them how to direct Symbian to the correct direction. Of course, this is just a joke, but it does not seem to prevent us from correctly thinking about this issue.

In 2006, I began to work as a real security software for the smart mobile platform (before that, I had tracked security trends on Palm and also had an original program ). At that time, the platform was dominated by Symbian (including s60 and uiq) and Windows Mobile. After experiencing painful learning, sustained frustration, and using codewarrior, which is said to be the best integrated development environment at the time, I boldly predicted the decline of the Symbian platform. Now, people with eyes can see this. I made this prediction based on two points. From now on, although the core issue is not touched, it is basically correct. These two points are: 1. Insufficient preparation for touch screen devices; 2. Difficulty in developing Symbian applications.

At the first point, it is completely a Nokia issue, not a Symbian issue. As we all know, Nokia has always been the actual maker of Symbian, and Nokia has always had a strange attitude towards touch screen devices for some reason. However, this is not Symbian's attitude towards touch screen devices. In fact, the Symbian system has been supporting Touch screen devices since the very beginning, and even the uiq subsidiary has been committed to the evolution of the upper bound surface of touch screen devices. However, Nokia has never tried to use the uiq interface to launch devices. On touch screen devices, it has only launched a 7710 model (the predecessor of 7710 was not productized ). 7710 of the interface system is named S90 by Nokia. As a matter of fact, even today, 7710 of interfaces cannot be outdated. If they can continue to evolve until now (there is another premise, that is, Nokia does have product personnel who care about the user experience on the touch screen), should not be inferior to other systems, at least not lose to s60 Fifth Edition. Interestingly, S90's UI prototype has been moved to maemo (hildon), and the latest development of the S90 has become the treasure of Nokia's high hopes.

On the second point, we all know that the learning curve of developing on the Symbian platform is steep. Even a veteran, for all kinds of descriptors (a thing that is almost unavoidable) as well as clearing the stack (another thing that is almost unavoidable), it is also difficult to fully master and write error-free code in a short time. I have always insisted that, clearing Stacks is a mechanism that makes it easier to introduce errors (error-prone mechanic ). Moreover, no matter how Symbian claims and advocates it, it is still an imperfect C ++ development platform, and cannot even be supported by STL for a long time. When developing in other languages, Symbian developers may have to bits their brains to implement a map while thinking more about what functions they want to implement and what kind of experience they will bring to users.

Development difficulties severely suppress application boom, especially for a large number of enthusiasts. The signature mechanism introduced since Symbian 9 makes this situation even worse. I cannot write any exciting applications (due to the limitation of capabilities). Even if I write them, it is difficult to promote them, this is because you have to pay a high fee to Symbian or its related certification/testing institutions before users use it. In this development ecosystem, I think it will be useless for Nokia to make OVI store better.

So now I think what is the root cause of the decline of Symbian? It is a conceptual issue. Both Symbian and Nokia seem to have been designing mobile phones with the idea of being an embedded device, rather than being designed with the idea of an open smart device. The direct consequence of this unopen idea is that it does not pay attention to the evolution of development platforms. In 2008, we could also hear the suggestion from Symbian to reply to the partner's request to improve the building system: why should we change it? It's working very stably? I have already said more than once that Nokia has missed the opportunity: At the beginning, Borland's Development Department was sold at very low prices, and Nokia was indifferent. You Need To Know That Borland has a deep foundation in development tools, c ++ builder has even been launched for Nokia mobile terminals! Nokia's shortcomings in this regard (for the basic software architecture, such as the operating system and compiler) have a profound impact, and maemo's fast forward is also one of the evidence, if it is not in a critical moment to form an alliance with Intel, the future of meamo is even more unknown. Even now, it is hard to predict how much meego can contribute to Nokia's mobile phones. Another mistake of Nokia is that it purchased Qt in the past few years. To be accurate, from a purely technical perspective, buying QT is not completely wrong. However, if you think about Nokia's purpose of purchasing QT, it can be regarded as a silly decision, nokia hopes to use it to quickly develop cross-platform applications, so that the number of applications on Nokia mobile terminals can grow rapidly. This fact proves that the technical decision makers of Nokia or Symbian have a serious C ++ complex, and they completely ignore the direction of mobile technology! Let's take a look at the current mainstream development platforms. None of them are controlled environments or almost controlled environments. Android Java, iPhone's objective-C, and Windows Phone are aware of this trend, so they are brave enough to migrate Development Technology to the. NET platform.

In conclusion, the decline of Symbian is not clear about the development prospects of Science and Technology, neither innovation nor identification or follow-up of innovation.

The last chance of Symbian. Symbian is indeed an excellent operating system (stable and secure), But Symbian has always lacked the clear understanding that Symbian is not an outstanding application development platform. It exposes too many systems to application developers without any obstruction during design or implementation. As a result, the latter has to face a strange system, compared with many other more mature operating systems, although similar in principle, it is quite different in usage and adds a lot of complicated operations (such as explicitly connecting to many servers, even file services are no exception), but also adopt a strange development language (the so-called Symbian C ++ ), the r class provided by the system does not even support auto-disabling associated open resource handles in the destructor. Currently managed by the Symbian Association, I still don't see any signs of turning over, either Symbian ^ 3 or Symbian ^ 4. If I want to provide a solution, it is to clarify and strictly control the interface level of the regional subsystem and application, and keep the C ++ development status at the underlying layer, the application layer is switched to other controlled languages or development tools that are more advanced and more productive. In this way, the application is isolated, the security risks to the system are reduced, the learning curve and cycle are greatly reduced, and the development efficiency is improved, it can attract more developers. The specific method, I personally think, either porting Dalvik to achieve compatibility with Android, or acquiring Red Five labs (http://www.redfivelabs.com) to integrate improvements using. NET development. Alternatively, you can also consider using the Flash technology in an all-round way. Unfortunately, I have a little understanding of this aspect and cannot determine whether it can meet the vast majority of development needs.

By the way, meego, I am not very optimistic about this system. Especially for Nokia mobile phones, it may be better for netbooks or other terminals. In terms of development, meego is now one step behind Symbian. the development environment does not support windows, and it is still the main development language C/C ++. According to the current situation, meego is likely to be the next Symbian.

 

Update/correction (January 1, July 31, 2010)

:

1. Nokia has developed a uiq System touch screen mobile phone with a model of 6708. It has been launched for the Chinese market and has poor market performance. By the way, this device is actually an ODM variant of BenQ P31.

2. The programming controlled language mentioned in this article should be more accurate and should be described as a controlled language/dynamic language. In terms of mobile platforms adopting this development model, the Blackberry platform, which has been widely used recently, is missing.

3. The error "meego development environment is not supported by Windows" mentioned in this article is explained simply by the previous maemo development experience that I have not tried Nokia's latest QT SDK.

 

Thanks to blue_light
He gave me a message to defeat the inertia and attached the above updates.