Nokia retreats into its manufacturer roots and gives the keys of his kingdom to Microsoft: Symbian, purposely created to hold Microsoft back from the mobile space, much to the ire of Gates, is gracefully terminated by a Microsoftie in exchange for a promise of a better mobile operating system in the future and hundreds of millions of dollars in kind. The transition will be so hard that Nokia will not provide any forecast for the full year and the stock has fallen 15% on a day: a master deal for Microsoft, a disaster for Nokia. No software, no soul.

The economics underlying mobile operating systems and their adjacent ecosystems are very different from the PC space: more complex, faster releases, costlier product recalls, energy constrained, lower margins.On a typical PC, >30% of its price goes to Microsoft; on a smartphone, less than 7%. That is, the incentives to get a perfect OS release from day one are much lower, and that for a company that is used to deliver good software after 2–3 versions. Only the much better post-handset-sale incentives of monetizing mobile apps through advertising can align the two companies to succeed on this new deal, except that this model is very disparate to what Microsoft is used to.

On a modern smartphone, the screen consumes more power than the CPU (>3x), so the power advantage of the ARM over Intel chips gets diluted, even taking into account that differences in use between users are huge, rendering power consumption characterization impossible (see Exhausting battery statistics). That ARM chips are so energy efficient it’s a history of unintended consequences and path-dependence: the packaging of the first ARM chips were made in plastic to keep costs as low as possible, instead of using the more common ceramic packaging, as in Intel chips, so they had to keep the power dissipation below a watt. And now, it’s also path-dependence what keeps Intel outside the mobile world: everyone will keep on using ARM chips, even if the power efficiency advantage gets irrelevant.

Will Intel dare to change from being a vertical integrated manufacturer to an open licensor, just to beat ARM, even if that means lower profit margins? Or will they risk seeing their ISA get less executed all over the world?

Symbian C++ programming has the steepest learning curve ever. Just before Stephen Elop burns the platforms, or not, let’s remember the tools introduced by Nokia that tried to reduce it, with mixed results:

• fshell is simply the most useful multi-purpose tool for the Symbian environment: the equivalent of bash + telnet + rlogin + a POSIX-like set of command-line tools. It will drastically reduce your compile-upload-debug cycle.
• Flowella will help you to quickly create mock-up prototypes for market research purposes.
• Qt Creator is Nokia’s biggest bet to save Symbian. Best suited for applications in which 80% of the code involves GUI tasks, to speed up development efforts. Unfortunately, few are the smartphones that have pre-installed support for it and Ovi isn’t distributing Qt apps for Symbian^3.
• Start with Open C/C++ if you are not proficient on Symbian C++ and don’t want to know why there are dozens of classes for strictly the same things.
• Symbian’s source code, which was the biggest open-sourced project ever. Huge but very well commented: if your application keeps on crashing and you think it should work, trust yourself, the bug it’s in the kernel and the source code will help you to track it down.