Header image
Insider News
Every Sunday


How Things Work


Arts or sciences? If it is sciences – and I guess it is – physics or biology? I always preferred physics not just because the Nuns in my convent neutered biology for religious and moral reasons, but because in physics you could see how things worked.

I’ve dabbled with code but never got my head around object orientation. The stuff I did was embedded. You could map directly from the chemistry producing binary, to how simple gates could be built into adders and multipliers, to operators and operands being numbers – and the evil delight of self-modifying code – to a program. You could see how it all worked.  Even if we did do stupid things at school like a program to calculate when Easter was.

In those simple programs you could understand how everything worked. The contribution of Symbian to the new Symbian foundation is 10 million lines of code. No-one will ever be able to read it all let alone understand it. Code in phones has got hard. Really, really hard. Motorola reportedly has 7,000 people working on its Motomagx version of Linux, rough estimates show that it’s cost Symbian something like $500m to get to where it is today with Nokia spending a similar amount on S60. Add in UIQ and what Nokia spent buying content for N-gage and you are over $1.2bn.

Software has replaced chips as the major cost in a 3G phone.

The result of software getting bigger and more complicated is that it’s much harder to develop. You switch from the physics model (where you know how everything works) to the biology model where you have to find out. Biology is all about testing. What was the effect of a chemical? Why didn’t the same thing happen in all circumstances? We don’t really understand much about living organisms. Doctors might say “take these pills and see what happens”. Today’s software developer might say “I’ll compile it and see if it does what I expect”.

But in one important way software can’t be like biology: a drug may take ten years to reach the market as it goes through all the testing. A phone carrier wants to test in under six months.

It’s long been easier to get phones through approval if they used a common RF platform but with a shift to software the testing priorities change.

At the press conference The Symbian foundation kept saying “This isn’t a reaction”, meaning but not saying, iPhone/Android. The journalists found that hard to take, but even if the other OSes didn’t exist it would be necessary for the industry to reduce the overhead of testing and integration by using a standard platform.

Consistency won’t make programming less like biology and more like physics, it’s too late to turn that clock back. But it will make programming less like a science and more like an art, which is how it should be.


Cat Keynes publishes her thoughts on the mobile phone industry every Sunday at www.catkeynes.com you can read the column  the previous Friday by subscribing here.



The internet is ad-funded and after many dotcom failures it seems to work. It works too for mobile phones if you get your target audience right. Blyk has expanded into Europe.

And E-plus is looking at it too.

When Orange launched in the UK they talked about being able to buy your own signal boosters. Everything comes to those who wait, as femtocells start to take off.

The 3G iPhone might be a lot cheaper now that Apple has bowed to the subsidised model in order to make good on the target of 10m units in 2008 but there is still a healthy margin.

Charging for incoming calls. As I said last week, yes that would be stupid.


<<< Previous Sunday's Following Sunday's >>>




[Home] [Archive] [Subscribe] [Advertise] [About Me] [Contact Me]