By mid-2005, Android had taken over and the future looked bright. But just six months earlier, things were not quite as rosy. In January of that year, the startup was in need of money and their first task was the same as for most startups: getting funding. After pivoting from a camera OS to an open source platform, they still have the daunting task of building a product, which means they will need more money to hire a large enough team to do the job.
So the company focuses on three things. First, they need a demo to show what’s possible. Next, they need to define their vision and create a pitch deck to help define that vision. Finally, they need to take the demo and slide deck on the road to tell their story to potential investors.
The first job for Andy McFadden (known to the team as “Fadden”) when he joined was fixing the demo, a prototype phone system that Brian Swetland and Chris White were working on. It’s not really functional (for example, it shows a product icon on the home screen which uses a set of hard-coded icons and stale data). But the demo represents a vision of what the product could be when it is implemented.
One of the apps that Fadden added to the demo was a simple calendar app. This early demo project will come back to haunt you. After several intervening years of working on things throughout the Android platform, he ended up helping out with the Android Calendar app. Time stands for no man…but calendar apps do.
The mobile advantage
As the team fleshed out their vision, they created a slide deck to explain it. These slides paint a picture of the opportunities they see for Android in the marketplace, as well as a picture of how Android will make money for investors.
The slide deck in March of 2005 had fifteen slides, which was enough to get the attention of VCs and Google.
The pitch deck is topped off by the second slide, which compares PC and phone products. In 2004, there were 178 million shipments of PCs worldwide. At the same time, there were 675 million phones; It has almost four times as many units as PCs, but with processors and memory as powerful as PCs were in 1998.
This potential in mobile application is a point that Dianne Hackborn, then at PalmSource and finally on the Android team, is still thinking about. The mobile industry is ready to take off because the power is finally there to find a real, powerful computing platform: Dianne says, “You can see the writing on the wall. The hardware is getting more powerful, and the market is bigger than PCs. “
The presentation also identified the problem of the development cost of mobile software. The cost of hardware is going down, but that of software is not, making it a larger and larger portion of each acquisition cost. But phone manufacturers are not experts in software platform development and do not have the skill set or the need to provide the increased capabilities needed to differentiate their software from that of their competitors.
The second important point in the pitch deck is that there is a gap, and opportunity, in the market for an open platform. That is, Android will be an operating system that is free and available to manufacturers through open source. Companies will be able to use and distribute this OS on their own phones, without looking to the software provider and without having to build it themselves. This open method was something that simply did not exist at that time.
Microsoft provides a proprietary OS that manufacturers can license and then deploy to their hardware. Symbian was originally used by Nokia, with some adoption from Sony and Motorola. RIM has its own platform, which it uses only for its BlackBerry devices. But there is no choice there for manufacturers who want a powerful smartphone without either building their own OS, putting significant effort into existing customization, and/or paying a high licensing fee.
Even more problematic, existing systems fail to provide an ecosystem for applications. Symbian provides some basic infrastructure for an operating system, but the UI layer is left as an exercise for the manufacturer, resulting in an application model for phones where applications written for a certain flavor of Symbian will not necessarily run on another difference, even on phones from the same manufacturer.
The Java programming language, known in the server and desktop PC world as “write once, run anywhere,” may have provided such cross-device application capabilities, but Java ME falls short of this in the mobile space. While providing at least the same language in all devices (like Symbian provides the same language of C++ for all its implementations), Java ME addresses various form factors and operating systems in phones by providing different features. of the platform, called profiles. These profiles have different capabilities, so developers need to change their apps to work on different devices, and often that approach fails when the capabilities are different on different devices.
Linux to the rescue!… Almost. Texas Instruments (TI) provides an open platform based on the Linux OS kernel. All manufacturers need is Linux itself, the reference hardware from TI, and then a huge host of other modules that manufacturers have to acquire, license, build or otherwise provide to create their own machine. As Brian Swetland said, “You can use TI TI’s OMAP chips to build a Linux phone. So you need TI’s OMAP and then forty components from forty different middleware vendors. You put all these together and you integrate them all and then you will have a Linux phone. And that’s just stupid.”
Android wants to provide the world’s most complete mobile device platform solution. It will be built on Linux, like TI’s offering, but it will also provide all the necessary pieces so that manufacturers will have a single system to get to build and ship their devices. Android will also provide a single programming model to app developers, so that their apps will run the same across all devices on which the platform runs. By having one platform that works on all your devices, Android will make phones easier for manufacturers and developers.