From Phonebloks to Project Ara to Vsenn, we've seen different designs and implementations of how a modular phone should look and work. Puzzlephone is another such project, started in Nokia's home town (Espoo), that much like Vsenn (another company from Finland) focuses on the middle ground between total customization (like Ara) and no hardware customization at all (like most smartphones today).
The Puzzlephone will have three main "modular" components: the "Spine," which consists of the LCD screen, speakers and the phone's basic structure; the "Heart," which contains the battery and secondary components; and the "Brain," which has the processor and camera. This platform will use a fork of Android to begin, but it could also use other open source operating systems in the future, if needed.
On one hand, out of all the modular phone concepts we've seen so far, this looks like the most practical one. It's not made out of a dozen pieces that could lead to a less stable body. The three main parts seem to be stuck together with screws, which isn't very aesthetically pleasing, but you can be sure they won't separate from each other even when dropped. Plus, these phones will likely be covered with cases by their owners anyway.
On the other hand, the components from the three main modules may not come from the same company, which means either companies making one component (such as the processor) will have to establish some partnership with other companies (such as the camera makers) to build together such a module, or they will have to build the whole module themselves (less likely).
A third alternative, and probably the most likely scenario, is that middleman companies will appear that would, for example, take Sony's camera and Qualcomm's processor and assemble them together into one module. This would spare the component makers from having to build their own modules for a modular phone; however, they also won't be the ones to reap many of the financial rewards. The modules will then likely carry the branding of the middleman module assemblers instead of the component makers (unless some "Intel Inside" type of deal is made).
In the end, it's good to see different modular phone concepts battling it out in the market to see which approach is best. The one that will win is likely the one to get the most momentum from the component ecosystem. If customers can quickly start choosing between a range of hardware optimizations for a certain modular platform, that's the platform that should get the biggest momentum, which in turn will get even more component makers interested in participating on that platform.
Whether this whole idea of a modular phone will ever enter the mainstream market remains to be seen. Getting into the mainstream will likely require some "killer app" (or component in this case) that can't be found in regular smartphones but appears in modular phones.
Puzzlephone, just like Ara and Vsenn, should start shipping a working platform sometime next year, and then we'll get to see which platform is the most appealing and/or has the biggest component ecosystem.