The OS is always greener...

In a report from a developer meeting, Nokia officials said they're moving to Maemo Linux as the OS for their high-end smartphones. That resulted in an entertaining little obituary in the Register by Andrew Orlowski (link). But then later in the day Nokia clarified that "we remain firmly committed to Symbian as our smartphone platform of choice" (link). That in turn led to a lively online debate about what Nokia actually said, and the challenges that Finnish people face when speaking English (check the comments here).

It's just one more chapter in the long and exquisitely awkward saga of Nokia and Symbian. From the outside I can't tell exactly what's going on at Nokia, and it's possible that Nokia itself doesn't know. It's a very large company, and various groups there can have conflicting agendas.

But I can't believe that there would be all of these repeated reports, leaks, and artfully-worded partial denials unless Nokia were de-emphasizing Symbian in the long run. The most prominent theory, which I believe based on things I hear through back channels, is that Nokia does indeed intend to move to Maemo at the high end. And, as we all know, in computing whatever's at the high end eventually ends up in the mainstream.

I'm sure Nokia has valid technical reasons for moving to another OS. Nokia has said that there are some things it wants to do with its smartphones that Symbian OS can't support. But still the change worries me. Nokia's biggest problem in the smartphone market isn't the OS it uses, it's the user experience and services layer in its smartphones. Moving to a new OS does almost nothing to fix that. It does force a lot of engineers to work on writing a lot of low-level infrastructure code that won't create visible value for users. It also forces Nokia to maintain two separate code bases, which will chew up even more engineers.

All of that investment could have gone into crafting some great solutions, the things that are the only way to pull customers away from Apple and RIM. At a minimum, it's a terrible shame that Nokia spent so much time and money on an OS that couldn't take it into the future.

(By the way, this focus on the OS doesn't apply only to Nokia. I hear a lot of buzz from operators and handset companies who believe that if they just pick the right OS they'll automatically end up with great smartphones. Android is the latest white knight for most of them, but of course Nokia's not going to depend on a technology from Google.)

There's an old joke in the tech industry about rearranging deck chairs on the Titanic. I don't think that applies to Nokia because they haven't hit an iceberg by any means. But I do have a mental picture of a sweet old lady who spends all her time every day cleaning the bathroom while the food is spoiling in the refrigerator.

6 comments:

Tan Miaoqing said...

It's not about OS, it's all about mindset and culture; it's about OS too, as you simply have better graphic technology with Linux which results in better user experiences

Michael Mace said...

Interesting. I don't think I've heard anyone explain the Maemo vs. Symbian thing in terms of team culture.

Thanks for the comment, Tan. It's an interesting perspective.

Jody Fanning said...

I think that it has been the stated intention for a long time that Symbian is moving down in the product line. But that also means that it will still be the main OS by volume for a long time to come.

And also the Symbian that everyone loves to hate is becoming increasingly irrelevant. With the move to QT as the main native development toolkit the underlying OS will have less to do with it. I think it will be the release Symbian^4 that will completely drop the current UI toolkit leaving only QT.

So eventually I think Maemo or Symbian it won't really matter much for most end-user applications.

So, as they say, Symbian is dead, long live Symbian.

Pauli Ojala said...

It seems to me that Nokia is trying to move their focus a couple of levels higher in the stack. They'd certainly much rather have people talking about the Qt framework and its cross-platform wonder powers than the Maemo vs. Symbian dichotomy.

Qt will be the standard toolkit for both Maemo and Symbian (it also supports Windows Mobile and a number of desktop platforms). In general the Qt APIs are very complete and pleasantly uniform; I don't think most mobile app developers will ever need to interact with any OS-specific layers below Qt. Hence the divide between Symbian and Maemo becomes much less important for the user-level ecosystem and more of a hardware engineering decision -- Symbian still does support more devices and has a much smaller RAM footprint.

Consider the iPhone's operating system: the fundamental Darwin/BSD layer is certainly important, but the user-visible magic happens in the Cocoa Touch framework. It isolates developers from the base OS. Apple could replace Darwin with a very different kernel and the large majority of iPhone software could be ported with just a recompile. That's where Nokia wants to be with Qt.

(Cocoa and Qt actually have fairly similar histories. Both were designed as cross-platform application frameworks -- Cocoa used to be called OpenStep and shipped on Windows NT, even -- and both provide the primary API for a well-known desktop environment.)

Anonymous said...

Having had worked at Nokia for 17 months, 6 in Espoo)], i must tell you, i think you hit the nail on the head when you speculate about this flap stemming from internal politics.

The maemo team is run in Helsinki, (Ruoholati office), with an extremely small agile team that certainly believes that they will finish off Symbian at Nokia.

The Symbian S60 effort, on the other hand, employs several thousand, (perhaps > 10K) people; much of the effort is headquartered in Tampere, (100 kilo from Espoo), and they operate with a hub/spoke model where the S60 OS team feed into 'device program' teams that actually 'own' the delivery of a specific device end-to-end. This is in sharp contrast to the Maemo guys, who own their product end-to-end.

And the execs? Well my experience was that they are very happy to have such competition going on. The fact of the matter is, though, the investment in terms of people in S60 is so large, (and employee-protecting unions so powerful in Finland) that it is highly unlikely that you'd see substantial change in this status quo in anything under a 10 year period.

QT is certainly a good move on their part, but it doesn't change any of the dynamics with regard to employee allocation.

Just MHO.

Michael Mace said...

Thanks, everyone. Really interesting comments.


jody wrote:

>>With the move to QT as the main native development toolkit the underlying OS will have less to do with it.


Good point. But in that case, here's a hint to Nokia: If the OS doesn't really matter, stop talking about it. Nokia feeds stories like this by discussing Maemo vs. Symbian in public. If QT is the future, everyone at Nokia should be saying constantly that the OS is irrelevant, and obsessively talking up the benefits of QT.


Pauli wrote:

>>In general the Qt APIs are very complete and pleasantly uniform; I don't think most mobile app developers will ever need to interact with any OS-specific layers below Qt.


QT has a great reputation, and I'm a big fan of OS-independent software layers. But Pauli, don't you think it's a little pathetic that folks like us have to go around explaining Nokia's OS strategy for it?


>>Consider the iPhone's operating system: the fundamental Darwin/BSD layer is certainly important, but the user-visible magic happens in the Cocoa Touch framework. It isolates developers from the base OS. Apple could replace Darwin with a very different kernel and the large majority of iPhone software could be ported with just a recompile.

And the users would never notice because Apple's smart enough not to make a big deal about issues like this in public.


Anonymous wrote:

>>Having had worked at Nokia for 17 months, 6 in Espoo


Extremely interesting comment. I should remind everyone that we have no way of knowing if an anonymous poster is actually who he/she claims to be. However, the details in the comment match things I've heard from Nokia people.


>>they operate with a hub/spoke model where the S60 OS team feed into 'device program' teams that actually 'own' the delivery of a specific device end-to-end.

That model is, in my opinion, one of Nokia's biggest problems in smartphones. Responsibility for a particular product is spread across a number of people. If you want to design a great, integrated systems product, you need unified management of it.