With the demise of Nokia’s open mobile efforts, and exodus of key players, there’s a lot of talk on what could and should have been. So here’s my little slice of the story, and I hope many will add to it in comments.
I didn’t know what to say the day that former Nokia colleague John Wigginton laid a small black device on my desk. He grinned as I gasped. I had no prior knowledge of the Nokia 770 Internet Tablet then; its existence had been kept a guarded secret for some time. But I was familiar with similar devices since I had been putting together a plan for mobile auditing of our cell phone production. John knew of this, and was leaving our group to take another position. I had the pleasure of being selected as the next factory QA shepherd of this little black marvel.
I had almost given up on my mobile auditing quest, haven run into too many cost and technical roadblocks. But here in my hand was an internet tablet, fueled by open Linux, and I didn’t have to purchase it. Solved?
The 770 wasn’t quite up to the task at first, but through no fault of the hardware; I needed VPN for one, and there was no package for the 770 yet. So this meant going beyond the scope of the usual factory QA engineer’s role and getting to know the software people. Maemo folks. After being bounced a bit through Nokia’s vast email ecosystem I was forwarded to Florian Rommel, who was porting a few Linux utilities including OpenVPN. I helped test the tool in order to satisfy my auditing project and in the process I was hooked. It wasn’t just about the one project anymore: it was about what this device could do.
So again pushing the boundaries of what factory QA guys should be doing, I expanded my involvement. I joined the Maemo developer mailing list, discovered InternetTabletTalk.com, and started learning. The first thing I discovered was the tremendous community hunger for this sort of product. ITT rapidly attracted open device castaways from across the platform universe. The 770 was a true hacking device: open, raw, and completely untethered to any commercial app ecosystem. If you could grasp Maemo’s then-crude packaging tools, you could bring in virtually anything from the Linux world.
And people did.
Nokia as a company seemed unprepared to handle the growing geek love. As a consequence I found myself in a dual (often self-conflicting) role: factory QA engineer tasked with ensuring the quality of shipped devices and member of an expanding, passionate open source community desperately begging for more support from Nokia. My first personal goal was to stretch far and wide the channels of communication between company and customers. The second was to push the boundaries of purpose for Maemo devices. I came up short in both, and unwittingly annoyed many Nokia colleagues in the attempts, but I continue to believe we could have done better on both counts.
To Nokia’s credit, though, we didn’t stop at the 770. Maemo chugged on.
One day in late 2006 my manager informed me that the 770’s successor was coming and I would need to immediately shift my focus to that. However, its existence was also to be kept a secret prior to launch. That fact alone piqued my interest… along with a hint that the next device would be a game-changer.
When I was handed 10 prototype N800s I immediately fell in love. Yes, even with that “retro radio” cover. Words cannot convey the emotions that swept through me as I handled these things. A computer I can hold in my hand. It was a geek dream come true!
By this time Florian had the VPN app working well, and I was using it on the N800 to finish developing my web-based auditing solution as well as test some secret Nokia productivity tools. I can say now that this involved ways of bringing some degree of office functionality to the device, mainly to manage email, view spreadsheets, etc. Amazingly, it worked well enough that some time later when I stupidly left my Thinkpad behind on a trip to Helsinki, a testbed N800 handled about 80% of my Nokia business needs. Yeah, I had to squint to read email, but the damn thing worked!
Well, mostly. I discovered prior to launch that some of the rotating web cameras were improperly flipping images. In my opinion the mechanism was too complicated and highly prone to failure. The hardware team was sure that I was seeing an anomaly that wouldn’t manifest in production devices. That turned out to be wishful thinking, as I’ll explain in a bit.
We found out late in the game that the N800 touchscreen was sensitive to severe sunlight. So I was given another 10 devices to test in an oven. Yes, an oven. After baking them for a bit, half of them developed warped screens. The reflective backing was too thin for extreme heat. Fine for near-polar Finland, not so much for Texas. So the thickness was doubled.
Eventually the time came to prepare for product launch, which would be at the 2007 Consumer Electronics Show (CES). I had the privilege/burden of having final responsibility for device quality. It was my butt in a sling if we didn’t get 200 pristine N800s to CES for review purposes and another 5,500 right behind that for actual product launch weeks later. Scary!
Many would see that as work, and really hard work at that. And it was: I labored for two shifts (on salary) almost every day for two weeks to make sure our auditors understood the product– what to pass, what to fail. Unfortunately, during that heightened inspection phase we discovered that the webcam issue not only persisted, it was worse than originally feared. Eventually I ordered a 100% screen of devices to make sure not one single bad one escaped the factory. I won’t share more details here, but it was bad enough that the hardware guys flew immediately from Oulu and headed directly to the Alliance factory to join our inspection efforts. They were shocked at the dropout rate… but talk about dedication! Ultimately we handled the problem and very few faulty devices escaped into the wild. At least from my factory.
But the N800 still experienced high return rates. Why? Not so much due to devices issues per se, but rather, the bewildering corporate philosophy behind them.
Nokia’s Maemo efforts, even during the N800’s development, were more R&D than full-scale development. The team and its resources were dwarfed by the bread-and-butter Symbian programs. In a sense that was good: some of the best innovation comes from resource-starved teams. What was missing was the next-step function, where a well-resourced team or set of teams renders that innovation into must-have products.
Somewhere up high in Nokia though there was a decision made that we would proceed on this plodding Step X of 5 development process. The hacker-beloved 770 was Step 1. The N800 Step 2. Very early on as the last-mile quality engineer I believed that we were on the wrong path. Instead of going straight from hackers to consumers, I thought it made sense to divert the journey through industrial and point-of-sale applications. These demanding environments would help us shake out the bugs that would fatally damage public perception of the consumer devices. After all, Linux was still struggling for widespread desktop adoption and the mobile space was less forgiving. We had some platform refining to do.
Consumers found the N800 to look more polished than it actually was. Things like a shortage of applications out of the box, poor online video performance and a failure to fully support the webcam (one of its best hardware features) soured consumer attitudes. They bought so many N800s that we were continually ramping up production to meet unexpected demand– and returned so many out of frustration that our Care personnel couldn’t handle them.
It was an awkward, ironic, exasperating time.
Meanwhile Nokia was being pushed by a rapidly-swelling and excited developer/enthusiast community. Most of these hacker types didn’t care about the faults consumers found with Maemo devices. In fact, many of them wanted to fix things. But too many aspects were out of their hands. Certain critical software bits were kept out of the public API, and to an extent understandably so: some were not just drivers of hardware features but critical drivers of Nokia profit as well.
And there were more hardware issues, such as the N900‘s tendency to catastrophically lose its micro USB connector– often with no abuse by the user (it turned out that some connectors had resisted solder adhesion, possibly due to poor cleaning and/or tinning). Again, industrious members of the community stepped up and began finding ways to mitigate and even correct the failures. After initial resistance Nokia finally acknowledged the seriousness and offered to repair or replace broken devices. But by then perception of the platform had taken another hit.
What I really want to focus on is the can-do spirit of the Maemo community. I have never experienced anything like it, and that’s speaking as a current and former member of many similar organizations. This is a collection of diehards who even now, as the remnants of the Maemo/MeeGo team are summarily dispatched from Nokia, are fighting to find a way to keep carrying the Maemo torch. I disagree with some of the suggestions but I admire the intent!
Speaking of MeeGo…
The news of MeeGo stunned the Maemo community. We didn’t know how to process it at first. Finally many of us came to understand that this new joint venture offered the most practical way at the time of getting a (mostly) open mobile platform into the mainstream. I don’t want to hash over the downsides; that’s public knowledge now and I don’t see any point in belaboring it. The MeeGo torch has now been passed to Tizen and we’ll have to see how that one plays out.
The best thing to come out of the MeeGo time was the release of two devices: the N950 (mainly to developers, and I was proud to participate) and the incredible Nokia N9. What amazes and frustrates me at the same time is that the excitement and interest around the N9 rival anything Apple has released… yet Nokia elected to dead-end the product line. This is something else I won’t spend much time on other to say that my black N9 is device of choice, despite an annoyingly fast-draining sealed battery, and I don’t think I’ll ever quite understand its orphaning.
Admittedly, MeeGo Harmattan was awkward naming for a platform, and the name alone underscored the complexities in partnering with another strong entity. Nokia and Intel ended up with different visions, to put the outcome too simply… and when the two parted ways Nokia elected to employ an internal project name as branding on its version of MeeGo. Clumsy branding didn’t kill the N9’s prospects however; Stephen Elop made it clear he saw no room for this product line in the new Windows Phone scenario and so what is arguably the best smartphone to be launched by anyone ended up with a half-hearted rollout in a somewhat puzzling list of countries.
I’m just glad to have one (thanks, Nokia Developer program!).
And of course there’s Qt
In 2008 Nokia purchased Trolltech for the Qt development solution. At first I was ignorant of what this meant. What the heck was this “Qt”?
Over time the purpose became clear, and I saw the beauty. Qt would provide a clear, consistent development experience to be deployed across the high and mid ranges of Nokia devices. Eventually there was speculation that it would even come down to the lower S40 strata. Nokia was finally trying to truly platform.
I can only imagine the internal political tension that resulted. I had moved into global logistics at the time of the Trolltech acquisition and was approaching Maemo more from a community outsider’s standpoint. The internal struggle between product teams and divisions is not something I want to dwell on anyway. I think we can all look at the changes Nokia was going through, what the industry was going through, and understand that there was a legacy getting in the way of much-needed changes. Qt was a disruptive force, and while it offered great things at a high platform level it also meant severe change for the pressured rank and file. The N9 shows the Qt promise realized, but we are also all aware that it was not as encumbered by legacy as were Symbian products.
There was recently a tweet from someone I won’t name mocking Qt’s prospects at this point. Such foolish mutterings can only come from those unaware that Qt’s future isn’t completely intertwined with Nokia’s. It’s been used to develop such amazing tools as Google Earth and AutoDesk’s Maya, and regardless of what Nokia does with the development platform I don’t see it dying. Let’s be patient and see what happens here. Maybe RIM won’t be the only mobile company recognizing its value.
I believe Qt will prosper.
The cold hard facts
Reality can be a bitter pill to swallow. When I was more involved with Maemo, especially as a Nokia employee, I dared hope that the platform had a chance. That it would succeed on merit alone, if by virtue of nothing else. That the massive horde of enthusiasts working toward its success would have their voices heard.
There was technically nothing stopping Maemo. Its end was wrought by politics, pettiness and short-term narrow thinking. I can’t entirely blame Nokia though– Maemo was a victim of a System, one that feeds off of the manipulated emotions of unwitting consumers. There were pockets across the planet that initially resisted this System, but it has been stoked by politicians and industry leaders into something that will only allow a few, merciless players to survive in the mobile space.
I hate to be so cynical, but reality is what it is. Still, I think something like Maemo can have a future, at least in certain niches. Will Nokia continue to clutch the Maemo DNA close and keep it closed, or dare we hope that it could be spun off into a separate venture? I’m too numb at this point to speculate further.
I would like to close on a more positive note by thanking the Maemo and MeeGo contributors, who are all too numerous to mention. But I do want to zero in on a couple.
First, Quim Gil, who tirelessly and tolerantly performed the challenging role of Community Lightning Rod for several years, through many changes and challenges. To say I have learned a great deal from his example is an understatement. If Nokia makes the mistake of letting him go, HIRE HIM. Quim is committed to making things happen and there are many companies who need his passion and focus.
Above all though I want to dedicate this article to my gone-but-not-forgotten friend Gary “lcuk” Birkett, who truly epitomized what it meant to be a Maemo contributor. Gary lived on the technical cutting edge and dared us to join him there. He consistently challenged Nokia and often surprised us all with what he made the devices do. Gary was also much-beloved by the community, as evidenced in an immensely popular farewell thread at maemo.org. I still miss you Gary: your enthusiasm, your persistence, your ability to engage the crowd. If we had more like you, especially in the upper echelons of Nokia, I daresay I would be writing a completely different article today, and certainly not an epitaph to a grand but failed experiment.
My possibly-naïve hope is that Maemo doesn’t really pass away, but finds a legacy in inspiring places like OpenTablets.org and CloudBerryTec. Just as I can’t bring myself to delete Gary Birkett from my contacts, I can’t accept the complete demise of Maemo. May it live on in some form or fashion.
Your comments are very welcome; please feel free to share your own Maemo thoughts and experiences. Thanks everyone.
Latest posts by Randall "texrat" Arnold (see all)
- Math is Hard, part 1 - 16 March 2014
- No Stranger to Fiction - 3 March 2014
- Down-Voting Disqus, or Why Binary Feedback Mechanisms Matter - 1 March 2014