Embedded magazine has produced an excellent multi-part article on the basics of programming embedded processors, all excerpts from Wayne Wolf’s book Computers As Components: Principles of Embedded Computing Systems Design . This is exactly the kind of educational material for which Embedded is known, and which is badly needed. Also, don’t miss their article on embedded programming in C, which is evidently becoming a lost art (not in my sphere!).

The Linux Foundation has opened up its Call for Participation (CFP) for the annual Collaboration Summit. The Summit will be held this year in San Francisco and is once again co-located with the Embedded Linux Conference (whose CFP just ended). This dual event is The Big Event for embedded Linux developers, so dust off your presentations and start begging for travel budget!

Today I’m going to talk about embedded design, the user experience, and business ramifications.

My friend Zonker, who is always insightful, just said this in his blog:

It’s amazing what people consider an “effort” now compared to 5, 10, or 20 years ago. […] People have little to no tolerance for what they consider inconvenient. Even if that inconvenience is, realistically, very minimal. It can cost thousands or even millions of users for a platform if people perceive it as inconvenient.

This is absolutely true, and doubly—triply—so in embedded systems design. User experience (UX in the industry) is king, more now than ever. As an example, let’s look at something we all now take for granted: the lowly telephone.

Believe it or not, 200 years ago, smoke signals and jungle drumming aside, it was not possible to contact someone electronically, electrically, or any other way besides (1) being there in person, or (2) writing them a letter. In 1844, 166 years ago, Samuel Morse demonstrated the first commercially viable telegraph system, building on a design Joseph Henry demonstrated in 1830. The UI was crude—a simple switch—and involved trained operators at both ends who understood Morse code, but it was a vast improvement over the physical transport of either people or parcels. Messages that previously took days or weeks to arrive now took minutes. Interestingly, the user experience for the average person was not much different from one they were already used to—the mail. Instead of dispatching a handwritten letter, people would dispatch a “telegram” via human messenger. This was the primary method of non-physical communication for the average person from the 1840s nearly until the turn of the 20th century. (Note: one could still send a telegram via Western Union in the United States until 2006.)

Telephones enabled individual users to actually talk to each other directly without need for an intermediary, vastly improving the user’s experience. Granted, most lines were “party lines” in which everyone could listen to the conversation, rendering secure audio communication an impossibility. (I can verify that party lines were still a reality in my ancestral village in northern Iowa until the early 1980s.) Still, it sure beat the pony express. Individual lines slowly became a reality through the mid-1900s, to the delight of teenagers everywhere.

Eventually came the mobile phone, now with Jetsons-like features, that fits in my pocket. The cost of sending a message or talking to someone far away is a very, very, very small fraction of what it ever has been before, throughout the history of electric communication briefly described above. The device itself can also play games, track your every movement and give me directions to get where I want to go, play music, and operate toy helicopters.

Yet the major difference between a no-holds-barred winner in the device marketplace and an also-ran (sorry, Palm, but it’s true) is its user experience. UX (user experience) (learn more here) is a far more important differentiator in a device than, say, performance, screen size, or available storage, up to the point where any of those incur costs in UX. And UX isn’t just the UI on the device, it is the ecosystem around it, as Apple uniquely proved (again) with the iPhone. In most ways, the iPhone didn’t provide anything that HP, Palm, Sony, and others hadn’t already provided before them. What they did was provide a solid UI and an ecosystem that made using the device seamless and fun.

The magic in UX is setting expectations and then meeting them. We must treat the users themselves as kings, otherwise the most clever devices fall by the wayside. Good ideas don’t win markets—treating the user like a valued customer is what does it.

Tomorrow, January 15, 2010, is the final day for submissions of presentations, BoFs, and demos to the Embedded Linux Conference. This annual conference, sponsored by CELF, is the premier event for embedded Linux experts in the western hemisphere. I have spoken at CELF for the past few years and I can say that it is a very friendly and collegial event, both for experts and for those new to embedded Linux, and that the CELF folks are awesome. I learned more in three days at ELC than I did in half a year of computer science classes in college, and it is all up-to-the-minute current.

This year the conference is being held in the Kabuki Hotel in San Francisco, as it was last year, and is again co-located with the Linux Foundation Collaboration Summit. Collab Summit is normally an invitation-only event, but ELC participants are given invitations.

Also keep an eye out for ELC Europe in the fall, and check out the calendar of open-source conferences.

The ever-vigilant Andreas Constantinou just posted an article titled Mobile Megatrends 2010 on the VisionMobile blog. Anyone involved with mobile computing, especially in the exploding embedded Linux space, should read this and follow Andreas. Seriously.

Most of what he covers is related to business issues, but I would posit that business issues are the number-one important issue in the mobile computing space right now. Anyone disagree?

Cavium Networks is announcing today that the acquisition of MontaVista is complete. MontaVista Software, Inc. is now MontaVista Software, LLC.

When we all found out in a company meeting on November 10 that MontaVista was being acquired, no one really knew what to think. We have been doing business with Cavium for years, providing operating systems for its industry-leading networking equipment, but it was a surprise to find that they wanted to own us. It was not readily apparent in those first few minutes why we would be of interest to a company whose rising star shines on hardware rather than software. However, over the past month we have all learned a bit about each other, and some of the big win-win scenarios I initially described in my blog post on the subject are absolutely true. (Yes, I consider myself to be prescient!)

Cavium is in a very unique position to explode onto a hungry marketplace with solutions in nearly every quarter for embedded and semi-embedded Linux. By embedded I mean hard embedded devices and appliances, automotive and aerospace control systems, networking equipment, and so on. These are the devices on which both Cavium and MontaVista have focused until now, and they will continue to be a very strong focus going forward, especially in terms of MontaVista’s Carrier Grade Linux offering.

By “semi-embedded” I mean devices that traditionally have used a hard embedded and/or real-time operating system (RTOS) but are now being recast in the marketplace as more general-purpose devices, including smartphones, netbooks, set-top devices, automotive “infotainment” systems (IVI), and the universe of new possibilities opening up with the proliferation of 3G and 4G cellular internet service. Mashups that were once the realm of science fiction are now commonplace. The combination of GPS and broadband internet means I can ask my cellphone what the weather is like and it will tell me, right where I am standing, with very few clues as to what I want. These systems are much more complex than traditional embedded devices, and must take input from a wide variety of sources, including a user whose expectations are being continually reset.

This is a market where all companies are newcomers, and both Cavium and MontaVista are well-positioned to dive in. There are other companies in similar positions, two in particular—not to name any names—but one is now tied to a massive single-platform system that is not ideally suited to hard- or semi-embedded, and the other is highly focused on very specific embedded applications. No one else in the embedded Linux sector, or frankly in the entire embedded marketplace, has the flexibility that the Cavium/MontaVista merger brings to the table. The combined force of a highly flexible hardware and software provider make this a very exciting place to be at this point in history.

I just updated my definitive list of open-source conferences for 2010. This is a list I maintain for all conferences that are important to open-source projects, Linux in particular, with an emphasis toward the embedded space. If you know of a conference I missed, please let me know in comments and I’ll be take a look.

I have added two new goodies this year. First is a note for calls for participation (CFPs) for conferences that accept presentations, as there is nothing as invigorating as standing up to talk in front of twenty or thirty people who are smarter than you are. Second, I also made a Google calendar of this list of conferences—please feel free to add it to your calendar, and edit it or send me edits. (If you can’t edit it for some reason, let me know.)

The hottest news conference-wise is for SCALE 8x, the Southern California Linux Expo in Los Angeles in February. The deadline for presentations was just extended to Dec. 24, so get them in! Also, CELF’s Embedded Linux Conference (April) is accepting presentations until January 15, and if you attend you also get to go to the Linux Collaboration Summit that is co-located at the beautiful Kabuki Hotel in San Francisco. And don’t forget Mobile World Congress in Barcelona (Feb), Embedded World in Nuremberg (March), and of course the Embedded Systems Conference in San Jose in April, in case you need t-shirts, pens, free magazines, or any of the really intense presentations on the roster, including Michio Kaku as the keynote presenter!

Looks like quite a year on the conference circuit.

I started writing this post back in May, but got sidetracked and never finished it, partly because it sounded like sour grapes from someone who has used a netbook and didn’t like the user experience. The thing is, I don’t use a netbook, never wanted to use one, and the few times I have seen them I didn’t like the user experience. That’s the real reason I didn’t finish the post—I didn’t want to comment on something I didn’t know well.

It looks like I’m not the only one. Jason Hiner of TechRepublic today announced the death of the netbook on his blog. While I think death may be a little premature to declare, I think doom is certain called for, despite Larry Dignan’s rebuttal on ZDnet. And I don’t want to leave out Laptop Magazine’s Mark Spoonauer, who also declares death and even provides a headstone.

The thing is, I really want to like netbooks. They would solve many problems in my life while introducing few others. I don’t mind small screens, can cope with small keyboards for a limited time, and I know precisely which apps require which hardware to run effectively. I even had a “netbook” once, a little Psion that fit my pocket, but I eventually abandoned it. What I don’t want is:

1. yet another device to feed—laptop and (non-smart) cell phone are plenty, thanks
2. a new algorithm for syncing local data, as not everything lives in the cloud (yet)(thankfully)
3. any device that solves 75% of a problem, yet requires 110% of my personal budget and replicates a device I already have (see #1)

Furthermore, I don’t think they are too expensive for what they are. Dropping prices will not sell more netbooks. Increasing performance will not sell more, either. Actually, I can’t think of a single thing that would cause me to augment my existing stable of computing power with yet another device, and that is the rub.

I am exactly the target market for netbooks, yet I don’t want one. That is why they will fail. There are a lot of people like me out there. We are already glutted with devices. Netbooks solve the same problems as notebooks, but at a higher price and with fewer options and lower performance. Smallness on its own is not a feature that grabs me unless it gets small enough that I can put it in my pocket.

All of this is based on my opinion. I am not presenting a very logical argument here because the market is not a logical beast. It is capricious and fickle, which is precisely why good ideas very often fail. To understand it, the logical thinker must shelve the logic and think like a fickle beast. The cost of failure is too high to depend on logic in a capricious world.

That being said, there are logical arguments, and Jason Hiner makes them quite well. The “that’s neat” response is not a long-term market driver. We encountered this at Transmeta back in 1999 trying to come up with a reference design for tablets—we called them “webpads”, and they bear a marked resemblance to today’s e-book readers, but they never took off (they were also far too expensive). For that matter, I think it remains to be seen whether e-book readers will take off, though as a device built to improve a specific user experience by using a people-friendly interface—e-ink in this case—they stand a good chance.

What I think might fit the sweet spot a little more closely is something like Dell’s new Latitude ON, which embeds a netbook-ish device inside a laptop. The tiny ARM-powered board shares keyboard, screen, and network connection and surfs and emails for 18 hours or more, and if I want more horsepower I can just push a button to boot or wake up the full laptop. That is a compromise I can live with. *

So, kudos to the netbook folks for trying. But what about MIDs? Are they in the same boat? I think that remains to be seen—the iTouch and Nokia N800 devices have certainly been popular among my geeky friends, but I have yet to see one in use that isn’t shadowed by the propeller on the user’s hat. It’s the “that’s neat” response again. I want to want one, but in reality I can’t justify it to my bank account or to my daily workflow. My immediate feeling is that the next generation of smartphones—especially the Motorola Droid—will make MIDs obsolete, but this will only happen in the US if they can break away from the usage models dominated by the carriers. The total cost of ownership is far, far too high. (See Droid and other smartphones compared by TCO.)

Here’s yet another angle: what is the logical difference between a MID that can read e-books, and an e-book reader that can browse the web? How about MID vs. smartphone? Which one would you buy, and why? Don’t be afraid to be illogical and capricious.

The main takeaway from all of this is that making a market in personal electronics is hard, and it is not getting easier or more obvious as technology improves. Making a market means building a product that provides a 100% solution for enough buyers to justify the cost of development, period. If that market is small or unsustainable, then it’s not a revolution, it’s a fad.

Code safely out there.

* Full disclosure: As it has now been officially announced, I can proudly say that MontaVista Montabello is running the Latitude ON, and that I helped work on the project. Actually I already said it, but now I can say it.

For all of my loyal fans, I will actually be talking at 2:30 CENTRAL time, not Pacific. That means 12:30pm PST, so get your popcorn early! Previous message included for clarity:

On Tuesday, November 17, 2:30pm CST, 12:30pm PST, I will be presenting a 5-minute lightning talk at TI’s ETech Days. The subject is booting MontaVista Linux 6 on the Beagle Board, companion to a recent white paper and pertinent to MontaVista’s very recent release of MVL6 for the Beagle Board and Zoom2 platforms from Texas Instruments. Come join us! And thanks to Jason Kridner for putting it together.

Also feel free to visit my personal blog, where I very recently discussed the Cavium Networks acquisition of MontaVista as well as Microsoft’s recent move to release part of .NET open-source.

On Tuesday, November 17, 2:30pm PST, I will be presenting a 5-minute lightning talk at TI’s eTech Days. The subject is booting MontaVista Linux 6 on the Beagle Board, companion to a recent white paper and pertinent to MontaVista’s very recent release of MVL6 for the Beagle Board and Zoom2 platforms from Texas Instruments. Come join us! And thanks to Jason Kridner for putting it together.

Also feel free to visit my personal blog, where I very recently discussed the Cavium Networks acquisition of MontaVista.