Previously we chatted about how we got into this PC mess, which started with the concept of something that was a lot like a user-focused mainframe and now needs to evolve into something that makes better use of hosted services. But this device needs to be transitional because we can't yet depend on wireless services, so full reliance on the Web for what is an increasingly a mobile experience just isn't practical for most of us.
Next Generation the Appliance PC
A year or so ago I spoke at the Linux desktop summit on what Linux needed to do to compete. Geoffrey Moore, author of Crossing the Chasm and a Linux user, spoke before me and we must have been channeling each other because we basically said the same thing (had I known that I would have done something else). What the market is signaling is the need for something different, and each vendor/platform has a potential roadmap which takes it down this path.
Apple's OS is based on the same concepts that created Windows in the '80s. It tries for a more appliance-like result, but the argument of whether an OS should be hardware-independent was lost by Apple and the market has clearly shown that Bill Gates was right. Recall, for much of the '80s, Apple, not Microsoft, was the larger and more powerful company. Its recent use of emulation technology and aggressive development of virtualization may indicate an alternative approach.
Apple's is the most appliance-like of the traditional PC platforms, and if it can leverage the Google relationship (which hasn't done much for them to date) Apple might have something interesting coming. Leopard has been delayed, evidently, until October for Windows Vista compatibility changes. The new Apple TV is an appliance and may also indicate a direction for future products from the company. The iPhone, while consumer-only, is also an indicator of this appliance-like move, and you can easily imagine how a device like this, once mature, could replace the laptop that many would rather not carry.
Linux, whose primary values are UNIX compatibility and the ability for users to access source code, actually goes off in a vector corporate users and CIOs don't want for the desktop, where even a little variance can cost millions in additional support cost. It is interesting how few people simply don't understand desktop economics of how just a little difference can result in huge budget overruns.
But Linux is common in thin clients and is growing in cell phones. In areas where wireless broadband can be relied on, the embedded version of Linux, properly wrapped with on-line and off-line services and applications, could forecast the future.
Google is attacking Windows dependencies and has no apparent desire to do a desktop OS -- in fact, it is working to trivialize it. This is a game-changing strategy and, because it will be incredibly hard for Microsoft to counter (and Microsoft is not doing well against Google), it could be the most successful long-term. Microsoft changed the game by trivializing hardware and IBM lost its lead as a result.
If Google can trivialize the OS and desktop applications, the entire desktop becomes commoditized. Microsoft might still be the dominant OS supplier, but PC cycle times would extend and appliances running embedded OS would combine to pull the cash away from the company and its OEM partners, unless they too switched models. This adds risk to the Apple partnership because if Google's approach works it will do as much damage to Apple as it does to Microsoft, basically obsolescing or commoditizing the existing PC market.
IBM IBM has a really interesting plan to retake the desktop, and it feels in many ways like a return to the mainframe. Its plan is to trivialize the OS and the PC and create a ground-up redesign of the entire environment based largely on WebSphere and, on paper, it is the most aggressive of the new offerings. However, in many ways, it may repeat the errors IBM has made about the desktop in the past as it, like Apple's offering, appears to single vendor-centric, and the initial risk that any company would take putting this in appears daunting.
From an IT perspective, it could be incredibly attractive, but from the perspective of users and operations it could be a nightmare to sell internally, and the end result may be more like Sun's failed SunRay 1 initiative than DOS or Windows. It is gutsy, though, and if it were successful, it would put IBM back in the center of things.
Of all of the PC OEMs, HP has the most options in development or service. It has bladed desktops, workstations and servers; it has thin clients; it has hand-held computers; and it has full PC desktop and laptop lines. Its leading advantage is if a company wants a next-generation platform based on Linux or Windows, HP can deliver it today and, with its PC line -- given the division is heavily staffed with ex-Apple people -- it can even provide the most Apple-like products in their consumer lines.
HP clearly sees the future coming, but is actually positioning for it rather than waiting for that future wave to break over their heads. What is particularly interesting is that even in a tight market, HP continues to push its R&D spend and, even with this added cost, is able to maintain market leadership status. Now that is applied R&D. HTC
Of the smartphone vendors, HTC stands out as the only one which is doing a PC, and it is starting at the smallest of classes the UMPC. With a wide selection of smart phones, many with keyboards and all based on embedded Windows, it could be the company to redefine mobile computing this decade. The experience they provide is not generic Windows because they push the envelope with regard to size and even industrial design. A company very early in this cycle, but one growing at a rate that is almost breathtaking to watch and where most of the products are both wireless and based on Windows.
The Microsoft Mobile platforms, embedded products, and Xbox all represent the idea of a thin client device that can be used ever more broadly in a world with rapidly increasing network bandwidth. The Xbox is particularly interesting because is actually provides a relatively, at least when compared to a typical thin client, robust platform, but clearly not one targeted at the general PC user yet, though it is trending that way in the home. Microsoft goes thin client in its set top box offerings as well, and its embedded product competes with Linux in traditional thin clients and point-of-sale terminals where it has actually done surprisingly well.
Strangely enough, embedded Windows is probably very close to what the market is looking for in a next-generation PC product in terms of foundation platform for the client side of the solution, but to get there Microsoft would have to do something very difficult politically and let this product grow to its true potential. Given the huge disparity in revenue between this and traditional Windows, that kind of move would be nearly impossible to make.
Next Generation in Pre-Birth
I actually believe that what will likely happen is someone will take embedded Windows and build the next generation PC platform by starting where Microsoft left off, channeling Apple for the UI, and connecting through a set of applications that will work both on-line and off during the transition to always-on broadband wireless.
I think there is room for a vendor to do what Apple did with BSD UNIX and create something based on Microsoft technology which will interoperate with the world of today and better embrace the world of tomorrow, but we are still waiting for that solution to appear and I doubt it is within the three-year horizon yet. But the demand is so clear that I believe it will appear in the next few years.
The ending platform will likely be very portable, but scale nicely on the desktop to use both local and remote resources to provide the performance you need, but no more, when you need it. With good battery life and a strong wired and wireless connection to the Internet, it is always connected and it likely will be the primary device you use while making cellular calls as well (even if it is just to ensure you are getting the lowest cost call routing).
The direction is clear, but will Apple and Linux stop watching Microsoft long enough to make a run for the goal, or will Microsoft, realizing it is falling behind, suddenly toss off a lot of its old baggage and embrace this future first? In the end, it will likely be a third party using a mix of technologies, like HP, who will make the first big move, but there is change in the wind and, when we look back in another 10 years, I think we will be surprised at just how obvious this change looks in hindsight.
As I finished this I realized there is a third part to this series. One in which I step into the future and look back and the events that led to what resulted. I found by doing this none of the OSs were important anymore and that we are on the forefront of a massive change. Look for part 3 in a few days.