Monday, October 6, 2008

The top five reasons why Windows Vista failed

Microsoft gave computer makers a six-month extension for offering Windows XP on newly-shipped PCs. While this doesn’t impact enterprise IT — because volume licensing agreements will allow IT to keep installing Windows XP for many years to come — the move is another symbolic nail in Vista’s coffin.
The public reputation of Windows Vista is in shambles, as Microsoft itself tacitly acknowledged in its Mojave ad campaign.
IT departments are largely ignoring Vista. In June (18 months after Vista’s launch), Forrester Research reported that just 8.8% of enterprise PCs worldwide were running Vista. Meanwhile, Microsoft appears to have put Windows 7 on an accelerated schedule that could see it released in 2010. That will provide IT departments with all the justification they need to simply skip Vista and wait to eventually standardize on Windows 7 as the next OS for business.
So how did Vista get left holding the bag? Let’s look at the five most important reasons why Vista failed.

5. Apple successfully demonized Vista

Apple’s clever I’m a Mac ads have successfully driven home the perception that Windows Vista is buggy, boring, and difficult to use. After taking two years of merciless pummeling from Apple, Microsoft recently responded with it’s I’m a PC campaign in order to defend the honor of Windows. This will likely restore some mojo to the PC and Windows brands overall, but it’s too late to save Vista’s perception as a dud.

4. Windows XP is too entrenched

In 2001, when Windows XP was released, there were about 600 million computers in use worldwide. Over 80% of them were running Windows but it was split between two code bases: Windows 95/98 (65%) and Windows NT/2000 (26%), according to IDC. One of the big goals of Windows XP was to unite the Windows 9x and Windows NT code bases, and it eventually accomplished that.
In 2008, there are now over 1.1 billion PCs in use worldwide and over 70% of them are running Windows XP. That means almost 800 million computers are running XP, which makes it the most widely installed operating system of all time. That’s a lot of inertia to overcome, especially for IT departments that have consolidated their deployments and applications around Windows XP.
And, believe it or not, Windows XP could actually increase its market share over the next couple years. How? Low-cost netbooks and nettops are going to be flooding the market. While these inexpensive machines are powerful enough to provide a solid Internet experience for most users, they don’t have enough resources to run Windows Vista, so they all run either Windows XP or Linux. Intel expects this market to explode in the years ahead. (For more on netbooks and nettops, see this fact sheet and this presentation — both are PDFs from Intel.)

3. Vista is too slow

For years Microsoft has been criticized by developers and IT professionals for “software bloat” — adding so many changes and features to its programs that the code gets huge and unwieldy. However, this never seemed to have enough of an effect to impact software sales. With Windows Vista, software bloat appears to have finally caught up with Microsoft.
Vista has over 50 million lines of code. XP had 35 million when it was released, and since then it has grown to about 40 million. This software bloat has had the effect of slowing down Windows Vista, especially when it’s running on anything but the latest and fastest hardware. Even then, the latest version of Windows XP soundly outperforms the latest version of Microsoft Vista. No one wants to use a new computer that is slower than their old one.

2. There wasn’t supposed to be a Vista

It’s easy to forget that when Microsoft launched Windows XP it was actually trying to change its OS business model to move away from shrink-wrapped software and convert customers to software subscribers. That’s why it abandoned the naming convention of Windows 95, Windows 98, and Windows 2000, and instead chose Windows XP.
The XP stood for “experience” and was part of Microsoft’s .NET Web services strategy at the time. The master plan was to get users and businesses to pay a yearly subscription fee for the Windows experience — XP would essentially be the on-going product name but would include all software upgrades and updates, as long as you paid for your subscription. Of course, it would disable Windows on your PC if you didn’t pay. That’s why product activation was coupled with Windows XP.
Microsoft released Windows XP and Office XP simultaneously in 2001 and both included product activation and the plan to eventually migrate to subscription products. However, by the end of 2001 Microsoft had already abandoned the subscription concept with Office, and quickly returned to the shrink-wrapped business model and the old product development model with both products.
The idea of doing incremental releases and upgrades of its software — rather than a major shrink-wrapped release every 3-5 years — was a good concept. Microsoft just couldn’t figure out how to make the business model work, but instead of figuring out how to get it right, it took the easy route and went back to an old model that was simply not very well suited to the economic and technical realities of today’s IT world.

1. It broke too much stuff

One of the big reasons that Windows XP caught on was because it had the hardware, software, and driver compatibility of the Windows 9x line plus the stability and industrial strength of the Windows NT line. The compatibility issue was huge. Having a single, highly-compatible Windows platform simplified the computing experience for users, IT departments, and software and hardware vendors.
Microsoft either forgot or disregarded that fact when it released Windows Vista, because, despite a long beta period, a lot of existing software and hardware were not compatible with Vista when it was released in January 2007. Since many important programs and peripherals were unusable in Vista, that made it impossible for a lot of IT departments to adopt it. Many of the incompatibilities were the result of tighter security.
After Windows was targeted by a nasty string of viruses, worms, and malware in the early 2000s, Microsoft embarked on the Trustworthy Computing initiative to make its products more secure. One of the results was Windows XP Service Pack 2 (SP2), which won over IT and paved the way for XP to become the world’s mostly widely deployed OS.
The other big piece of Trustworthy Computing was the even-further-locked-down version of Windows that Microsoft released in Vista. This was definitely the most secure OS that Microsoft had ever released but the price was user-hostile features such as UAC, a far more complicated set of security prompts that accompanied many basic tasks, and a host of software incompatibility issues. In order words, Vista broke a lot of the things that users were used to doing in XP.

Bottom line

There are some who argue that Vista is actually more widely adopted than XP was at this stage after its release, and that it’s highly likely that Vista will eventually replace XP in the enterprise. I don’t agree. With XP, there were clear motivations to migrate: bring Windows 9x machines to a more stable and secure OS and bring Windows NT/2000 machines to an OS with much better hardware and software compatibility. And, you also had the advantage of consolidating all of those machines on a single OS in order to simplify support.
With Vista, there are simply no major incentives for IT to use it over XP. Security isn’t even that big of an issue because XP SP2 (and above) are solid and most IT departments have it locked down quite well. As I wrote in the article Prediction: Microsoft will leapfrog Vista, release Windows 7 early, and change its OS business, Microsoft needs to abandon the strategy of releasing a new OS every 3-5 years and simply stick with a single version of Windows and release updates, patches, and new features on a regular basis. Most IT departments are essentially already on a subscription model with Microsoft so the business strategy is already in place there.
As far as the subscription model goes for small businesses and consumers, instead of disabling Windows on a user’s PC if they don’t renew their subscription, just don’t allow that machine to get any more updates if they don’t renew. Microsoft could also work with OEMs to sell something like a three-year subscription to Windows with every a new PC. Then users would have the choice of renewing on their own after that

Intel and Yahoo! to Launch “Widget Channel”

Intel and Yahoo! last month announced plans to launch the “Widget Channel,” an application framework for TV and related consumer electronics devices that are based on the Intel Architecture. According to the companies, the Widget Channel will allow viewers to access Internet applications that have been designed for TV, while watching programming. It will be powered by the Yahoo! Widget Engine, a fifth-generation applications platform that will support a line-up of “TV Widgets”–small Internet applications that can be accessed by the remote control and that are intended to complement and enhance the TV viewing experience, by providing informational and entertainment content, and community features. The companies say that the Widget Channel will also allow developers to use Javascript, XML, HTML and Adobe Flash technology to create their own TV applications, thus “extending the power and compatibility of PC application developer programs” to TV and related CE devices. In addition to supporting the Yahoo! Widget Engine, Yahoo! says it will provide consumers with Yahoo!-branded TV Widgets that will be based on the various services it offers on the Internet.
Intel and Yahoo! say that the Widget Channel’s mini-apps will, among other things, enable consumers to access Internet videos, track stocks or sports teams, interact with their friends, access news reports, find out additional information about the programs they are watching, and share content with friends and family. The companies say that they will be easily personalizable, because they will be based on Internet services such as Yahoo! Finance, Yahoo! Sports, Blockbuster and eBay that are designed to be customized by end-users. “TV will fundamentally change how we talk about, imagine and experience the Internet,” Eric Kim, general manager of Intel’s Digital Home Group, said in a prepared statement. “No longer just a passive experience unless the viewer wants it that way, Intel and Yahoo! are proposing a way where the TV and Internet are as interactive, and seamless, as possible. Our close work has produced an exciting application framework upon which the industry can collaborate, innovate and differentiate. This effort is one of what we believe will be many exciting new ways to bring the Internet to the TV, and it really shows the potential of what consumers can look forward to.” Added Marco Boerries, EVP of Yahoo!’s Connected Life arm: “On the PC and mobile devices, Yahoo! is a leading starting point for millions of consumers around the world. Yahoo! aims to extend this leadership to the emerging world of Internet-connected TV, which we call the ‘Cinematic Internet.’ By partnering with leaders like Intel, we plan to combine the Internet benefits of open user choice, community, and personalization with the performance and scale embodied in the Intel Architecture to transform traditional TV into something bigger, better and more exciting than ever before. By using the popular Yahoo! Widget Engine to power the Widget Channel, we intend to provide an opportunity for all developers and publishers to create new experiences that can reach millions of TV viewers globally. Yahoo! plans to enable the Cinematic Internet ecosystem, which will benefit consumers, device makers, advertisers and publishers.”
According to Intel and Yahoo!, the Widget Channel will be powered by a set of platform technologies that include, in addition to the Yahoo! Widget Engine, core libraries designed to exploit the capabilities of the Intel Architecture. The Widget Channel framework will use a number of established Internet technologies to significantly lower the barrier-to-entry for developing applications optimized for the TV, the companies say. In order to spur development of widgets for the Widget Channel, Intel and Yahoo! plan to make a development kit available to manufacturers of TV and other CE devices, as well as to advertisers and content providers. The Widget Channel will also include a Widget Gallery, the companies say, to which developers will be able to publish their TV Widgets across multiple TV and related CE devices, and through which consumers will be able to browse and choose the widgets they would like to use.
The companies say that a number of partners are planning to develop and deploy TV Widgets, including Blockbuster, CBS Interactive, CinemaNow, Cinequest, Comcast (see below), Disney-ABC Television Group, eBay, GE, Group M, Joost, MTV, Samsung Electronics, Schematic, Showtime, Toshiba and Twitter. In addition, the companies say they are working with “industry members” to promote the development of open standards that would help grow the TV Widget ecosystem: as part of these efforts, they are sharing an early version of a development kit for the Widget Channel with a “selected” group of developers.
The Widget Channel software framework is designed to run on Internet-connected TV’s, cable set-top boxes, optical media players and other consumer electronics devices that are powered by a newly launched family of system-on-a-chip (SoC) media processors based on the Intel Architecture. The first of these Intel Architecture-based SoC’s is the Intel Media Processor CE 3100, which is billed by the company as a highly integrated chip that includes a high-performance Intel Architecture core and other functional I/O blocks to enable HD video decode and viewing, home-theater-quality audio, 3D graphics and the fusion of the Internet and TV experiences. Intel says it plans to release an Intel Media Processor CE 3100-based hardware development system, called the “Innovation Platform,” which will provide the initial development and validation environment for developers of widgets for the Widget Channel.
Comcast, the US’s largest cable MSO, announced last month that it is working with Intel to bring IP-based applications to TV, using the new Widget Channel application framework. The companies say that they expect to begin integration testing of the Widget Channel framework in the first half of 2009 on Comcast’s EPG, using tru2way/OCAP technology. “The Widget Channel enables interactive applications and tru2way technology has opened the door for these types of innovations to work in the cable industry,” Comcast Cable CTO, Tony Werner, said in a prepared statement. “We’re looking forward to working with Intel as we continue to bring our customers new features and services that further enhance their viewing experiences.” Comcast is billing its plans to implement the Widget Channel on tru2way as an “important milestone in the evolution of the tru2way ecosystem,” and says that the Widget Channel framework will complement tru2way technology and broaden the interactive TV developer community. (Note: for more on the Widget Channel, see article in this issue, and listen to [itvt]’s radio interview with Patrick Barry, VP of connected TV at Yahoo!’s Connected Life division, in Issue 7.99.)

High-speed RAM can damage Nehalem i7 processors

According to the Inquirer, Intel is advising motherboard and RAM vendors about their new X58+Core i7 combo and to abide to a strict memory voltage limit of 1.65 volts limit. If the advice is not followed, they have warned that the CPU will get fried.
This came to light first when admins on the XFastest forums < http://www.xfastest.com/redirect.php?tid=14549&goto=newpost > posted several photos of the ASUS P6T Deluxe motherboard with a big sticker over the DIMM slot warning that anything more than 1.65 V will destroy the CPU. ASUS has admitted that the story is true.
This doesn't seem to be really a problem since DDR3 JEDEC specs state that it should operate at 1.5 volts. But the problem may arise for overclockers with performance memory since many memory vendors offer faster RAM kits that operate at higher voltage levels. As an example, OCZ Reaper PC3-14400 operates at 1..9V, Mushkin's XP Series uses 1.9-1.95V while Corsair's Dominator high-end takes you all the way up to 2.1V.
While Mushkin will re-design their kit specifically to suit the X58/Core i7 combo and should be out sometime next month, the other memory vendors have not commented anything beyond stating that their kits are pending certification.
Intel is yet to explain why the memory voltage would damage the CPU, though we speculate that this could have something to do with the on-die (integrated) memory controller of the Nehalem. Though it is yet to hit the Indian market, people importing the Core i7 should make sure that they get a compatible RAM kit or they will have to underclock it