eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
The foremost companies of our industry this year look to be having trouble executing on plans to provide technologies promised in ballyhooed promotions.
But as customers and partners complain about changes and failed strategies, the response from corporate executives remains the same: Trust us!
That simple, dont-worry-be-happy message—so easy to give as well as to receive when things are going well—becomes less credible and more worrisome with each tweak of a product roadmap.
Both Intel and Microsoft provide leading examples of this dilemma since the companies have either canceled or pared back major technology directions this year.
For example, heres a rundown of some of Intels recent twists and turns:
- In January, at the Consumer Electronics Show in Las Vegas, Intel said it would produce LCOS (liquid crystal on silicon) processors for high-definition televisions in 2005. Analysts said the chip—as well as the backing of an industry leader such as Intel—would shake up the economics for this growing market. Intel said the chip would let vendors offer 50-inch HDTVs costing less than $2,000.
The company was expected to ship sample chips to partners in the summer, but nothing was released. In mid-August, Intel said it would delay the product.
The company instead was “evolving its development plans and wont bring our initial product to market this year,” a spokeswoman told ExtremeTech at the time. “We are heading down a path of developing technology that will enable clear product differentiation with improved picture quality.”
Well, that evolution came to a dead end this month as the company pulled the plug on the chip and involvement in the market.
- First slated for release this fall, a 4-GHz version of the Pentium 4 architecture in late July was reported delayed until early 2005. But Intel this month said it had canceled plans for the single-core processor.
The company now points to its advancement of performance, power-handling and production capabilities with forthcoming dual-core processors, due later in 2005 and 2006. The transition to multiple cores aims to address the growing problems of current leakage and power consumption found in the faster chips made using the companys smaller fabrication process.
Next Page: Putting on a game face.
Game Face
While putting a game face on the subject, this shift away from straightforward clock performance is mostly new to Intel, having already been tackled by competitors Advanced Micro Devices and IBM. The acceptance by customers and different market segments to this transition remains unclear, analysts said.
- In the summer, Intel admitted finding a “root flaw” in the chip sets supporting its 64-bit Nocona processor aimed at workstations and servers. The problem involves the integrity of the PCI Express data handling and will be fixed with an update by the end of the year, the company said.
- Intel has touted its support of various technologies into chip sets to partners and analysts. As a major backer of Wi-Fi and WiMAX wireless standards, Intel heavily promoted its “Grantsdale” chip set, which would let a laptop function as a Wi-Fi access point.
But this summer, the company admitted that a manufacturing flaw in the chip would prevent customers from gaining the advertised functionality. And Intel later stated it will never provide the support.
Still, a company spokesman described this move as a “business decision,” adding that system vendors “told us that they didnt need this feature at this point.”
These shifting roadmaps and mistakes have sapped developer and partner confidence in the company, which has been admitted by Intel executives.
At the Intel Developer Forum in early September, chief operating officer Paul Otellini apologized to a keynote audience, calling the problems “some fumbles.” He said future goals will be more closely aligned with capabilities.
However, from the continuing shifts in plans since that admission, and the companys reported hazy approach to shipment forecasts, it looks as if little has changed behind the Intel curtain.
Lets face it, offering a target of a year for a product will improve shipment performance statistics; its just not very useful data. Consider if airlines started counting their on-time performance for flights by the day instead of by the minute, we wouldnt have much to judge by, would we?
Next Page: Microsofts sorry record.
Microsofts Sorry Record
Meanwhile, on the Microsoft front, CEO Steve Ballmer addressed a large audience of IT professionals at last weeks in Orlando, Fla., about security concerns about the Windows platform as well as the companys plans for its next-generation client and server version of Windows, called Longhorn.
He said a “Longhorn wave” is headed our way, with some parts arriving sooner and others later. And others even much later.
A year ago at Microsofts Professional Developers Conference, company officials detailed the many new features—some 700 or more—due for Longhorn and reaffirmed its schedule, albeit years in the future. Developers were handed a CD with a technical preview.
That plan became hash over the course of the next 12 months, and the entire project was reorganized to become a “wave” rollout rather than the previous tsunami strategy. Rather than acknowledging a development and marketing failure, Ballmer would have us believe that this change (what some call “Shorthorn”) is a good thing. And thats likely so given the raw state of the components.
To be honest, Microsoft should find a new name for its upcoming releases of Windows, since “Longhorn” is now a misnomer. The Longhorn project has had so many facelifts over the past year or two, its difficult to recognize.
Also evident at the Gartner event was the confidence gap between customer expectations on security and the marketing messages delivered by Microsoft executives.
“Weve learned more about security than anyone else in the world,” Ballmer said at ITxpo last week. “We need to focus in on a few things. We need to engineer in fewer vulnerabilities going forward. We have new development tools to spot security vulnerabilities. We will release those to users.”
A “few” things? Who is Ballmer kidding? Nobody.
“Trust is not a word that I would use” in relation to Microsofts promises on security, one developer said following the keynote address, adding that she had no reason to trust Microsoft because it “hasnt delivered anything to date” to improve the condition of security.
The condition of Windows security is a joke, or it would be one if the issue didnt cost every user on the Internet and impact the workflow of every IT manager in large and small enterprises every single day. Or as revealed last week, cause PC makers to try to improve things on their own with a third-party security software bundle.
So, how can these companies regain our trust? Simple: Tell us what will be done and when. And then do it.
Of course, thats the toughest part.