Tuesday, October 11, 2005

WSJ Article: How to Ship Better Software

Robert Guth of the WSJ recently wrote a fascinating article with respect to Microsoft's legendary challenges shipping Longhorn, WinFS, and quality software products. The article, titled Code Red, Battling Google, Microsoft Changes How it Builds Software, is an insider's review of Windows problems and the team assigned to make the product more modular, extendable, and easier to test. If you can, read it.

Not long ago, Detroit took 5-7 years to take a car from blueprint to the dealer's lot. Innovation fell prey to the inefficiencies of the Big 3's product development processes and customers abandoned US cars in favor of Asian manufacturers who responded more quickly to consumer tastes and sold higher quality products. Guth's article positions MSFT as the GM to GOOG's Toyota and underscores MSFT's inability to be first to market with innovative (desk top search, ad words, tabbed browsing, maps, etc), (endless patches and security warnings) quality products. Windows proved to be too large a boat anchor to allow MSFT to predictably ship products ahead of competitors.

The article gives credit to Jim Allchin, Window's top executive, for an effort to refactor Windows and, more importantly, the tools, culture, and processes of the Windows development organization. Allchin and Amitabh Srivastava set out to improve quality via new tools that automated unit testing, rejected checked-in code that failed quality checks, improved build processes, system tests and coverage, and a culture traditionally more focused on feature additions than architecturally integrity and quality software.

As a former BOD member of Klocwork, I know first hand that the enterprise and ISV market suffer from poor development processes, a lack of automated source code analysis tools, and a culture of missed ship dates and brute-force solutions. The market is waking up to the need to fix problems at "day zero" and to maintain architectural integrity as products mature. Failure to do so results in products that make innovation very challenging and impossible to maintain.

While MSFT is a poster-child for buggy products, the industry as a whole can benefit with a new generation of tools that improve software quality and reduce the cycle time for new releases that meet customer needs.

Hummer Winblad is an active investor in development tools solutions, with investments in companies such as Akimbi, Palamida, and others still in stealth.

If you know of other companies of note and interest attacking this problem, please send them my/our way.