While I think we can all agree with Scoble that rewriting 60% of an operating system just isn’t going to happen without some serious additional delays (triggering a massive revolt among the peasants of Windowstan), that assertion only answers half of the question. Namely, “Will 60% of Vista be rewritten before it is released?” I don’t think so.
The other half of the question, which scares me to the point of causing echoes in my underwear drawer, is “does 60% of the Vista code need to be rewritten?”
“Need” is a relative term. I need Internet access. People on the street couldn’t care about that, they need food and shelter. When you talk about need, you’re implicitly talking about the “or else”. I can live without Internet access, but I need Internet access or else I can’t conduct my business. With software, when you say code needs to be rewritten, the “or elses” probably include issues with usability, support, and extensibility. In Smarthouse’s follow-up article, it appears that the biggest problems in Vista revolve around the Media Center Edition, but it’s hard for me to believe that those issues could even touch the 60% of code mark.
As a software developer myself, I worry that the 60% number that the “insider” tossed out to Smarthouse may be pretty close to the truth about how much code needs repair in order to be truly reliable and maintainable — but that a large portion of that 60% will be stamped “good enough” by Microsoft in order to avoid further delays in shipping Vista. I fear that only the critical issues are being addressed now, and that the rest will be left for cleanup in service packs.
Back in 1990 (after the release of Windows 3.0), I was one of those who predicted that Windows would take over the market as the platform of choice for business applications. Windows was very buggy then, and many of my clients scoffed at me when I spoke of their future conversion to Windows as a certainty. Now all of those applications are running on Windows platforms. Over the years, the quality of the Windows releases has improved (most notably with the NT architecture and its descendants), with a few bumps in the road along the way. As quality has improved, so has the length of time between releases of the OS. It seems to me that Microsloth has gradually adopted a more “enterprise” (read “red tape” ) culture in order to avoid bugs, but in the process they have not only lost their agility (ability to rapidly deliver quality innovation) but have also created a software development bureaucracy in which the focus can be lost, ultimately causing more problems than it evades.
If that is the case, they better turn things around before consultants like me start recommending other platforms for the desktop. Let’s hope that the recently announced restructuring helps. If Vista turns out to be another Windows95, I might have to defect from Windowstan to the United States of Linux.