Please release me

I'm still staggeringly busy, trying to do my own job a few months into a new role, covering for a colleague who was beaten up and supporting colleagues who took over my erstwhile duties. This workload and other events beyond it have left little time to comment with any value on subjects I would otherwise have addressed.
But here I am blogging at a desk designed for someone six inches shorter than me at the Queen Elizabeth conference centre in Westminster, attending a software quality and systems conference. This event is no where as boring as it might appear; I'm scrawling in an interval not during a keynote. I've heard a number of interesting propositions and we're promised a giant Scalectrix set later...
But I digress.
Microsoft sent someone here direct from HQ, a team leader on the Visual Studio release whose presentation was well received by most, though by no means all. Should Hackneyed be rephrased Redmond?
This particular team leader revealed that Microsoft has one defect for every ten lines of code, but that 5% of known bugs remain before release. Given that just the test suite of Visual Studio will require a bare minimum of 10,000 lines of code, this means it will be sold with at least fifty bugs; imagine then how many there will be in Longhorn, the successor to Windows XP, or how many were released in the Microsoft Office suite.
The speaker's justification for this state of affairs was the pressure of time to market and that if you have millions of users they will always find a bug. This, I felt, was a complete cop out. I appreciate that they're a lot less complex, but I develop web sites. These too have millions of users, are developed in a scale of months rather than years, but no one charges hundreds of pounds to use them to get a message asking you if you want to report a problem. They just don't return to the site.
Microsoft runs a roughly equivalent number of testers to developers on each project – if only we could charge services at those rates – but there appears to be an assumption that everyone will want to contribute to the company's success. People beat down Redmond's door to work for or partner with one of the world's largest companies, but that doesn't mean they want to pay for a tool that does only part of what it should and will break down once people hack its flaws.
The customer defines quality. If this generic customer ever evolves its expectations then, as with Firefox, Microsoft will inevitably lose its market share, particularly to other products that have similar defects but are free and constantly updated. Why pay to inherit someone else's problem and test it for them? There's hope for open source yet.