Posted by: Anonymous Coward
on May 20, 2004 11:10 PM
This article effectively disposes both the AdTI thesis and its presenter, thank you very much. So much for Ken Brown. As for the microkernel debate, Tanenbaum has a point that monolithic architectures are probably less secure. However, experience has shown that a true microkernel results in excessive kernel/user mode transitions for many demanding applications, including common ones like games and computer browsers. Of course, Microsoft exacerbated the problem on Windows by bundling rich desktop apps as "architectural components" of the OS in a (mostly successful) attempt to defeat competitors and antitrust regulators. Maybe there is a middle ground between today's OS architectures and the microkernel designs of the '80s.
microkernel debate redux
Posted by: Anonymous Coward on May 20, 2004 11:10 PM#