At some point, the torches and the pitchforks are going to come out and the masses of PC users are going to start banging on someone’s castle door.

• For example, there are folks who gritted their teeth and went ahead with a Vista rollout, assured that they were getting something significantly more secure than XP. According to eWEEK, "the top reason survey respondents gave for adopting Windows Vista was that the new operating system is perceived as being more secure." Those buyers will not be pleased to learn that thorough tests find security on both XP and Vista to be comparable and "wafer thin." Even critics of those tests concede that "Vista’s security is most certainly not a ‘slam
  dunk’…The mantra that Vista is an
  evolutionary step in security should be met with better results than this."

  That’s a serious problem, given the latest demonstration that users exhibit amazingly perverse behavior that makes thick-client security a constant concern.

• There are people who bought into Microsoft’s positioning of Office as a development platform, relying on the promise of portability of their custom code between Windows and Macintosh versions of those applications.

  Those buyers are now being told by Microsoft that their Macintosh customizations will require a migration step called "move to AppleScript" as part of any adoption of Macintosh Office 2008.

• There are resource-starved administrators of public-sector computing facilities who are just now getting the memo that current federal rules require massive improvements to their information archival and search capability.

  Congratulations, you moved your organization from paper-based to PC-based databases and flat-file tools: now you have to make their desk-by-desk data collections discoverable, whether you had budget for that or not.

It may take the math of catastrophe theory to predict the resulting "tipping point" behaviors, but my gut says that sudden massive changes in technology choice can not be far away. Less state on the client device, and less dependence on desktop OS APIs, seem likely to make a cuspy transition from abstract niceties to urgent goals.