My first impressions on this stuff, in a lecture from Michael Leworthy, was 'hmmm sounds like they're making it easier for managers to be control freaks and harder for developers to actually get things done.

But something won me over.

When you attempt to check an item in to the version control system, you are told if your check-in violates the checkin policy. (e.g. you haven't associated it with a workitem, or you haven't run code analysis... this is all configurable and you know some people will go policy nuts). But, there is always no matter what, the option to over-ride and JUST CHECK-IT-IN ANYWAY.

Sure you have to provide a reason why you ignored the policy rules (e.g. 'The policy that i must make my manager a coffee before checking a file in, is rude and inapt'), and the policy-breaking check-in may cause an email to be generated for your manager, telling him or her what just happened -- but at least this humane hole in the process reflect reflects a kind of real-world philosophy that sounds much nicer than the old draconian software-enforced check-in policies I've seen elsewhere (I'm looking at you IBM).

secretGeek's ticket, accomodation and breakfast cereal at Teched is brought to you by the very intelligent folks at advantech software.