Unless you're storing huge, uncompressable binary files, we've found that disk performance is the expensive part of the maintainance story, not disk capacity. TFS servers with lots (hundreds or thousands) of active users often need a dozen spindles to keep them fed, especially the TempDB. Brian Harry's blog has touched on these factors: http://blogs.msdn.com/bharry
Rest assured Destroy is very high on the list of vNext features.
Microsoft is conducting an online survey to understand your opinion of the Msdn Web site. If you choose to participate, the online survey will be presented to you when you leave the Msdn Web site.