I'm not immediately familiar with the private request collection but I
assume there is still only one of these private collections for the
entire request. If that is the case, that doesn't really address the
issue of it being bloated with unrelated keys with possible name
collosions-- it only mitigates it slightly.
I think Seth's idea of a well-named struct for all data needed for a
certain view at least addresses the latter issue.
~Brad