[Design] Time tracking; Offline replication and data backup; Policies

[Paul Saitta]

(sorry about the duplicate)

  One relatively small feature which might be quite useful is time
tracking -- a lot work environments need or use this, and I can see the
potential for interesting uses of this when combined relatively seamlessly
with email archives, calendar archives, etc to form an easily minable
historical record.

  A question I have (which I assume has been considered) is how offline
replication will be handled -- if I want to check someone's calendar, and
they run the application on a (currently disconnected) laptop, where does
that data come from?  This isn't a new problem, but it has interesting
implications in a serverless environment.  A corollary is the problem of
backing up data, especially in larger installations -- it would seem to be
beneficial to provide for data backup in a more managed manner than simply
assuming or hoping that the filesystem of the machine the program is
running on gets backed up.

  All of this gets more complicated if users use the same installation for
both business and personal purposes -- resource separation might become an
interesting problem.

This brings up the further issue of policies -- for example, say a
corporate environment dictates that all email older than three months be
deleted, while non-work email belonging to the same user is separate.  How
are policies communicated in a manner consistent with the p2p concept of
the application (trust hierarchies?), and how are separations like that
maintained?  How much meddling into these sorts of issues is desirable?

/P.

-- 
Ideas are my favorite toys.