On Fri, 10 Sep 2010, Graham Murray wrote:
> Or the stats could do the opposite and calculate the hash for all
> submitted anonymized domains and aggregate the counts with any
> submissions which anonymized that domain and the html show the
> non-anonymized together with the totals for all reporters, both those
> who anonymize and those who don't.
Also, v2.2.0 contains a feature that prefixes all the hashed strings with
an additional string we don't get as part of the report. That means we
can't reverse-engineer the hash, so your data are truly anonymous (but
also can't be correlated with anyone else's data to find trends that might
be interesting).
We're happy to have whatever data you feel comfortable providing.
Naturally the un-hashed data is the most useful in terms of finding
interesting trends or specific use cases, but it's completely
understandable if you'd prefer to obscure the names of domains from which
you receive mail.
Received on Fri Sep 10 2010 - 06:49:57 PST
This archive was generated by hypermail 2.2.0+W3C-0.50 : Fri Sep 10 2010 - 13:50:01 PST