Hi Janos,
Please define 'extremely fine grained' control, and what variables you
want to grant to your users.
Well for example there is the spamSubjectPrefix setting, which I would
like to let the user decide. We do virtual hosting, and our customers
like to have as much control as possible, so being able to set that
would be nice.
Another option they might want to change is the kill level which seems
to be represented by spamOverallLimit and spaminessOblivionLimit.
Also whether or not to use anti-spam and anti-virus should be a
user-definable option in our setup. On the other hand, I don't want
them to enable training mode etc, but we'll make sure out user interface
doesn't allow for this at all.
reading the docs, I saw that clapf only seems to supports policy
groups and assigning a group to a user.
Yes, and by reason. Some variables have effects on clapf performance,
and I believe they are for the admins only. Users may create such a
policy causes unexpected results. In that case you should expect more
complaints from users that the system is broken/fails/sucks/...
So I designed the policies out of the users' scope.
I see, but like for us, all management of user profiles is done through
a custom developed front-end, which can pretty much shield off anything
we don't want to expose. So if we know what options are "dangerous", we
can simply not bring them to the user. :)
Frankly, I'm a bit shocked that you really need 2k policies, though you
are free to create as many policies you want, however I believe that in
most cases ~5-10 policies are good to go.
No I probably won't need 2K policies. If it isn't possible to fix it
within clapf, we'll code it so that if a user chooses a combination of
options that isn't in a profile yet, a profile will be automatically
created. That way we limit the number of options, as well as allowing
the customer to change whatever he wants. Of course, to make this work
nicely, we would need to know what the performance killers are.
Also, in the LDAP properties list, I don't see an option to enable or
disable virus filtering in a policy. Say that a user just wants to keep
spam out of his mailbox but doesn't care about getting viruses, how
would I do that?
Kind regards,
Gerry.
