On tir, 2004-08-10 at 23:21 +0200, Simon Josefsson wrote:
I'm sad to see that ManageSieve apparently won't be on the WG agenda.
It was suggested ManageSieve was considered a "hack". Could someone
explain that?
since it was I who said it, I guess I should come forward. I only meant
that managesieve is a stopgap measure. the proper method would be ACAP.
I'm trying to understand what it is about that is a
"hack". To me, ManageSieve much less of a hack compared to
transferring Sieve from clients to servers over LDAP or ACAP. I've
explained it before, but my reason for that is that to be useful, the
protocol transferring Sieve will have to support syntax validation of
the script, and crafting that on to LDAP or ACAP is to me a very big
hack.
ACAP caters for dataset class specific validation rules. you're right
there is a problem, since the ACAP server can't know which extensions
the server supports, and indeed, there may be more than one server (ACAP
consumer) using the same values.
managesieve assumes that it is run on each server instance. I don't
think it is problematic that ACAP is used similarily. (it can partly be
hidden by referrals, ie. multiple servers for different uses at a site
can appear as one to a user.) the user can still easily migrate
settings between ACAP servers.
--
Kjetil T.