We were without JC and Lucia, so we weren't really able to make progress
on the issue of CalNet friendly names being private. We had a
general discussion of the topic anyway. Tom and Mike agreed that it
didn't make any sense to coordinate a supposedly private namespace with
a public one. Jerry pointed out that you could give people other options--
a web form could default to the uclink/socrates name, but let people choose
something else. This might not violate the privacy guidelines, but
we'll have to continue the discussion later.
We moved on to the matter of whether we should automatically generate
addresses. There are a lot of questions to be answered there. If we
generate them but don't publish them, spammers may still find them (Mike).
People reacted badly to the roll-out of SurePay--should this service be
opt-in only? There are FERPA regulations too. (Tim) Currently we get a
student's e-mail address as part of their application, it's stored in
BearFacts but not in LDAP, but it may be stored in LDAP in the future.
FERPA's restrictions are applicable only to third-party access; the
institution or department that has an address can use it, it just can't
disclose the address to a third-party. We're not sure how this applies
to creating a new forwarding address, but we should think about it.
(Mike) We could look at it different than assignment. Maybe if you mail
to first.last@Berkeley.EDU it does a lookup--if it finds an unambiguous
match, it mails that person, if not, it bounces back with a list of
close matches from LDAP. You still have the privacy issues of course.
(Jerry) Why would we need to auto-generate? Why not just make it opt-in?
(Mike) The service is very scalable, so we don't need to make a big hardware
investment up front. That means you don't need to worry about how many
people use the service initially; if it's not many, big deal. (Tom) The
problem with opt-in is that you'll get many of the students, but
faculty/staff will be very slow to adpot it. You'll get maybe 20% of those
groups, while a much larger portion of them would benefit from using such an
address. If they're given an address they can change, they'll be more
likely to get involved with the system. (Jerry) Yeah, but do we care?
If someone doesn't want to take the effort to do it you can't force them.
You'll get stale addresses if you make people choose something they're not
going to use.
(Gordon) You could coordinate everything very closely, so people are only
managing one address. The @Berkeley.EDU address could be created at the
same time as the uclink address, and go away at the same time. This would
be the easiest way to manage the service. (Tom) The easiest way to manage
the service would be to ignore all namespace conflicts. If you try to
coordinate @Berkeley.EDU creation and deletion with uclink creation and
deletion, you'd have to change lots of processes and write lots of code.
A clean namespace, you'd only need to write the code to set your address,
which you'd have to write anyway. (Janet) So why aren't we doing that?
(Tom) And what about people who aren't on uclink--how are you going to
coordinate with the hundreds of mail servers out there that the
@Berkeley.EDU address gets deleted when the local mail account gets
deleted? (Gordon) I'd be willing to make everyone get uclink accounts.
(Mike) Bad idea.
(Tim) What about using the account name the person already has in
LDAP? If they're using tim@uclink, make their address tim@Berkeley.EDU.
(Tom) You'll get conflicts with departmental mail servers. (Tim) Maybe
you could use the LDAP addresses as a first pass, then send mail to
people who have conflicts telling them to choose a different address
for @Berkeley.EDU.
We were past noon and losing the room at this point. I've been asked to
write up something for BC&C, deadline Jan 7th. Let me know if there's
anything you think should be addressed in that article.
Our next meeting is tentatively scheduled for 1:30 PM on Monday, January
7th, in 261 Campbell. Have a good holiday.
-- Tom Holub (tom_holub@LS.Berkeley.EDU, 510-642-9069) College of Letters & Science 249 Campbell Hall
This archive was generated by hypermail 2b29 : Mon Dec 17 2001 - 15:52:20 PST