> But we do encourage (promote) an interface that forces
> off-campus authentication to our Summon instance.

With an explanation that it's because of pirates! :-)
https://auth.lib.unc.edu/ezproxy_auth.php?url=http://unc.summon.serialssolutions.com/search?s.q=

>  And one we would need to revisit if we looked to Summon (or some other
> product) as a catalog+periodical literature hybrid.  Right now we have
> separate discovery layers

That's a useful factor to add to the discussion, thanks. We use Summon
for "everything", but I can imagine that the discussion about
requiring authentication could go differently if it was only used for
articles.

I was pretty sure that I remembered seeing some research showing that
authentication was a barrier, i.e. a certain proportion of users don't
actually bother to continue. I did some brief-ish searches (taking the
opportunity to compare Summon, EDS, Primo and Google), but the best I
could come up with was this (an evidence-free statement):
http://www.codinghorror.com/blog/2007/06/removing-the-login-barrier.html
(I also found a medical paper [1] with a small study that appeared to agree.)

David


[1] Tjora, Tran & Faxvaag (2005). "Privacy vs Usability: A qualitative
exploration of patients' experiences with secure internet
communication with their general practitioner", Journal of Medical
Internet Research 7:2.