> ....If your discovery strategy is
> predicated on having your scholarly IR harvested and presented to the world
> through a separate discovery tool and the vast bulk of your document views
> are coming from Google and Google Scholar users, does this lessen the
> 'compelling experience' requirement?...
Not at all. The reality is that the vast majority of users are not going to
start with a library portal, so no matter what you do, it needs to play
well with Google.
For singular items like individual documents and images, making things work
with different solutions isn't too bad. However, as the use case turns to
identifying all photos that satisfy a particular need, slicing and dicing a
dataset consisting of thousands of large files in a complex hierarchy,
exploring a large format book, relating audio/textual/visual archival
resources, etc, there are issues with getting the stuff efficiently into
the repository with appropriate metadata, how to maintain what's in the
repository (which includes tasks such as batch assignment of metadata or
updating items) in addition requirements the users may have with regards to
navigating or interacting with the items.