Do you know if users were doing *Google* searches (as opposed to Google Scholar searches)?
I didn't turn up any bitstream results in Google Scholar, but there are plenty in Google proper.
For instance, I get "about 908" results by searching Google for
For examples of searches that retrieve both bitstreams and metadata records, a search for
yields "about 6190" results.
Here are two examples of a bitstream appearing with higher rank than its metadata record. In these cases I couldn't find the metadata record in the search results. YMMV.
1) The 14th result in a Google search for
"indiana geological survey" site://scholarworks.iu.edu/dspace/
is a bitstream:
I could not find the metadata record ( http://hdl.handle.net/2022/432 aka https://scholarworks.iu.edu/dspace/handle/2022/432 ) in the (about 187) search results.
Results in Google Scholar are, of course, different. (No site search is possible there.)
"indiana geological survey" author:j-steinmetz
gives the metadata record as the third result.
This is very interesting... Google Scholar's indexing of item metadata
vs bitstreams has evolved over the years. Early on they thought users
were absolutely not interested in seeing item records -- they want to go
directly to the bitstreams -- but I think that belief has changed and
now they are definitely indexing both item metadata and bitstreams.
In this case, I wasn't able to reproduce the problem in Google Scholar.
I searched by terms that were only in the metadata and only in the PDF,
and in all cases GS took me to the item record first... I couldn't find
a search that took me directly to the bitstream (although I've seen that
behavior in the past). Do you know how your user got to that result?
Anyway, assuming there are still cases where GS takes the user directly
to the bitstream, I think they'd be interested in your feedback. The
particular bitstream you provided is a great example of why a user might
need to go through an item record... the bitstream has absolutely no
embedded metadata or other context to help the user figure out what
they're looking at.
Maybe Rob Tansley can also comment, or we can take this up again with
Anurag at Google Scholar.
> We've had complaints from users who have found out-of-context DSpace
> documents with Google searches, such as
> _https://scholarworks.iu.edu/dspace/bitstream/2022/1333/1/7(1)81- 82.pdf_
> There is, of course, a DSpace metadata record, but the Google search
> retrieves the document itself with no link to the metadata record.
> We're considering disallowing access to the bitstream path so that
> users will not be confronted with raw documents with no metadata
> context, but do not want to lose the full-text searching
> functionality. Has anyone found a solution to this problem?
> Thanks in advance for help.
> Julie Bobay, Director for Scholarly Communication Initiatives
> Indiana University Bloomington Libraries
> Wells Library E1060
> Bloomington, IN 47405
Associate Director for Technology
Dspace-general mailing list