[open-bibliography] Sharing bibliographic user data
adrian.pohl at okfn.org
Mon Nov 15 15:36:41 GMT 2010
I am very happy that LibraryThing is thinking about opening up some
data. This discussion drove me to finally contact the people behind
BibSonomy whether they would open up the BibSonomy data. In
genreal, they are interested but by their response some more questions
were raised concerning opening up user-generated data.
BibSonomy is a social bookmark and publication sharing system. Until
now BibSonomy offers an API to reuse the data in library catalogs
etc. and they provide full dumps for download by researchers. To
get these dumps you have to sign a license agreement. As said, they
are open to releasing data under an open license but have some legal
- It's not they who collect the data but the users so the question
is: Is it that every user must agree to opening the aggregated data?
- Or can the service providers themselves open up data without asking the users?
- If it's ok for the service providers to open up the data: Which data
are they allowed to release under an open license? All data that is
publicly available anyway? Or should user-specific data be deleted to
some extend? E.g., is it OK to leave user names in place or should
they be anonymized? BibSonomy also has real names and even some
CVs? Could they make these open without asking?
To get open data from social reference management services like
BibSonomy, book lovers sites like LibraryThing or user-generated data
from library catalogs we should provide some answers to these
 See mine for example: http://www.bibsonomy.org/cv/user/acka47 or
this one with photo: http://www.bibsonomy.org/cv/user/kw
2010/11/4 Ben O'Steen <bosteen at gmail.com>:
> I think what you are asking about is not sharing data, but creating a
> set of the data and sharing that as a separate resource (like a zip file
> of csv or rdf)
> For example, given an API in which usage data can be retrieved in real
> time, many will just 'copy-by-reference' - pulling in the data on demand
> and so reflect the data as it is at any given moment.
> If you release a snapshot of data (to help lower API load) then you can
> remind the reusers to respect that the data has a best-before-date,
> especially for things like labels, lists and tags and to remain as in
> sync as practical.
> If your users will be upset with the perceived lack of control
> downstream, then perhaps you cannot freely share the data? Perhaps a
> survey or a mailshot to all your users illustrating what you would like
> to do may help find out exactly how they may react?
> On Thu, 2010-11-04 at 10:57 -0400, Tim Spalding wrote:
>> Thanks for the reply. I broadly agree. Let me put part of my question
>> in concrete terms.
>> 1. Most users on LibraryThing make their tags and reviews public.
>> 2. LibraryThing exposes their tags on library catalogs (our
>> LibraryThing for Libraries product). Users don't object to this. (If
>> they do, they can opt out on the reviews side.)
>> 3. BUT users know that the data is coming from LibraryThing. If they
>> change their review, it will soon change on the libraries too. If they
>> removed it, it would get removed.
>> 4. I believe that if LibraryThing merely released the data, with no
>> obligation to refresh it, people would be VERY upset.
>> See? The problem isn't openness. It's respect for the user. Users are
>> happy with their data going other places, but they still think of it
>> as theirs, and want to make sure they can change or remove it.
>> How is that solved, in an open context?
>> open-bibliography mailing list
>> open-bibliography at lists.okfn.org
> open-bibliography mailing list
> open-bibliography at lists.okfn.org
More information about the open-bibliography