From Mike Conlon, VIVO Project Director
Learning is fun! Register now for VIVO Camp, April 6-8, Albuquerque, New Mexico. Join Violeta Ilik, Paul Albert, Mike Conlon and Graham Triggs for an introduction to everything VIVO. If you are considering VIVO, planning a VIVO implementation, have started and need to know more, please plan to join us in Albuquerque! Modeled after the highly successful Fedora Camps, VIVO Camp will offer high quality instruction, breakout sessions, opportunities to have your questions answered. Should be a great way to meet leaders in the VIVO community, share experiences, and learn more.
Triple Pattern Fragments coming to VIVO. A triple pattern fragments (TPF) server has been added to VIVO for the next release. You can try it now on OpenVIVO at http://openvivo.org/tpf (OpenVIVO is a VIVO anyone can use. The VIVO Project uses OpenVIVO to demonstrate VIVO and features coming to VIVO. Do you have a profile in OpenVIVO? You should – just sign on to OpenVIVO using your ORCiD username and password. A profile will be created for you. Don't have an ORCiD username and password? You should. Just go to http://orcid.org and sign up.)
What is TPF? At the 2016 VIVO Conference, Ruben Verborgh of Ghent University gave an outstanding talk on Triple Pattern Fragments, "The Future is Federated." You can find his talk on OpenVIVO here: http://openvivo.org/display/doi10.6084/m9.figshare.3680310. Ruben has many papers and presentations on triple pattern fragments and the semantic web. His OpenVIVO profile has an extensive bibliography with many links to full text and presentations.
The basic idea is intriguing. Triples are great for representing data, but SPARQL is not that great for sharing data. Open SPARQL endpoints (servers providing open access to a collection of triples) are rare. SPARQL can be difficult to learn and write, and it is very easy to make mistakes that produce the wrong data, or lead to very excessive demands on the SPARQL server. Institutions running open SPARQL endpoints often must reset their servers due to such queries. The maintenance of such servers becomes a burden to sharing data.
TPF provides a solution. TPF servers provide data, but only to very simple queries that can be quickly resolved. In this way, TPF servers can easily be maintained. The simple queries can be combined by applications to get answers that might have required complex SPARQL and significant processing time for a SPARQL server.
Following Ruben's talk, it became very clear that VIVO and Vitro are outstanding platforms for sharing data via TPF. A TPF server has been added to OpenVIVO. It responds to TPF requests for triples. We will explore opportunities to use TPF in OpenVIVO. TPF has the potential to "leave data at rest." Software using TPF can query data across a collection of VIVOs, rather than gathering data into a central place to make cross site searches.