Posting to Autocat
On 2/24/2015 11:24 PM, Galen Charlton wrote:
I agree with you that it will require an ongoing commitment from a number of institutions, large and small, to maintain both long-term archives of metadata and a sufficient number of redundant copies for the sake of availability. Furthermore, this will require some planning — LOCKSS, as a concept, works best if ongoing conscious action is taken to verify that enough copies of something we want to preserve actually exist. Fortunately, the profession already has a lot of experience dealing with such issues — both positively  and negatively. , .
However, one of the things I find appealing about Linked Data is that one is*not* solely reliant on the graces of centralized institutions to maintain ongoing access to the metadata. It’s easy  to hoover up dumps of triples and keep them indefinitely, and server capacity to do things like host a copy of the entire VIAF dataset and do interesting local processing is also relatively cheap.
Hi Galen (by the way, I have great admiration for your work, especially with Koha!)
I want to emphasize that I strongly believe that libraries need to adopt Linked Data, but everyone–especially librarians–should realize it will be a huge commitment with no promises. I am an historian, so I look at these issues through an historical lens. For instance, when I look at diagrams such as this one of the linked-data universe (http://en.wikipedia.org/wiki/Linked_data#mediaviewer/File:LOD_Cloud_Diagram_as_of_September_2011.png) in addition to seeing how much there is, it leads me to wonder how much of it hasn’t been updated in the last few years because people have grown tired of it, or how much has just been shut down altogether. Freebase was bought by Google and is now being stopped, while it looks as if the very center of it all, dbpedia, will metamorphose into Wikidata. If that is happening at the very center, what is happening in the other areas?
Another example is the research that showed that Microsoft Academic Search had not been updated for a long time (http://arxiv.org/abs/1404.7045. I don’t know what the current thinking is but such occurrences shouldn’t surprise us)
At the same time I have to confess (that is, it is my opinion) that I have yet to see anything that has yet been created with linked data that I think is such a wonderful or transformative advance. For instance, probably the best known examples of linked data are Google’s Knowledge Graphs and we read continually about how great the Knowledge Graphs are, but so far (again, in my opinion) it is all hype and spin. The Knowledge Graphs present elementary information that is not all that different from reading an entry in a normal encyclopedia.
In the library sphere, it is true that putting our data into the Googles and Yahoos as Bibframe or schema.org will make it more “widely available” but it will also present incredible additional demands from us if our data is to be useful practically to the public. Otherwise, it will languish at the bottom of the Googles and the Yahoos and remain just as unavailable as it is now in our “separate silos” in our catalogs. I have discussed this at some length in other posts.
Nevertheless, this doesn’t mean that nothing can be made. I remember a great talk by Tim Berners-Lee at
http://fora.tv/2009/10/08/Next_Decade_Technologies_Changing_the_World-Tim-Berners-Lee who gave some fascinating examples of what can happen when you share your data.
I live in Rome, so almost every day I see remnants of a fabulous and beautiful past that was lost and ruined. When libraries go into linked data, I wonder whether we will be re-enacting the landing on Plymouth Rock by the Mayflower Pilgrims (a success) or the establishment of Jamestown (a disaster).
We don’t know what will happen. I agree we must take that step because doing nothing certainly leads to oblivion, but we must also be aware of the further consequences and responsibilities.
So far, I haven’t seen much discussion about the potential pitfalls. I also like to think there may be other options …