Posting to Autocat
On 2/12/2015 9:38 PM, Galen Charlton wrote:
On Thu, Feb 12, 2015 at 3:07 PM, J. McRee Elrod wrote:
Looking forward to linked data are we?
Actually, yes. Besides the consideration that any competent designer of a library discovery system that ingests RDF triples would of course implement local caching to guard against big SPARQL endpoints having service glitches, a world where all metadata of interest to libraries is web-accessible and has clear (and open) licensing terms, would make it /easier/ to maintain multiple archives of it.
For example, even today I can download a copy of the entire VIAF dataset  and do what I want with it, including keeping a copy of it for long-term preservation or building a new service using it. Various other library datasets are readily available .
This is a significant improvement over the days where it was necessary to drop a few thousand dollars to get one’s own copy of the NAF and SAF.
This is absolutely true, in theory. We must wait to see what happens in practice. Of course, while each library can keep a copy of VIAF, it is expensive to do anything with it. To implement linked data will demand a long-term and significant investment from different organizations to create a series of mirror sites, thereby easing the load on a single server, and creating redundancies for those times when something actually goes down.
All of the expertise and technology exist today and is being done by all kinds of organizations, but behind it all needs to be a long-term investment of money. This is something that is difficult to come by these days, especially the “long-term” aspect. For instance, in 2010 Google bought Freebase, which was considered to be one of the major parts of the linked data universe and Google has decided to shut it down, apparently after taking what they needed, and move it all to Wikidata. (https://gigaom.com/2014/12/16/google-is-shutting-down-its-freebase-knowledge-base/) Wikidata is a relative newcomer (began in 2012) and is paid for by “community contributions”. In this way, Freebase joins the long list of tools discontinued by Google (http://en.wikipedia.org/wiki/List_of_Google_products#Discontinued_products_and_services) It shouldn’t be any surprise that there is no button in Wikidata that says “Ingest Freebase data” and it is turning out to be a lot of work to move Freebase into Wikidata. (http://dataliberate.com/2014/12/google-sunsets-freebase-good-news-for-wikidata/) Having major parts of the linked-data universe being “community funded” should also give us pause. While Wikipedia may be fine, I hope that Wikidata will be too. Finally, what all of this means for dbpedia is still unclear (at least to me).
Putting all of these issues aside however, it is still very positive that –at last–the NAF and SAF are available to the public–but it is clear that we are only at the beginning of a long, and expensive, process to create something useful in a practical way for the public.