Michael Keller, University Librarian at Stanford University, gave an excellent review of and argument for linked data. He said that the problem with information today is that there are too many silos, too many search engines, in too many places and looking different from one another.
We do not make it easy for users to get what is local and what is easy to get. We give them better interfaces and show them other tools, all of them good, and suggest that our clients use them to broaden their search. We routinely do not refer clients to the web. Our OPACs and the other services do not refer to more than a tiny fraction of pages available on the web. We therefore confuse our readership.
Some of us provide our readers with lots of secondary databases–too many–in some cases over 1000! Selecting a database to search is something of an art. We have made it very difficult. This shows only a very minuscule portion of what is available, and our users have to search through it on their own.
Our users, authors, teachers, students, and consumers, need us to help them find a better way. Ideally, librarians and publishers will make facts about what we have discoverable on the web. But:
- There are too many stovepipe systems, The landscape of discovery and access is a shambles and cannot be mapped in any logical way.
- There is too little precision with inadequate recall. Most of this problem lies in limitations in the design and execution of the infrastructure that supports discovery and access.
- We are too far removed from the Web. Together, our metadata collections make up a big chink of the “dark web”, and it is clear that visibility on the web promotes dramatic increases in discovery and access.
Our working environment consists of consumers and users, publishers as intermediaries, and libraries as warehouses. Then there is the Internet, and the web of pages of information. The web of data is the next big thing to empower individuals. The next phase is the Linked Data phase, leading to the semantic web, where the machines will understand meaning and act on it.
The construction of the linked data environment and the recipe for creating it includes identifying entities embedded in the knowledge resources, tying them together with named connections, and publish the relationships as crawlable links on the web, then build and use apps supporting discovery by the web of data. Here is a simple example.
RDF triples are a way to describe objects and ideas on the web; URIs allow machine interaction among Web objects. But our metadata standards are closed. Passive metadata is searchable by word, but it is in silos. The search results are refinable, but final–there is no way to go beyond them. In contrast, semantic metadata is open, dynamic, interactive and responsive, and leads to other queries and other views.
Many publishers and societies have begun to make use of linked data. They aggregate content in their own realms and beyond, and provide actionable, constantly updated links and compellig services tying users to them. Here are a few publishers that have adopted the semantic web.
For publishers, a library’s content is king, but if users cannot find it, there is a problem. Publishers must make their content visible. Aggregation is very important.
The Linked Open Data Value Proposition by the Stanford/CLIR Linked Data Workshop held in June 2011 is an encouragement. Google is using Stanford’s bibliographic facts and Web resources to create linked data pages using the Freebase Open Data Browser.
On Monday of this week, LC announced a bibliographic framework for the data age.
We are headed to the semantic web where uniquitous computing and mobility are essential.
Columnist, Information Today and Conference Circuit Blog Editor