Nature.com has just released a new public link interface (resolver base URL is http://www.nature.com/openurl?) into its content using the ANSI/NISO standard “OpenURL Framework”. This now allows its users to build Web links using familiar citation-based metadata elements which complements the existing support for Web linking using the DOI, or Digital Object Identifier – a persistent and managed identifier. It allows users freedom from any proprietary link syntax. Here’s how it works.
An OpenURL link is a simple URL with a familiar querystring which is conformant to open and public specifications and provides a description of the object requested using either known identifiers (such as DOIs or PubMed identifiers, etc.) or metadata fields (from registered vocabularies). This differs from most URLs which either provide an internally hard-coded database key, or else a proprietary service structure, and thus must generally be considered to be “opaque”, i.e. they are not open to further scrutiny, other than to blind guesswork (which can be a hazardous undertaking and is also subject to change at no notice). An OpenURL, on the other hand, is defined by standards documents, and so is a reliable – and moreover “provable” – information structure.
An OpenURL can be built by a third party and targeted at one service provider, or repointed at will at another service provider. If both service providers support the OpenURL standard then both links will be successful. There is no longer any requirement to learn each and every service provider’s own proprietary API (or application programming interface). OpenURL is thus one of the great levellers within the Web information space – a veritable “glue” technology – and is being actively deployed by the digital library and scholarly publishing communities as they begin to ramp up their interoperability initiatives in support of academic and professional information sharing. (For an excellent recent review, see “Why OpenURL?” in last month’s D-Lib Magazine.)
While OpenURL is not defined as a network wire protocol – e.g. there is no prescribed response from a given link server – typically it has been used within library contexts to address the so-called “appropriate copy” problem, whereby a library with multiple subscriptions to online services could find itself (and its users) badly caught out if it did not have a subscription for a given link provider (i.e. the one issuing the link). (See this D-Lib paper for a review of the “appropriate copy” problem.) OpenURL technology instead allows for the downstream communication of an object description which an institution’s link server can then mediate (or broker) and provide the best links for its own local users using local knowledge of subscription entitlements. This is a truly remarkable – if not astonishing – development of a cooperation between publishers and their customers, which does nothing less than allow the customer to determine the final destination of network endpoints, i.e. the customer chooses what it is that is “published”.
There are two strains of OpenURL available: an early de facto “SFX” specification which is a very simple hardwired set of URL querystring keys, and the full-blown ANSI/NISO information standard Z39.88-2004: “The OpenURL Framework for Context-Sensitive Services”. The Nature.com link server can handle either form, and more interestingly can handle a combination of both since there is no possibility of key collisions between the two forms so they can both co-exist on the same URL.
The Nature.com link server (resolver base URL is http://www.nature.com/openurl?) aims to link its users directly with content hosted on its platform. The recently developed CrossRef link server also has this same aim. Institutional link servers, meanwhile, may choose to “maximize” the links made available to end users – and this is something that Nature.com could pursue at some later date. OpenURL is further supported by Connotea – Nature.com’s well-known online reference management utility – which allows its users to register their own favoured link servers and so, where possible, Connotea will build outward-bound links to these link servers.
OpenURL is very definitely a technology on the move. Watch this space.
(Ooh, and, by way of being my first post here and all of that, my name is Tony Hammond and I work in Nature’s New Technology team. A strange fascination with identifiers and metadata, a long association with other climes – hot and cold, and a peculiar conviction that England might at least make the quarter-finals. It could be.)