One of the things we want to do with this blog is to explain some of the technologies and services we offer, and we’re going to start off with the OpenURL standard and Link Resolvers.
In a nutshell, A Link Resolver takes a reference/citation for a journal article (or any item), and provides various ways to access it.
The common scenario is:
- A user finds a record/reference for a item on a journal article indexing service, such as Google Scholar, Web of Knowledge or Scopus
- The service’s website has a link to the Link Resolver (for Sussex called ‘Find it @ Sussex’) next to the the reference. The link includes parameters about the article including its title, ISSN, the year/volume/issue of publication and page numbers.
- When the user follows the link, The Link Resolver will then show a webpage, which shows a link to an online version of the article, if we subscribe to one, and alternative means of accessing it as well.
How does this work?
The source website
First things first, the link has to appear on the original website (Web of Science, Scopus, etc). This isn’t black magic. These sites have agreed to provide a link to a Link Resolver (which in turn makes their website more useful).
The parameters are the same no matter which Link Resolver is being linked to, because it is an open standard (OpenURL), just the hostname changes. For Sussex it is http://sfx.lib.sussex.ac.uk:3210/sfxlcl3/
These sites know which Link Resolver to provide a link to as they know who is providing a user’s access (because they are subscription based). We provide details about our Link Resolve, its URL and name, when we subscribe to these resources.
The link itself takes the form of a base url (eg http://sfx.lib.sussex.ac.uk:3210/sfxlcl/) and then a list of parameters as defined by the OpenURL standard.
An example OpenURL:
Which is a link for: Toxicological evaluation of live attenuated, cold-adapted H5N1 vaccines in ferret. Vaccine [0264-410X] Vol 25 (52) 2007 pp 8664-8672
The Link Resolver
So a user clicking the link will create a HTTP GET request to the Link Resolver, passing parameters about the article in question. How the Link Resolver works internally will vary on the software, but basically it will use rules and a knowledgebase to decide (a) if we have access to the journal for that year (b) the URL the user needs to access the full text.
How does a Link Resolver know what we have access to (and from which site)
Link Resolvers will use a database consisting of thousands of journals. The knowledgebase is broken down by ‘providers’ (vendors) which are normally publishers. Within providers are ‘packages’, bundles of journals with certain coverage dates. These mirror the packages publishers typically make available to Libraries and Universities to subscribe to.
Example:
- Let us say we subscribe to the ‘Social Sciences’ package with a particular Publisher, which provides access to a collection of journals for volumes after 2001.
- This contains a journal ‘Economics Today’.
- Hopefully the knowledgebase will contain details of such a package, and we can flag we have access to it.
- A user finds an article on Web of Knowledge from ‘Economics Today’, published in 2005. They follow the ‘Find it @ Sussex‘ link which takes them to the Link Resolver.
- The link contains details about the article (journal, year, issue, etc) and the Link Resolver knows we have access to it, because it knows the journal (and volume) are in the package above, which we have flagged we have access to.
The final stage is making a link to the full text of the article. This is normally done following some rules for each publisher and package. For our example (fictional) publisher package, the url may take the form: http://journals.publisher.com/$ISSN/$year/$pageNumber.
So the Link Resolver, knowing this format, can construct a URL with the information it has about the article.
More than just the full text
A Link Resolver can do more than just link to the full text, especially if we don’t have access to it! It has the details of the journal you want, so it can do quite a few things with this, such as provide an import file for Reference Management software such as Endnote, or offer to search other catalogues and websites with the details. Potentially Link Resolvers could work with much more than just journal articles, they have started to add (e)books to their knowledgebase and could really act as the middleman for any type of document.
Providing a website to search and browse for journal titles
If you think about it, the knowledgebase will have details of all the journals we have access to, and the coverage dates. So as an offshoot, the same system provides the web interface to our online journals search/browse section of the Electronic Library. Using the data we maintain here.
One of the useful things about Link Resolvers is that they can help avoid broken URLs (something quite annoyingly common with publisher websites) and journals changing between different suppliers. So long as a link goes to the Link Resolver, it will continue to work unlike one which goes straight to a article which could break.
So, keep a look out for those Find it @ Sussex Buttons.
And if you’re interested, have a look at the OpenURL specification (version 0.1 is much easier to understand than version 1.0!).
Filed under: APIs, Developer, Services Tagged: | linkresolver, openurl, sfx
[...] have no access to the fulltext because they are taken to the publisher website. This is where a link resolver can be handy. The second issue is that the link will not use a ‘proxy’ for allowing [...]