You don’t use reference content the way you read a novel, and that is true of  print and digital manifestations: reference is navigated differently.  How does reference work? Why can finding an answer to a reference question be so satisfying – or so maddening?

I spent many years creating, managing and editing reference works (dictionaries, encyclopedias, chronologies, you name it), and although I was responsible for the transition of several reference works from print to digital, I did think for a while that reference publishing would disappear as reference print publishing died out, ever since Wikipedia took over the encyclopedia market. Certainly, the general-purpose A-Z encyclopedia disappeared in print form.

But it is clear that reference publishing is not dead. It has changed, with digital delivery, and it has different requirements, but still has a value. There are criteria for evaluating reference works, in whatever form they are delivered. It is interesting to reflect how reference works succeed or, in my opinion, fail. These may be print or digital works, but in some cases the principle is the same.

Where reference has been transformed is by developments in digital information retrieval. Thesauri and taxonomies have become the basic building blocks of many content-based reference websites. Wikipedia turns out to be in some ways more valuable as a machine-readable tool (DBpedia) than in its own right.

So reference continues, albeit dramatically changed in form. Nonetheless, the principles of creating a taxonomy (and a good reference work constitutes a taxonomy in a specific domain) remain fundamental. It is possible to evaluate reference works on very different topics, from the point of view of their construction and validity as ontologies. Of course it’s impossible to compare reference works in two separate subjects. There is no meaningful comparison between, say, a dictionary of geometry and a companion to eighteenth-century politics: you certainly can’t compare the subject matter. Nonetheless, some comparisons can me made. How easy is it to find the information? Does the content make assumptions about the user’s prior knowledge? Are there sufficient links and cross references? Is it clear what the scope of the work is? Does it require (and does it have) an index? For all these things, it is possible to say that a work is better or worse than another. Not many people make such comparisons, but that doesn’t make the exercise any less useful. For that reason, this site will look to compare and to evaluate from a reference perspective widely differing titles. Most importantly, could this work be used as the basis for a digital resource?

For some of the principles on which these reference ideas are based, see my Top Ten tips on reference publishing . These were based very much on practical considerations of compilation for publishers, but the guidelines still serve. They don’t address more fundamental concerns about what reference publishing is all about, in an age of linked data  – that will need another, different, ten tips.