It is one thing to digitise a collection for the benefit of researchers: people arriving with a clear objective in view who are happy to fill out a catalogue query form to find what they need. It’s quite another challenge to open up a vast archive of material to uninitiated, casual visitors and encourage them to browse and stumble upon things that they would never have thought to go in search of. Using the first type of website resembles a trip to the British Library, where readers have to fill out a request for each book they wish to consult. The second type is more like a well-stocked bookshop with irresistible table displays and an inviting sofa where you can happily while away an hour, thumbing through the armful of books that caught your eye (and you’d only popped in to grab the latest Nigella Lawson for your foodie brother-in-law).
I’ve worked as a designer and a developer on Tate’s website for five years. My major focus has been the digital presentation of our art collection which we entirely redesigned as part of the new Tate website, launched in April 2012. Over the next twelve months I will be part of the team working to extend this online collection to incorporate the archive material now being digitised for the Transforming Tate Britain: Archives & Access project. It’s a hugely ambitious programme that presents challenges at every stage – and high among them is how to ensure that the end product has that ‘bookshop-sofa’ feel.
The Tate website already features a public catalogue of all the Tate Archive’s holdings. We could have looked into extending this down to the level of individual items and simply inserting all our new catalogue content and photography in situ. This might have satisfied the aim of publishing our archive in digital format, but would it really have achieved the goal of opening up access to this material to new audiences? The risk in such an approach was that our newly catalogued records could be consigned to a digital vault, screened off behind a daunting search form, where only expert users – knowing what they were after and how to ask for it – would ever be likely to find them.
Our preferred approach from the early planning stages of this project has been to build on previous work to digitise the artworks in the Tate collection. We decided that all Tate’s digitised collections, whether of art or archive material, should be brought together in a single, easily searchable database and presented in the same accessible, visually appealing style.
It seems fundamental to us that a visitor interested in, say, Paul Nash should be able to use one simple search to retrieve not only the paintings by Nash in the Tate collection but also the letters of his in the Tate Archive that have been selected for this first phase of digitisation. The case for a unified point of access becomes even clearer when you consider that many of the items in our archive collections can be considered artworks themselves. When our archive material goes online next year, a visitor searching for Graham Sutherland will find both the 56 large-scale works in our art collection and, in much greater numbers, the drawings from Sutherland’s sketchbooks, now held in the Tate archive. Normally spread between various physical locations, these will be brought together for the first time, virtually, in a single set of search results.
By speaking of a ‘single database’ I don’t mean that we necessarily set our hearts on a fully unified IT system behind the scenes. Our art and archive collections are managed by different departments in the museum and the primary cataloguing data for these collections reside in completely separate source databases. Projects that undertake to merge multiple source databases can defeat even the hardiest of project managers. Our task for this project is merely to mask these institutional divisions from our online visitors, so that what they experience is a single pool of information where digital records for around 150,000 artworks and archive items can not only coexist but speak a common language.
The vocabulary of that common language resides in the metadata that describes our collection objects. Metadata is a term with many definitions but always carries the sense of ‘data about data’: terms that describe information or assets – usually in a standardised, machine-readable format. In order to unite Tate’s art and archive collections we needed to find a common way of expressing who made an object or artwork, when it was made, what type of object or artwork it is, and what subject topics it relates to.
Most of these metadata terms only work effectively if they are drawn from a single controlled list, otherwise you can end up with multiple versions of the same artist’s name attached to different collection items, or ‘apple’ and ‘apples’ being entered as distinct subject terms. Hence, Tate’s system architects decided that our archive catalogue records, although entered by the cataloguers into a specialised archive database, would need to be pushed through a series of intermediate systems already being used for publishing our art collection online, picking up the appropriate, controlled metadata on their way to becoming website pages. (It’s a complex process which I won’t even attempt to explain here!)
Metadata is critical to enabling our visitors to navigate a pathway through the collections we are making available online. Leafing through the entire set of available records is not a realistic option. A shared system of metadata will allow people to filter those 150,000 records into a more digestible set of artworks and archive items relevant to a particular person, subject, period or art movement.
Returning to my bookshop simile, metadata is like a shelving system. A collection database without decent metadata is about as useful as a bookshop whose entire stock is shelved in order of ISBN number. Normally, we rely on books being arranged by subject matter or author to get a handle on what’s available. But a digital collection has the advantage of allowing users to pick between rival shelving systems. The workers in a physical bookshop have to choose whether to arrange their fiction books in one sequence by author, to divide them up by genre, or to split ‘classics’ from contemporary fiction. In the online world, a visitor can break down the same collection successively by author, subject or date … or indeed use any combination of these.
Good cross-collection metadata doesn’t just offer website users a menu of options when they enter our collection through the front door route. In the real-life internet environment, a huge proportion of the people who land on any individual collection record page on the Tate website don’t arrive there as a result of using our site’s collection search but parachute in from an internet search engine or a link on another website. If each of those pages is a navigational dead end then the chances of these visitors discovering anything else in Tate’s collection before they leave the site again are slim.
When we developed the Art & Artists pages for this website we put all our metadata out on display, to encourage visitors landing on any artwork page to ‘find similar artworks’ (by clicking through to lists of artworks that share a particular characteristic of the work at hand) or to browse a sample of ‘other works of art you may be interested in’ (see screenshot below). The latter feature takes a bunch of significant metadata about the artwork – including artist, medium, date and subject matter – throws it at a smart piece of search engine software, and gets back a list of artworks that share the most characteristics in common. Both these features are there to encourage unforeseen onward journeys from individual pages, so that our visitors discover material that may be unknown to them but suits their tastes or interests.
Our intention in 2014 is to extend both these features to embrace our digitised archive material. The ultimate goal is that a visitor looking at a sculpture by Barbara Hepworth on the Tate website, having plunged straight in from a Google image search (unaware, perhaps, that Tate even holds artist archives), could be alerted to a relevant letter by Hepworth, written in the same year and now also viewable in our digital collection. Once our curious visitor has gone on to read the letter, new sets of suggestions might lead her to explore a whole chain of further archive items and artworks. If we succeed in pulling this off then bookshop-sofa status could well be in our grasp.
I hope this has given you a flavour of how we are approaching this project from the website end. I’d love to read your own thoughts on how to make large online collections accessible to the uninitiated. Please do leave a comment below, sharing links to any good examples you’ve seen on other websites. Alternatively, you might critique our existing collection pages. We know that these are less than perfect in terms of usability and we will be seeking to improve them as we introduce archive items into the mix. All your feedback will serve as welcome grist for our web design mill.