This week I’ve been at the Digital Libraries conference in London. I was part of the Doctoral Consortium on Monday and as well as attending the main conference, was lucky enough to be part of the workshop: The Search is Over. If you are even remotely interested in how we can open up museum and library collections and archives, develop alternatives to search or even just visualisation, I encourage you to read the papers presented (and not because one of them was mine) and check out the Twitter feed of the workshop hashtag tsio2014. I am only sorry the keynotes are not included on the site (at the moment) as they were all superb. There is talk that the ‘Search is Over’ concept/discussion will be taken elsewhere, I can’t encourage you enough to get yourself along.
If you’re interested, a modified version of my presentation text is below. The more formal paper is available on the Search is Over website (as a PDF).
Motivation: burying the books
The motivation for this research is something that is very common: the sense of loss when books are taken away from you. My university, the University of Technology in Sydney, has just opened their Automated Storage and Retrieval System (ASRS) – we call it the LRS. They’ve moved a large proportion of their books into the ASRS which is essentially a large hole in the ground. So, for a lot of the collection we can no longer get to it physically. We can look them up in the online catalog and request books from the ASRS, which are then retrieved by a robot.
The books are stored in large bins where they are arranged in a “best fit” manner. They’re arranged, for want of a better word, according to size and fit. RFID tags enable the robots to find them. They are OUT OF ORDER.
To get access to these books I need to use the existing interface. This interface stands between me and that book. You could say it stands between me and the collection. Except a lot of the collection these days is digital. The library holds many books in electronic versions, so I’m never going to get to a physical book in those cases. The interface IS the collection, and it will only continue to become more so: the less printed books libraries acquire, the more we see their collection as interfaces.
This doesn’t seem all that radical – we are, after all, at a digital libraries conference. The reality though is, interfaces don’t do it for us the way books on shelves do. Not because I can’t touch the book itself, but because I can’t see what’s next to the book I went in to retrieve. I can’t have a physical relationship to the collection. I can’t build an understanding of what’s in the collection, how big it is, or how lacking it might be, by standing amongst it.
Even though I can still order books from the ASRS, the retrieval moment is lost. A robot picks it up for me and in doing so takes away what is for some of us a small frisson of excitement as we get the opportunity to cast our eyes over the books around the one we want. It takes away my opportunity to browse the books on shelving trolleys, that lucky dip of the library. The physical aspects of the collection contribute to the decisions I make about the overall collection, about sections and individual books.
What UTS is doing is not unique. It’s also understandable and reasonable, given the demands of library patrons for more work space and the difficulty the library faces storing an ever expanding set of books. Many scholarly and research libraries are in the midst of changes that will affect access to their book collections. Take for instance
- the New York Public Library
- the University of Sydney library, and
- the (State) Mitchell library in New South Wales, Australia
In these three cases there has been a public display of displeasure at the library plans. Library patrons expressed feelings of loss: not always a loss of books as merely objects but the loss of an experience.
Some of them saw these changes as an attack on their idea of what a library should be (you can’t have a library without books) while others were concerned with the loss of serendipitous browsing or their inability to get to the book beside the book. In the case of the New York Public Library they got them to abandon their plans. The Mitchell Library made some compromises. The Sydney Uni library has moved much of their stack to storage, leaving a new, streamlined, low-fat, slim stack, with only marks on the floor that reveal where the shelves once were. Kind of like stretch marks.
It’s not that people aren’t accustomed to using library interfaces, they are, but not in isolation; they use interfaces in conjunction with the shelves. Searches don’t always end with the retrieval of a single item. We build further understanding of the collection through an engagement with the books themselves. Judgments and decisions are made within the retrieval moment, not only when using the interface to locate items. For example, when I’m looking at the shelves I can see which books are the new ones, which are the old, which ones have multiple editions, which ones are available for short loans (and thus everyone will want them). I can do some of those things through current interfaces but not to the same extent.
So why aren’t current interfaces doing it for us?
Current library interfaces aim to get us to exactly what we want. They assume we know what we want and how to put that into a query. Since Google came along and made searching the web a doodle, libraries have been under pressure to perform. The single search box has become the pinnacle of information retrieval interfaces. Which is a bit unfair on libraries.
Libraries are caught between a rock and a hard place: the structured world of the library catalog and the ad-hoc, user-centric, free text world of Google. They are limited by catalog data and things like proprietary software but they labour under the expectations of users who want their search to work like Google.
Despite being digital and ostensibly having the subsequent ability to sort information in almost infinite ways, library interfaces narrow users options. They are bound by the rules of library classification systems, without the benefit of the physical context they provide. The Dewey system curates a collection when it’s on the shelves, it ceases to have this ability once my access is reduced to search only.
So, given all this, as the name of this workshop attests, we need to start thinking more broadly about how we access collection. We also need to understand not only how interfaces operate and how they can be ‘usable’ but also how they shape our relationship with library collections.
Opportunity: the affordances of the miscellaneous
Some people, and I’ve been guilty of this at times, think all this moving books and reducing shelf space is a loss. It’s a BAD THING. For interface designers though it provides an opportunity. It is an opportunity to explore what is afforded by the potential of the ‘miscellaneous’. Here I am referring to the ‘miscellaneous’ as described by David Weinberger. The beauty of the ASRS is: once books are no longer publicly accessible, they need not be in an order that enables them to be found.
So, to recap Weinberger’s miscellaneous quickly: the way in which we impose order and thus access to physical objects and information has been freed by the digital. In Weinberger’s three “orders of order” we have moved from a “first order” limited by matter, through a “second order” where information about objects is maintained separate from the objects themselves through to the “third order” which is miscellaneous. Order is defined dynamically. For example, iTunes provides us with the ability to sort individual tracks in multiple ways, instead of being locked into a stipulated track order by the physicality of a vinyl record or CD.
In an automated storage system, where books are stored by size and retrieved by robot, the catalog doesn’t need to relate to the storage system because patrons don’t have to locate the book itself. This provides us with the potential to explore what could be done when you break the relationship of the library’s classification system to the physical arrangement of books on the shelves, when patrons no longer need to understand their physical location. It gives us the opportunity to explore what we can do in the interface when the classification system gets out of the way.
Once patrons can’t retrieve books from library shelves, the classification system no longer functions as a way-finding tool. In the case of the UTS Library, the Dewey number acts as a direction for the patron, and a location for the book.
This patron-centric view of the library, where the processes and systems involved in getting books in to the library and storing them are of no real concern, provides us with the opportunity to explore what can occur within an interface when the patron no longer needs directions and the book no longer needs a location. Of course, we are limited by the data at our disposal and the connections it affords but forgoing the idea of an overarching order still frees us up visually.
Current interfaces are underpinned by the library convention of overarching order. The overarching metaphor is one of dominant order. They reinforce the idea that there is a collection that is arranged in a certain way and the patron can negotiate this collection by searching or by browsing once search results are returned. They reinforce the cataloging concept of a single entry: that in cataloging an item it is possible to find the ‘proper place’ for the item.
Alongside the search field some interfaces provide the ability to browse using a pre-defined view of the collection through a single parameter such as geographic location or time; they offer faceted views of search results. Sometimes you can combine these things but the techniques used to communicate these views are simple lists. In these interfaces you can turn the collection around and view it from different angles but you can’t change the relationships between individual parts.
These interfaces are not exploiting the flexibility of the third order. In terms of a design approach it is suitable to ask what the possible affordances of the “miscellaneous” are in this instance. For example, are library patrons reliant on a sense of an over-arching order for their conception of the ‘collection’? Do they build their ideas of what the collection is and how they can access it through their understanding of and relationship to the dominant order? A different conception of order, as enacted within a library interface, is necessary for us to approach interface metaphors that move beyond the transactional and limited browsing metaphors offered in current interfaces.
An interface theory
In terms of a framework through which to approach the concept and theory of interface, I am using the work of Bolter and Gromala and Johanna Drucker particularly, at this point in the process.
Bolter and Gromala’s analysis of the interface through the lens of digital art provides a valuable framework through which to assess current library interfaces and develop alternatives. Particularly, their argument that we need to look at interfaces, rather than merely through them. Interfaces are representations of knowledge in and of themselves, not merely gateways to knowledge. This research will explore the visual presentation of library interfaces through this lens.
Johanna Drucker’s argument that interface design and theory has been dominated by a scientific approach and her call for a ‘humanities approach to interface theory’ is also proving important. For Drucker, there is an assumption underlying current theories that interfaces are ‘neutral’. There hasn’t been sustained consideration of the metaphors used within them and the assumptions that underpin them. Interfaces are viewed as pragmatic and instrumental rather than rhetorical and persuasive. I would like to explore whether or not this is entirely the case and if so, how can the development of speculative prototypes contribute to such an approach to interface.
So, the plan for this research is to carry out a practice-based exploration of potential library interface metaphors through a set of speculative interface prototypes.
The murmuration: a metaphor for discovery?
Prototypes developed during this research will initially seek to explore order and its role in library interfaces by employing the metaphor of a bird murmuration.
I want to take a quick moment to explain why a murmuration and how as I have noticed that sometimes it can be taken in a literal manner, rather than as a metaphor.
A flock of birds, rather poetically known also as a murmuration, is essentially a swarm. It is self-organising. Larger patterns emerge from interactions between individual actors, rather than according to an overall system. Order is generated by emergence.
In terms of a flock of starlings, for instance, individual birds respond to their environment, things like predators will make individual birds move and then ones surrounding those birds move in turn. Each responds to their immediate environment.
I am very aware that much work has been done on flocking or swarming behaviour in relation to information, as well as a number of other systems. I am interested in the visual, metaphorical and poetic properties of the murmuration.
There are a few areas where I think the murmuration could offer potential:
A moving and porous collection
When you think about the library collections we have now they are porous, there is a blurriness around their boundaries. They are also moving collections. What is in and out of the collection isn’t as easy to define as when we had only books on shelves. You can link to almost anything from the online catalog – for instance the UTS catalog links to academic research blogs – does this mean they are now in the collection? The swarm changes shape and size continually. It’s not a defined group. If we start to think about the collection in this way we begin to appreciate that library collections are continually moving, both in terms of what is in and out of the collection and in terms of patron engagement with it.
The mumuration could be used to experiment with exploration of the collection and conceptions of order within the collection itself. Each piece of the murmuration can be viewed from different angles, connections can be made, undone, remade.
I also would like to explore the unpredictable nature of a swarm. A swarm alogrithm, so I believe, is unpredictable. In terms of a speculative interface I think this property is valuable. What would happen if we allowed this unpredictability to play out in an interface? Not randomness per se, but something that allows for chance.
The murmuration is an incredibly poetic image that appeals to us visually, as well as physically through it’s movement. Watch this video to see what I mean. I mean, LOOK at that thing. It’s gorgeous. For some of us, that’s what we get when we browse the shelves. Something very much beyond the transactional, the single search box and the functional.
I think the murmuration allows us to ask: How does this representation change the library patron’s experience and perception of the collection? What affordances are created when the collection is experienced as a murmuration? And, what are the likely consequences of this change?