The project so far and a proposal
This project began as a one-semester piece. It has turned into something larger and where it will end up is not certain. I would like to turn it into a larger research piece for either a research Masters or PhD. I would consider it to be in early proposal stage.
The two of us who are doing the Info Viz studio were tasked with telling a story about the UTS Library collection using information visualisation. The UTS Library is undergoing changes under the stewardship of Mal Booth. The most significant for me was the relocation of the majority of its collection to an underground automated storage and retrieval system, or the LRS. This is like the old closed stack concept, except robots get your books, not library assistants.
Of course, my first reaction was: “But how will people browse? I need to touch the books!” Won’t this limit the ability of patrons to find something serendipitously? If you can’t see where the book is how you can see other things around it that may be of interest?
This is not an uncommon reaction.
Whether or not the inability to touch and see books will limit the potential for serendipitous discovery is uncertain. What the LRS does open up is an opportunity to explore whether a) people actually want to discover: is it a valuable and critical exercise and b) what other ways can discovery be facilitated in a digital environment.
When we were doing our preliminary work mapping the field some questions came up for me:
- Does anyone browse Dewey?
- Has the single search box culture led to a group of scholars who will go no further?
- Are some people natural finders, others natural browsers?
- Has browsing become a romantic ideal?
- Have we lost the drive to discover?
- Is it really a case of near-enough-is-good-enough for uni students and what effect might this have on scholarship?
My wish to answer these questions was fed by my long-standing interest in classification and its relationship to how we find things. Not so much how we search but how do things appear for us to find? Can I create an interface that will, almost surreptitiously, give people things to discover? Search could be a starting point but can I, ambush almost, the behaviour of searchers to give them opportunities to discover and find? To give them some magic? For instance, what if we made things harder to find?
One of my initial ideas was a big interface that people could touch. It would be an interface to the library collection but not necessarily a catalog. I imagined the collection arranged in small neat rows, until someone touched one. Then the collection would scatter and re-sort according a to a rule set by the user or some other, more random rule.
I thought of this as category play. Can the interface provide connections between categories that were random, unexpected but still useful. The data would be provided by the catalog but obvious hurdles came up immediately, like the possible need to provide essentially a whole new set of metadata in order for the novel connections to be displayed.
I then went off and did a whole lot of reading about swarm theory and information flocking. (What if each piece of the collection was given it’s own ‘bird’, it’s own set of rules to govern behaviour and then set loose? What would the catalog do? How would it arrange itself?)
I began envisioning information flocking, information exhibiting emergent behaviours that would lead to beautiful visualisations that look something like this:
I began hand plotting representations that looked like this:
But after looking at the data I had from the library catalog and talking to the wonderful people at the library who run the online catalog I realised I was trying to force the information into a pattern that it would not fit. I had to just let it do what it wanted to do.
- performed a search on a keyword, in this case, football
- plotted out each result on an x-y axis where y = 001 – 999 and x = the bits after the decimal point. So you would be able to see everything in the 700s along a horizonal line.
- put that data in Keynote to produce a prototype for a touchable version of a catalog.
It looks like this:
This is the top 200 results, in order, when searching for “football” in the UTS library catalog. It is an early prototype for a touch interface to enable discovery.I am very keen to allow a physical interaction with the results, to enable sorting and discarding through sweeping movements. One aim is to reveal connection by visually portraying location.
Possible directions from here
As I said, this is a very early prototype. One area I would like to investigate is to delay access to close detail. For example, can Franco Moretti’s ideas of ‘distant reading‘ be used to think about how we engage with search? Can we delay ‘close reading’ of the results to spend more time looking at it from a distance in order to see connections and hopefully, to discover? It might also be fruitful to think about it with Moretti’s emphasis on geometry as opposed to geography: diagrams as opposed to maps. The focus is on connection and pattern instead of definitive location.